www.nature.

com/scientificreports

OPEN Efficient implicit constraint

handling approaches
for constrained optimization
problems
Iman Rahimi 1, Amir H. Gandomi 1,2*, Mohammad Reza Nikoo 3, Mohsen Mousavi 1,4 &
Fang Chen 1
Many real-world optimization problems, particularly engineering ones, involve constraints that make
finding a feasible solution challenging. Numerous researchers have investigated this challenge for
constrained single- and multi-objective optimization problems. In particular, this work extends the
boundary update (BU) method proposed by Gandomi and Deb (Comput. Methods Appl. Mech. Eng.
363:112917, 2020) for the constrained optimization problem. BU is an implicit constraint handling
technique that aims to cut the infeasible search space over iterations to find the feasible region faster.
In doing so, the search space is twisted, which can make the optimization problem more challenging.
In response, two switching mechanisms are implemented that transform the landscape along with
the variables to the original problem when the feasible region is found. To achieve this objective, two
thresholds, representing distinct switching methods, are taken into account. In the first approach,
the optimization process transitions to a state without utilizing the BU approach when constraint
violations reach zero. In the second method, the optimization process shifts to a BU method-free
optimization phase when there is no further change observed in the objective space. To validate,
benchmarks and engineering problems are considered to be solved with well-known evolutionary
single- and multi-objective optimization algorithms. Herein, the proposed method is benchmarked
using with and without BU approaches over the whole search process. The results show that the
proposed method can significantly boost the solutions in both convergence speed and finding better
solutions for constrained optimization problems.

Keywords Constraint handling, Multi-objective optimization, Evolutionary algorithm, Boundary updating,

Switching point

Constrained optimization problems arise naturally in most disciplines where finding optimized feasible solu-
tions, especially with multiple objectives, is challenging. Despite the considerable variety of techniques developed
in optimization fields and other disciplines to tackle these problems, the complexity of their solutions calls for
alternative solution methods. Moreover, the computational cost has become another major concern due to the
complexity of modern real-world problems. Also, the demand for optimal design and its applications in engi-
neering and industry have become even more significant to satisfy the need for more strategic designs in modern
engineering ­practices1. Engineering optimization problems usually consist of constraints that may be physical,
geometrical, or operational; handling such constraints to find a single feasible solution is a challenging t­ ask2,3.
For multi-objective problems, constraint handling is critical since a set of feasible solutions, i.e., Pareto front set,
is sought. To tackle real-world constrained optimization problems, the constraint handling technique (CHT) has
been combined with an evolutionary algorithm (EA) to achieve constrained evolutionary algorithm optimization
(CEAO)2. Most CHTs proposed in the literature are explicit methods, e.g., CHTs by penalty or other fix-ups2.
The CHTs come in two main forms: explicit and implicit. Explicit techniques involve the explicit definition of
constraints as part of the problem formulation. They are designed to respect and enforce constraints explicitly,
guiding the search towards feasible regions of the solution space and addressing violations directly.

1
Data Science Institute, University of Technology Sydney, Sydney, NSW, Australia. 2University Research and
Innovation Center (EKIK), Óbuda University, 1034 Budapest, Hungary. 3Department of Civil and Architectural
Engineering, Sultan Qaboos University, Muscat, Oman. 4School of Civil and Environmental Engineering, University
of New South Wales, Sydney, NSW, Australia. *email: [email protected]

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 1

Vol.:(0123456789)
 www.nature.com/scientificreports/

For example, several existing CHTs in the literature are based on feasibility and infeasibility r­ egions4–6, pri-
ority ­assignment7, and tournament selection and a selection ­operator7,8. Some studies also provide surveys of
­CHTs8–11. In contrast, implicit techniques do not necessitate the explicit form of constraints. They handle con-
straints inherently during the search for the optimal solution, avoiding the need for additional objective function
evaluations. Implicit CHT simplifies the problem formulation, but it may face challenges in handling complex
or nonlinear constraints effectively.
For instance, Diwekar and ­Rubin12 proposed an implicit constraint handling technique for the ASPEN chemi-
cal process, which is based on a mixed integer mathematical programming approach consisting of master and
subproblems. Since most of the relations in the above-mentioned problems are i­mplicit12, proposed a method
for partitioning the variables to significantly decrease computational time.
Raghavan et al.13 developed an implicit constraint handling technique for optimization based on the proper
orthogonal decomposition (POD) of shapes, further producing a bi-level reparameterization approach for struc-
tural geometries. Uemura et al.14 proposed a real-coded genetic algorithm for implicitly constrained black-box
optimization, in which the method employs the weighted mean of the best individuals in a population to find
the optimal solution.
Mirabel and L ­ amiraux15 proposed a method to handle all constraints explicitly and implicitly that deals
with manipulation planning, where the constraints are solved explicitly as much as possible. Implicit handling
is alternatively employed with few variables in case all constraints cannot be handled explicitly. Nomura et al.16
introduced a natural evolution strategy called DX-NES-IC for implicit handling of constrained black-box opti-
mization, which demonstrated better performance than DX-NES, xNES, and CMA-ES.
The current study extends the work of Gandomi and ­Deb6 by extending and applying boundary update (BU)
to single-objective and multi-objective optimization problems (MOOPs). As novel strategies, two switching
mechanisms are suggested that transform the landscape and variables to the original problem when the feasible
region is found. Then, the optimization process is continued without the BU method.
In the work by Gandomi and ­Deb6, the authors proposed a novel approach, boundary update (BU), as an
implicit CHT that updates variable bounds by directly using the constraints and then applying them to several
single-objective optimization problems. Since, the BU method is an implicit CHT, it should be coupled with
an explicit CHT, that is feasibility rules in our s­ tudy17. As for the strategy without BU, only an explicit CHT
(feasibility rules) is applied to solve the problem. The BU method is an implicit constraint handling approach
that aims to cut the infeasible search space over iterations to find the feasible region faster. Although the BU
method directs the search operators to the feasible space and reaches the first feasible solution reasonably fast,
it twists the search space, making the optimization problem more challenging. The current study tries to tackle
the above-mentioned problem by proposing two switching approaches.
The current study augments existing research by introducing two novel switching mechanisms. In the initial
method, called Hybrid-cvtol, the BU approach is employed until constraint violations reach zero, ensuring zero
violations across the entire population. Subsequently, the algorithm transitions to the optimization process
without utilizing the BU approach. The second switching mechanism, called Hybrid-ftol, in this study involves
employing the BU method during the initial phase until the objective space remains unchanged for a specified
number of generations. At that point, the optimization problem shifts to a state without utilizing the BU approach.
The remainder of the paper is organized as follows. Section “Proposed approach” presents the proposed
approach. Section “Research methodology” discusses the research methodology. Section “Numerical examples”
provides numerical examples. Finally, Sects. “Discussion”, “Limitation”, and “Conclusion” give the discussion,
limitation, and conclusion.

Proposed approach
A constrained optimization problem can be formulated in Eqs. (1–3) as follows:
Maximize(Minimize)F(x) = (f 1 (x), . . . , f t (x)) (1)

s.t.hi (x) ≤ 0f ori ∈ {1, . . . , n} (2)

g j (x) = 0forj ∈ {1, . . . , m} (3)

LB ≤ x ≤ U B (4)
where F(x) is the objective vector that consists of several objectives (t is the number of objective functions); n
and m are the number of inequality and equality constraints, respectively; and x is the decision variable bounded
by the lower bound (LB) and upper bound (UB) vectors. These equations yield several Pareto optimal solutions
rather than producing a single ­solution18. In the case of single-objective optimization, a single solution can be
found as a solution to F(x) = f1 (x). In the optimization process, the boundaries of decision variables, represented
by lower bounds (LB) and upper bounds (UB) vectors, undergo dynamic changes as the algorithm progresses
through iterations. This dynamic nature enables the optimization process to adapt and explore a continually
evolving solution space, illustrating the inherent variability and adaptability of decision variable boundaries
throughout the iterative stages. Hence, the iterative nature of optimization involves the dynamic evolution of
the i-th decision variable boundaries throughout each iteration.
The proposed method uses the constraints to narrow down variable space and then forces the algorithm to
focus its search in the feasible region by limiting the viable search space for the variable(s). In the BU method,

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 2

Vol:.(1234567890)
www.nature.com/scientificreports/

the boundaries change iteratively and are updated during the optimization procedure. Mathematically, it could
be written as f­ ollows6:

(5)
   
∃i ∈ {1, . . . , m} : [∀j ∈ {1, . . . , n} : xi ≥ li,j x�=i ∪ xi ≤ ui,j x�=i ]

where li,j and ui,j are dynamic lower bound and upper bound for the ith decision variable, respectively. Updating
the bounds involves the following scenarios:
If lbi = −∞andubi = +∞ : (6)

lbui = li,j (x�=i ) (7)

ubui = ui,j (x�=i ) (8)

Else:
lbui = min(max li,j x�=i , lbi , ubi ) (9)
   

ubui = max(min ui,j x�=i , ubi , lbi ) (10)

   

where (lbui , ubui ) are updated boundaries. BU begins by selecting a repairing variable that can handle the most
significant number of constraints without overlapping other repairing variables. As such, if there is more than
one candidate, the one that handles the greatest number of constraints is selected. Also, if there is still another
candidate for variable selection, one variable is chosen ­randomly6.
If a repairing variable handles the first ki constraints, then:
lbui = min(max[li,1 x�=i , . . . , li,ki x�=i , lbi ], ubi ) (11)
   

ubui = max(min[ui,1 x�=i , . . . , ui,ki x�=i , ubi ], lbi )(whereki ≤ m) (12)

   

If another repairing variable (xr ) is defined:

lbui = min(max li,1 x�=i,r , . . . , li,ki x�=i,r , lbi , ubi ) (13)
     

ubui = max(min ui,1 x�=i,r , . . . , ui,ki x�=i,r , ubi , lbi ) (14)

     

lbur = min max lr,ki +1 x�=r , . . . , lr,ki +kr x�=r , lbr , ubr (15)
       

ubur = max min ur,ki +1 x�=r , . . . , ur,ki +kr x�=r , ubr , lbr wherer ∈ {1, . . . , n}andr �= i (16)
       

In the BU method, a repairing variable could be substituted with a generalized semi-independent v­ ariable6
and then rewritten with lower and upper bounds as:

(17)
   
x ∈ x1 , . . . , xh , xh+1 , . . . , xn → mx ∈ p1 , . . . , ph , xh+1 , . . . , xn

where p1 , . . . , ph are semi-independent variables; and mx indicates the set of selected repairing variables.

(18)
   
x ∈ x1 , . . . , xh , xh+1 , . . . , xn → mx ∈ p1 , . . . , ph , xh+1 , . . . , xn

xi = lbui + pi (ubui − lbui )fori = 1, . . . , hwhere0 ≤ pi ≤ 1 (19)

After selecting repairing variables, the search operator will be applied to the problem, the boundaries of
non-repairing variables will be checked, and the mx-vector will be updated. During the solution procedure, the
boundaries of repairing variables will be updated, and then the semi-independent variables ( pi , i = 1, . . . , h) are
remapped to the actual variables using updated boundaries. In the end, those constraints that were not involved
in the repairing variable boundary, along with fitness values, will be evaluated using actual variables. Algorithm 1
illustrates the BU method in a constrained optimization problem.

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 3

Vol.:(0123456789)
 www.nature.com/scientificreports/

Initialization
While the criteria are not met
Applying the search algorithm
Applying the BU method and updating the mx-vector
Check the boundaries of non-repairing variables
Update the boundaries of repairing variables
Remap p variable to the original variables using updated boundaries ( , )
Evaluate the constraint violations
Evaluating fitness function
End while

Algorithm 1.  Implementation of the BU method in a constrained optimization problem.

As an illustrative example, the following multi-objective optimization problem (MOOP) is considered to
explain BU:

minf1 (x) = (x12 + x22 ) (20)

minf2 (x) = −(x1 − 1)2 − x22 (21)

g1 (x1 + x3 ) = x1 + x3 ≤ 1 (22)

−2 ≤ x1 ≤ 2 (23)

−2 ≤ x2 ≤ 2 (24)

−2 ≤ x3 ≤ 2 (25)
Either x1 orx3 can be selected as a repairing variable of the single constraint. All the variables are in the range
of − 2 to 2. From the variable selection s­ trategy6, as explained earlier, x3 is selected as a repairing variable. The
inequality constraint is solved with respect to x3 as the lower and upper bound functions. The repairing variable
( x3) is substituted with the mapped variable, p, in the range of 0–1. Therefore, the lower and upper bounds of
the variables are {− 2, − 2, 0} and {2, 2, 1}, respectively. When the search algorithm is applied to the problem, the
repairing variable is substituted with the mapped variable, and the boundary of the repairing variable is mapped
for the rest of the search ­cycle6. The original variable can be calculated by the following formula to compute the
objectives and constraints:
x3 = lbu3 + pi × (ubu3 − lbu3 ) (26)
where ubu3 andlbu3 are the lower and upper bounds for the repairing variable x3, respectively. The updated bounds
can be written as follows:
lbu3 = min(max(lb3 , 1 − x1 ), ub3 ) (27)

ubu3 = max(min(ub3 , 1 − x1 ), lb3 ) (28)

The repairing variable is remapped to the actual boundary when the fitness values are evaluated. Figure 1a–d
compare Pareto-optimal fronts found by the BU method coupled with a CHT (here, feasibility rules) and by the
method without the BU method (only the feasibility rules approach is considered) for a population size of 100
for different generations. Figure 1 shows that the BU method produces more non-dominated solutions than
the approach without BU. However, the number of non-dominated solutions is not always a good indicator
for comparison. In the following sections, other indicators are explained in detail to compare the algorithms.
Although the BU method directs the search operator to the feasible space and reaches the feasible region in
a reasonable time, it twists the search space, which can make the optimization problem more challenging and,
hence, does not always bear good results and could potentially result in premature convergence. Consider the
following example:

Maxf (x) = x12 + x22 − 2 × x1 − 2 × x2 + 2 (29)

s.t.g1 (x) : − 3 × x1 + x22 − 5.5 ≤ 0 (30)

 

g2 (x) : −(x12 + 2 × x2 − 4.5) ≤ 0

g3 (x) : 0.8 + x13 − x2 ≤ 0

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 4

Vol:.(1234567890)
www.nature.com/scientificreports/

Figure 1.  Pareto-optimal fronts obtained with and without the BU method.

0 ≤ x1 ≤ 4 (31)

0 ≤ x2 ≤ 4 (32)
In case of selecting variable x2 as a repairing variable, the design space for the original variables x1 and x2
(Fig. 2a) is converted to the new design space (Fig. 2b). Twisting landscape may result in challenges for optimi-
zation problems, such as premature convergence. Some of these issues are presented in the following examples.

Figure 2.  Contour plot of the search space.

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 5

Vol.:(0123456789)
 www.nature.com/scientificreports/

Therefore, two distinct hybrid methods are introduced, each incorporating its own unique switching mecha-
nism. In the initial approach, the BU method is employed until all constraint violations are reduced to zero,
ensuring the absence of violations across the entire population. Subsequently, the algorithm transitions to the
optimization process, excluding the use of the BU approach. In the second switching mechanism, the BU method
is utilized in the initial phase until the objective space remains unchanged for a specified number of generations.
Following this, the optimization problem transitions to a state where the BU approach is no longer employed.
In both approaches, the optimization process is halted once the criteria and threshold tolerance are met. In the
subsequent phase, the optimization problem is solved without the BU method, using the final population from the
previous phase as the initial population for the new optimization problem. Figure 3 illustrates that the handling
of boundaries and constraints occurs after the application of search operators.
During the first phase, the proposed methods simultaneously check constraints and examine variable bounda-
ries, making the optimization process an iterative search procedure for handling constraints and boundaries
concurrently. If the criteria are met, the second phase of the optimization process is conducted without utilizing
the BU approach, constituting a sequential iterative search procedure where constraint handling occurs after
boundary checking. The implementation of the proposed hybrid method is detailed in Algorithm 2.

Figure 3.  Optimization process of the proposed method.

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 6

Vol:.(1234567890)
www.nature.com/scientificreports/

Initialize the parameters

Set a threshold tolerance for constraint violation and objective space changes
While all constraints are handled
Find the repairing variable (s)
If there is more than one candidate
Select the one (s) that handle/s the greatest number of constraints
If there are still other candidates
Choose randomly
End if
End if
End while
While the criteria are not met (here threshold tolerance: constraint violation should reach zero/ no change in objective
space in the last n generations)
Applying the search algorithm
Update mx-vector
Check the boundaries of non-repairing variables ( ℎ+1,…, )
For i=1 to h (repairing variables)
Update the boundaries of repairing variables,
Remap p variable to the original space for post-processing
End for
Check the constraint violation
End while
Stop the optimization process using the BU method
Initialize the algorithm using the final population of the previous optimization process
Set a termination criteria (e.g., number of evaluations, number of generations, time, etc.)
While the criteria are not met
Apply the search algorithm
Optimize the problem without the BU method
Evaluate fitness function
End while

Algorithm 2.  Implementation of the proposed method (hybrid methods).

Research methodology
To evaluate the proposed method, Python and Pymoo ­libraries19 were used, which include well-stabilized algo-
rithms. The specific algorithm parameters are presented in Table 1. Furthermore, because the implemented
algorithms act stochastically, each of these algorithms were ran 31 times with random initial point(s).
Several examples, including benchmarks and real-world problems, were solved using three methods (with-
out the BU method, with the BU method, and hybrid method). The presented examples are of different types:
constrained by one, two, and three objective functions. For the problem with one objective function, the pro-
posed method was implemented using evolutionary strategy (ES), differential evolution (DE), stochastic ranking

Algorithm Parameter settings

Evolutionary strategy Number of offspring = 200, rule = 1.0/7.0
Differential evolution Variant = "DE/rand/1/bin", crossover constant (CR) = 0.3, Weighting factor (F) = 0.8
Stochastic ranking evolutionary strategy Number of offspring = 200, rule = 1.0/7.0, gamma = 0.85, alpha = 0.2
Biased random key genetic algorithm Number of elite = 200, number of offspring = 700, number of mutant = 100, bias = 0.7

Table 1.  Specific parameter settings for the algorithms used in this study.

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 7

Vol.:(0123456789)
 www.nature.com/scientificreports/

evolutionary strategy (SRES)20, and biased random key genetic algorithm (BRKGA)21. For two objective func-
tions, the BU method was applied using NSGA-II17,22. For three and five objective functions, some other EAs were
used. As such, NSGA-III23, U-NSGA-III24, AGE-MOEA25, ­NSDE26, and N ­ SDER27 were applied to optimization
problems. The following metrics were evaluated to assess the accuracy of the obtained results.

• Number of non-dominated solutions: The BU method was coupled with an explicit constraint handling
technique and compared to the approach without BU (only feasibility rules were applied to the optimization
problem). Since the BU method aims to reduce the search space, the number of non-dominated solutions
was determined for comparison.
• Constraint violation (CV): Compared to unconstrained optimization problems, constrained optimization
problems are more challenging since a large proportion of infeasible regions appears in the search space
(meaning the hit ratio is low), especially for highly constrained problems where these regions lead to some
difficulties, such as feasibility-hardness and convergence-hardness. Therefore, even finding one feasible solu-
tion is a significant achievement. To test the effectiveness of the proposed method, the first feasible solutions
found by the methods were compared and the whole population was tracked against generations.
• Performance indicators: Generational Distance (GD)28, Generational Distance Plus (GD+)29, Inverted Gen-
erational Distance (IGD)30, Inverted Generational Distance Plus (IGD+)29, and Hypervolume (HV) are the
indicators that were analyzed.
• Running ­metric31: It is possible to trace the difference in the objective space followed by generations.

Moreover, the population was fixed at 100, and the maximum number of experiments was set to 11.

Numerical examples
The proposed approach was applied to several single-objective and multi-objective optimization examples,
including benchmarks and real-world problems. The presented examples include problems constrained by one,
two, and three objective functions. The performance of the three strategies was evaluated using an example that
has linear constraints. Following this numerical example, the proposed approach was evaluated to solve a sur-
rogate model for a car-side impact design problem. This design problem consists of two versions of optimization,
i.e., single- and multi-objective optimizations, and is an example of a black-box optimization problem. The speed
reducer problem, a single-objective constrained engineering optimization problem, was also considered. For the
multi-objective optimization problem, two benchmarks, namely OSY and BNH, and three real-world constrained
optimization problems, including welded beam design, Cantilevered Beam design, and multi-objective car-side
design problems were solved via the three strategies.
Several metaheuristic methods have been developed and implemented to reduce the total runtime of the
optimization ­problems2,32, but still produce results that are almost as accurate and precise as conventional solving
methods. Also, real-world engineering problems involve some type of optimization that is often constrained,
most of which are considered MOOPs. No single solution exists for a MOOP; instead, different solutions gener-
ate trade-offs for various objectives. Furthermore, MOOPs arise naturally in most fields, and solving them has
been a challenging problem for ­researchers32,33. EA methods have been identified as more effective in tackling
the challenges that arise from MOOPs, for which the form of the Pareto-optimal front (discontinuity, noncon-
vexity, etc.) is not i­mportant34,35. Moreover, most multiobjective evolutionary algorithms (MOEAs) use the
dominance ­concept36,37. In this work, for the single-objective optimization problems, Genetic Algorithm (GA),
Differential Evolution (DE)38, Evolutionary Strategy (ES)39, Stochastic Ranking Evolutionary Strategy (SRES)20,
and Biased Ranking Key Genetic Algorithm (BRKGA)40 were considered. NSGA-II41 is applied to bi-objective
and multi-objective optimization problems with three objective functions and to multi-objective optimization
problems with more than three objective functions, which are called many-objective optimization problems.
NSGA-III23, U-NSGA-III24, and Adaptive Geometry Estimation-based Multi-objective Evolutionary Algorithm
(AGE-MOEA)25 were further implemented for optimizing the problems. For all problems, the BU method was
coupled with feasibility rules proposed by D ­ eb41. Table 2 presents the studied examples in this work to assess the
effectiveness of the proposed method in comparison with the other methods.

Problem Type of problem Example No. of objective/s No. of constraints No. of variables
1 Benchmark G1 ­problem22 1 9 13
2 Real-world Single-objective car-side ­problem42 1 10 11
3 Real-world Speed ­reducer43 1 11 7
4 Benchmark BNH5 2 2 2
5 Benchmark OSY44 2 6 6
6 Real-world problem Welded beam ­problem45 2 4 4
7 Real-world problem Cantilevered Beam design ­Problem46 2 10 10
8 Real-world problem Multi-objective car-side ­problem47 3 10 7

Table 2.  Test problems considered in this study.

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 8

Vol:.(1234567890)
www.nature.com/scientificreports/

Single optimization problems

In this section, three single-optimization problems, namely G1, car-side design problem, and speed reducer, are
considered to be solved using the different strategies discussed above.

G1 problem
The G1 problem is a famous mathematical constrained optimization problem that includes 13 decision vari-
ables, a quadratic objective function, and nine linear constraints. The problem is an excellent example of a highly
linearly constrained problem, with a feasibility ratio of 0.111%6 (see the Appendix). In other words, it is a highly
challenging to find a feasible solution for this constrained problem. The three different strategies were applied
to this problem. According to the variable selection strategy, x10 , x11 , and x12 were selected as repairing variables
for handling all nine constraints. The global optimum for this problem is -15, and the optimization results by
using GA, DE, ES, SRES, and BRKGA with the BU method, without the BU method, and hybrid methods are
presented in Tables 3, 4, 5 and 6 and Fig. 4. The Friedman rank test at a significant level of 5% was used to detect
differences in treatments across test attempts (Table 4). Also, the Post-hoc Nemenyi Friedman test was performed
to identify exactly which groups have different means (Tables 5, 6).
The comprehensive analysis of algorithm performance on the G01 optimization problem, characterized by
13 continuous decision variables, nine linear constraints, and a quadratic objective function, unveils intricate
relationships between problem features and algorithmic behavior. Each algorithm’s response to the problem’s
unique attributes provides valuable insights into their adaptability and efficacy.
The hybrid methods, consistently showcasing superior performance across diverse algorithms, reflect an
inherent versatility that aligns with the multi-faceted nature of the G01 problem. The distinct advantage of the
hybridization approach becomes particularly evident in the handling of constraints. The BU method, serving

Algorithm BU step Best Mean Median Worst St. Dev

Nonea − 13.83 − 12.87 − 13.07 − 11.99 0.68
BU − 14.43 − 13.94 − 14.17 − 13.02 0.48
GA
Hybrid-cvtol − 15.0 − 14.87 − 14.95 − 14.58 0.15
Hybrid-ftol − 15.0 − 14.99 − 14.99 − 14.99 0.00
None − 10.08 − 8.83 − 8.74 − 8.10 0.66
BU − 14.02 − 12.61 − 13.46 − 9.55 1.62
DE
Hybrid-cvtol − 14.80 − 14.718 − 14.710 − 14.66 0.047
Hybrid-ftol − 14.35 − 14.02 − 14.05 − 13.68 0.21
None − 10.75 − 9.24 − 9.38 − 8.13 0.89
BU − 10.89 − 9.89 − 9.99 − 8.98 0.65
ES
Hybrid-cvtol − 13.41 − 10.47 − 9.84 − 8.63 1.63
Hybrid-ftol − 10.77 − 9.92 − 10.0 − 8.98 0.57
None − 7.40 − 5.68 − 5.85 − 3.63 1.24
BU − 9.05 − 8.27 − 8.18 − 1.04 0.44
SRES
Hybrid-cvtol − 7.96 − 5.10 − 5.04 − 1.04 1.68
Hybrid-ftol − 8.38 − 5.51 − 4.54 − 3.27 1.84
None − 13.83 − 12.46 − 12.20 − 11.30 0.89
BU − 14.43 − 13.56 − 13.85 − 11.54 1.03
BRKGA
Hybrid-cvtol − 14.90 − 14.88 − 14.88 − 14.86 0.01
Hybrid2-ftol − 14.93 − 14.87 − 14.88 − 14.85 0.02

Table 3.  Statistical results of different methods on G1 problem (The best performing method (objective
function values) is marked in bold). a “None” means no BU method or hybrid is used and only feasibility rules
as an explicit constraint handling method is worked alone.

Algorithm p-valuea Statistic value

GA 0.002 14.62
DE 0.002 14.03
ES 0.06 7.28
SRES 0.25 4.07
BRKGA 0.005 12.59

Table 4.  Summary of the p-value of the Friedman rank test over all runs. a All statistical test in this study was
performed at a significant level of 5%.

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 9

Vol.:(0123456789)
 www.nature.com/scientificreports/

Hybrid-cvtol vs. Without

Hybrid-cvtol vs. BU BU BU vs. Without BU
Algorithm p-value h p-value h p-value h
GA 0.45 ~ 0.03 + 0.59 ~
DE 0.12 ~ 0.001 + 0.45 ~
ES 0.90 ~ 0.25 ~ 0.15 ~
SRES 0.22 ~ 0.59 ~ 0.90 ~
BRKGA 0.12 ~ 0.01 + 0.87 +

Table 5.  Summary of the p-value of the posthoc_nemenyi_friedman over all runs. Posthoc_nemenyi_
friedman: +, ~, and – presents the first method performs statistically significantly better, equal, and worse than
the second approach. Significant values are in bold.

Hybrid-ftol Hybrid-ftol
Hybrid-ftol vs. Without vs. Hybrid-
vs. BU BU cvtol
Algorithm p-value h p-value h p-value h
GA 0.12 ~ 0.003 ~ 0.87 ~
DE 0.87 ~ 0.12 ~ 0.45 ~
ES 0.90 ~ 0.06 ~ 0.90 ~
SRES 0.59 ~ 0.90 ~ 0.90 ~
BRKGA 0.20 ~ 0.03 + 0.90 ~

Table 6.  Summary of the p-value of the posthoc_nemenyi_friedman over all runs (continue). Significant
values are in bold.

as an implicit constraint-handling technique, demonstrates a keen ability to navigate the infeasible search space
effectively. This characteristic proves invaluable in problems like G01, where constraints play a crucial role in
shaping the feasible region. The hybrid methods, by strategically integrating the BU method, adeptly leverage
this constraint-handling mechanism, enabling faster convergence to the feasible area.
Considering the explicit constraints and the quadratic objective function in G01, it becomes apparent that
the algorithmic adaptation to these features significantly impacts performance. Genetic algorithms (GA), when
combined with the BU method (as hybrid methods), exhibit a noteworthy convergence pattern (as illustrated in
Fig. 5). The initial phase, focused on constraints handling, ensures the algorithm converges to the feasible area
efficiently. Subsequently, the transition to the optimization process without BU capitalizes on the algorithm’s
newfound knowledge of the feasible region, resulting in a streamlined path to the final solution. This adaptive
approach to the problem’s structure showcases how algorithmic behavior is intricately linked to specific features,
allowing for a dynamic and context-aware optimization process.
The inadequacy of the SRES algorithm in this context sheds light on the sensitivity of algorithmic performance
to the problem’s characteristics. The SRES algorithm, designed for stochastic optimization, might face challenges
when confronted with the structured nature of the G01 problem. The lack of significant improvement suggests
that algorithmic suitability is not universal but rather contingent on the problem’s intricacies. This observation
underscores the importance of aligning algorithmic choices with the specific features and constraints of the
optimization problem at hand.
Furthermore, the reduction in standard deviation observed in the hybrid methods implies a more stable
and reliable convergence behavior. This heightened stability is particularly valuable in the presence of complex
problem features, such as the quadratic objective function, where rapid fluctuations in convergence patterns
might hinder progress. The hybrid methods’ ability to maintain a consistent trajectory towards the optimal
solution underscores their resilience in handling the intricacies introduced by specific problem characteristics.
In summary, the algorithmic performance on the G01 problem intricately relates to its unique features. The
hybrid methods showcase adaptability and versatility in addressing constraints, demonstrating a keen under-
standing of the problem structure. The nuanced convergence patterns, strategic switching mechanisms, and
sensitivity to problem-specific attributes provide valuable insights into the dynamic interplay between algorithmic
behavior and the intricacies of the optimization problem.

Car‑side problem
The car-side impact design problem, a notorious challenge in engineering, stands out as a complex and time-
consuming optimization task, as depicted in Fig. 6. This problem, adaptable to both single-objective and multi-
objective formulations, adds layers of intricacy due to its quadratic objective function and explicit models.
Specifically, when addressing the single-objective version with the goal of minimizing car weight, the constraints
and objective function take on quadratic forms, amplifying the intricacies involved. The thicknesses of the inner

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 10

Vol:.(1234567890)
www.nature.com/scientificreports/

Figure 4.  Violin plot of implementation of using different algorithms constraint handling methods (G01
problem).

floor side ( X3) are strategically selected as decision variables to solve the explicit models, and their bounds are
dynamically adjusted to accommodate the interconnected variables, adding an additional layer of complexity.
The results of this formidable optimization challenge are presented in Tables 7, 8, 9 and 10 and Figure 7,
offering a quantitative assessment of algorithmic performance. Statistical tests (Tables 8, 9, 10) and visualiza-
tions in Figure 7 highlight the superiority of the hybrid methods, particularly when combined with the genetic
algorithm (GA). This dominance is established through the comparative analysis, showcasing that the hybrid
approach consistently outperforms both the individual methods and, notably, requires fewer function evalua-
tions (FE) to attain the final solution.
Significantly, the examination of the results reveals nuanced differences between the differential evolution
(DE) and evolutionary strategy (ES) methods, with the hybrid method showcasing slightly superior outcomes.
The incorporation of the BU method, along with the hybrid methods, demonstrates a tangible advantage, result-
ing in enhanced solution quality. The visual representation of the switching point for GA in Figure 8 provides
further clarity on the optimization process. The algorithm strategically transitions from the BU-assisted phase
to the standard optimization phase, showcasing an adaptive mechanism designed to optimize the convergence

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 11

Vol.:(0123456789)
 www.nature.com/scientificreports/

Figure 5.  Constraint violation implemented by GA on the G01 test problem.

Figure 6.  Schematic of car side impact design ­problem42.

path. However, the study identifies limitations within the context of the car-side impact design problem. While
GA, BU, and hybrid methods exhibit notable improvements, the performance of biased random key genetic
algorithm (BRKGA) and stochastic ranking evolution strategy (SRES) fails to show enhancement when incor-
porated into the hybrid framework. This suggests that BRKGA and SRES may not be suitable for addressing the
intricacies of this particular problem, underscoring the need for algorithmic adaptability in the face of diverse
engineering challenges.
In conclusion, the car-side impact design problem proves to be a formidable optimization challenge, and the
hybrid methods, particularly in conjunction with GA, emerge as the most effective approach. The adaptability
of the hybrid methods, strategic switching mechanisms, and the incorporation of the BU method collectively
contribute to enhanced solution quality and computational efficiency. Nonetheless, the study sheds light on the
context-specific nature of algorithmic suitability, emphasizing the importance of tailoring optimization strategies
to the unique characteristics of complex engineering problems.

Speed reducer
The examination of the speed reducer design problem, featuring one objective function and seven continuous
decision variables, has yielded multifaceted insights into the interplay between algorithmic approaches and
problem-specific characteristics. With the overarching goal of minimizing the total weight of the speed reducer,
the complexity of this task is compounded by a combination of linear and nonlinear constraints. Strategically
leveraging three repairing variables—specifically, the face width and shaft diameters ( x1, x6 , and x7)—proves
instrumental in managing these constraints effectively.
The results, meticulously presented in Tables 11, 12, 13 and 14 and Fig. 9, reveal a consistent trend: the hybrid
approaches consistently outperform standalone methods, yielding the best objective function values across
all algorithms considered. This dominance suggests that the hybridization process contributes to algorithmic
robustness, offering a versatile strategy to navigate the intricate optimization landscape presented by the speed
reducer design problem.
Notably, the algorithms (GA, DE, and BRKGA) exhibit improved performance when integrated into hybrid
frameworks, signifying a synergistic effect where the strengths of individual algorithms complement one another.

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 12

Vol:.(1234567890)
www.nature.com/scientificreports/

Algorithm BU step Best Mean Median Worst St. Dev

Nonea 23.889 24.474 24.246 25.295 0.486
BU 23.538 23.558 23.563 23.568 0.011
GA
Hybrid-cvtol 23.522 23.551 23.557 23.566 0.015
Hybrid-ftol 23.566 23.714 23.602 24.189 0.238
None 21.920 22.370 22.142 24.389 0.661
BU 20.797 21.842 21.786 23.128 0.696
DE
Hybrid-cvtol 20.750 21.784 21.782 23.067 0.618
Hybrid-ftol 21.318 21.714 23.602 22.556 0.238
None 20.432 20.770 20.815 21.168 0.251
BU 20.147 20.453 20.540 20.585 0.161
ES
Hybrid-cvtol 20.147 20.584 20.434 21.386 0.422
Hybrid-ftol 20.446 20.746 20.727 21.064 0.234
None 20.518 20.673 20.728 20.760 0.092
BU 20.246 20.424 20.436 20.528 0.101
SRES
Hybrid-cvtol 20.249 20.425 20.436 20.528 0.100
Hybrid-ftol 20.395 20.477 20.432 20.580 0.077
None 18.165 18.904 18.90 19.572 0.493
BU 20.088 20.246 20.202 20.590 0.178
BRKGA
Hybrid-cvtol 19.631 20.082 20.124 20.639 0.346
Hybrid-ftol 19.359 20.127 20.20 20.534 0.411

Table 7.  Statistical results of different methods on car side design impact problem. a “None” means no BU
method or hybrid is used and only feasibility rules as an explicit constraint handling method is worked alone.
Significant values are in bold.

Algorithm p-value Statistic value

GA 0.001 15.00
DE 0.025 9.339
ES 0.162 5.125
SRES 0.0175 10.125
BRKGA 0.020 9.734

Table 8.  Summary of p value of the Friedman rank test over all runs.

Hybrid-cvtol Hybrid-cvtol vs. BU vs.

vs. BU Without BU Without BU
Algorithm p-value h p-value h p-value h
GA 0.59 ~ 0.001 + 0.06 +
DE 0.72 ~ 0.02 + 0.26 ~
ES 0.90 ~ 0.31 ~ 0.20 ~
SRES 0.87 ~ 0.20 ~ 0.03 +
BRKGA 0.87 ~ 0.15 0.02 +

Table 9.  Summary of the p-value of the posthoc_nemenyi_friedman over all runs. Significant values are in
bold.

The statistical tests conducted (Tables 12, 13, 14) affirm the significance of these improvements, providing a
robust validation of the observed enhancements and instilling confidence in the findings.
A closer examination of GA within the hybrid methods reveals its particular suitability for this problem, con-
sistently outperforming standalone and BU-assisted methods. This positions GA as a robust choice for addressing
the complexities inherent in the optimization landscape of the speed reducer design problem.
The incorporation of the BU method consistently leads to a substantial reduction in the number of function
evaluations required to reach the final solution. This reduction in computational cost is a valuable contribution,

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 13

Vol.:(0123456789)
 www.nature.com/scientificreports/

Hybrid-ftol Hybrid-ftol
Hybrid-ftol vs. Without vs. Hybrid-
vs. BU BU cvtol
Algorithm p-value h p-value h p-value h
GA 0.59 ~ 0.59 ~ 0.06 ~
DE 0.90 ~ 0.09 ~ 0.90 ~
ES 0.59 ~ 0.87 ~ 0.73 ~
SRES 0.90 ~ 0.03 + 0.87 ~
BRKGA 0.90 ~ 0.06 ~ 0.90 ~

Table 10.  Summary of the p-value of the posthoc_nemenyi_friedman over all runs (continue). Significant
values are in bold.

Figure 7.  Violin plot of implementation of using different algorithms constraint handling methods (Carside
problem).

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 14

Vol:.(1234567890)
www.nature.com/scientificreports/

Figure 8.  Switch point for GA to normal optimization without the BU method.

Algorithm BU step Best Mean Median Worst St. Dev

Nonea 2997.353 2998.827 2998.933 2999.647 0.645
BU 3119.764 3126.639 3127.78 3129.697 2.792
GA
Hybrid-cvtol 2997.002 2997.964 2998.023 2998.779 0.539
Hybrid-ftol 2996.371 2996.422 2996.404 2996.542 0.055
None 2847.208 2982.631 2996.641 2997.054 42.848
BU 2706.172 2900.272 2881.668 3131.199 157.048
DE
Hybrid-cvtol 2706.141 2838.565 2828.955 2996.729 108.803
Hybrid-ftol 2718.692 2996.422 2996.404 2742.44 0.055
None 3001.0 3023.363 3025.0 3039.00 12.694
BU 3120.0 3123.727 3123.0 3129.00 2.561
ES
Hybrid-cvtol 3007.0 3018.000 3018.0 3030.00 6.281
Hybrid-ftol 3005 3015.428 3015 3028.00 6.694
None 3012.0 3031.272 3025.0 3063.00 17.669
BU 3108.0 3123.181 3125.0 3128.00 5.236
SRES
Hybrid-cvtol 3015.0 3039.000 3038.0 3076.00 17.832
Hybrid-ftol 3108 3122.57 3125 3128 6.298
None 2997.996 3001.692 3001.084 3006.936 2.761
BU 3117.466 3123.452 3123.855 3127.798 3.386
BRKGA
Hybrid-cvtol 2997.312 3000.977 3000.109 3006.877 3.254
Hybrid-ftol 2875.699 2946.35 2951.593 3036.718 56.361

Table 11.  Statistical results of different methods on speed reducer problem. a “None” means no BU method or
hybrid is used and only feasibility rules as an explicit constraint handling method is worked alone. Significant
values are in bold.

Algorithm p-value Statistic value

GA 0.0001 21.0
DE 0.0018 15.000
ES 0.0031 13.799
SRES 0.0003 18.71
BRKGA 0.0031 13.799

Table 12.  Summary of the p-value of the Friedman rank test over all runs.

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 15

Vol.:(0123456789)
 www.nature.com/scientificreports/

Hybrid-cvtol
Hybrid-cvtol vs. Without BU vs.
vs. BU BU Without BU
Algorithm p-value h p-value h p-value h
GA 0.019 + 0.46 ~ 0.46 ~
DE 0.29 ~ 0.06 ~ 0.87 ~
ES 0.06 ~ 0.90 ~ 0.03 ~
SRES 0.014 + 0.90 ~ 0.026 +
BRKGA 0.03 + 0.90 ~ 0.06 ~

Table 13.  Summary of the p-value of the posthoc_nemenyi_friedman over all runs. Significant values are in
bold.

Hybrid-ftol Hybrid-ftol
Hybrid-ftol vs. Without vs. Hybrid-
vs. BU BU cvtol
Algorithm p-value h p-value h p-value h
GA 0.001 + 0.019 + 0.46 ~
DE 0.03 + 0.003 + 0.76 ~
ES 0.002 + 0.82 ~ 0.70 ~
SRES 0.90 ~ 0.02 + 0.014 +
BRKGA 0.002 + 0.70 ~ 0.82 ~

Table 14.  Summary of the p-value of the posthoc_nemenyi_friedman over all runs (continue).

emphasizing the efficiency gains afforded by the BU method, a crucial consideration in engineering optimiza-
tion scenarios.
Despite the positive outcomes, the study highlights that algorithms like BRKGA and SRES do not exhibit
improved performance within the hybrid framework, suggesting their unsuitability for this particular problem.
This observation underscores the importance of algorithmic adaptability and tailoring strategies to specific
problem characteristics.
In conclusion, the insights gained from the speed reducer design problem analysis underscore the effectiveness
of hybrid approaches in enhancing algorithmic performance. The adaptability of GA, efficiency gains with the BU
method, and the nuanced trade-offs between algorithms offer valuable guidance for tackling real-world engineer-
ing optimization challenges. These findings contribute not only to the understanding of algorithmic behavior in
this specific context but also pave the way for broader applications in engineering design optimization.

Multi‑objective optimization problems

The following subsections provide two benchmarks and two real-world optimization problems to assess the
effectiveness of the proposed hybrid method for multi-objective optimization problems.

Osyczka and Kundu (OSY)

Osyczka and ­Kundu44 introduced the OSY test problem, a widely recognized benchmark featuring two nonlinear
objective functions and six linear and nonlinear constraints. As highlighted earlier, the BU method is a central
focus of this work, particularly in its ability to directly handle constraints and its subsequent integration with
other explicit constraint-handling techniques. The BU approach strategically reduces the feasible search space by
altering one or more variables, compelling the algorithm to concentrate on exploring regions that adhere to the
defined constraints. For the OSY problem, a specific repairing variable selection strategy is employed, targeting
variables x1, x4 , and x6 identified as effective for handling all constraints inherent in the problem.
Following the application of the BU approach, NSGA-II is employed to navigate the complex landscape of
the OSY problem. NSGA-II, known for identifying viable Pareto fronts, initially achieves an entirely feasible
population after 1500 evaluations. In contrast, the coupling of the BU method with the algorithm accelerates this
process, achieving a fully feasible population after only 300 evaluations. Subsequently, the optimization process
seamlessly transitions to routine optimization without the continued use of the BU method.
The evaluation of performance indicators, presented in Tables 15, 16, 17 and 18 and Fig. 10, offers a compre-
hensive understanding of the algorithmic performance. It becomes evident that the hybrid methods, particularly
when incorporating the BU approach, consistently outperform other methods. Statistical analyses, including the
Friedman rank and posthoc tests, validate this observation by demonstrating a significant improvement in all
performance indicators (excluding IGD+) when the hybrid approach is implemented.
Examining Fig. 10 through the lens of a violin plot further illustrates the superiority of the hybrid methods.
The best values for all performance indicators, with the exception of IGD, show marked improvement when the
BU and hybrid methods are employed. This enhancement in performance underscores the effectiveness of the

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 16

Vol:.(1234567890)
www.nature.com/scientificreports/

Figure 9.  Violin plot of implementation of using different algorithms constraint handling methods (speed
reducer problem).

BU method in combination with the hybrid approach, providing valuable insights into the algorithmic strategies
that lead to improved convergence and solution quality for the OSY problem.
In summary, the OSY problem, coupled with the BU method and hybrid approach, serves as a compelling
case study. The specific variable selection strategy, the seamless transition between BU-assisted and routine
optimization, and the consistent improvement in performance indicators highlight the adaptability and effec-
tiveness of the proposed approach. These findings contribute not only to the optimization of the OSY problem
but also offer broader insights into the potential of the BU method and hybridization in addressing complex,
constrained optimization scenarios.

Bin and Korn Test Problem (BNH)

BNH is a test problem proposed by Binh and K ­ orn5, characterized by two conflicting objectives and two con-
straints, introduces a challenging landscape for optimization. The nonlinear nature of the objectives and con-
straints adds an additional layer of complexity, requiring sophisticated algorithmic strategies to navigate the

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 17

Vol.:(0123456789)
 www.nature.com/scientificreports/

Performance indicator BU step Best Mean Median Worst St. Dev

Nonea 0.354 0.652 0.697 0.727 0.12
BU 0.680 0.712 0.698 0.772 0.032
IGD
Hybrid-cvtol 0.636 0.655 0.651 0.677 0.014
Hybrid-ftol 0.412 0.484 0.499 0.503 0.030
None 0.504 0.520 0.526 0.549 0.013
BU 0.503 0.573 0.506 0.754 0.107
IGD+
Hybrid-cvtol 0.491 0.495 0.495 0.500 0.003
Hybrid-ftol 0.400 0.433 0.439 0.439 0.013
None 0.031 0.042 0.046 0.050 0.007
BU 0.002 0.114 0.007 0.403 0.17
GD
Hybrid-cvtol 0.006 0.009 0.010 0.013 0.002
Hybrid-ftol 0.032 0.057 0.038 0.176 0.048
None 0.022 0.034 0.037 0.043 0.006
BU 0.000 0.113 0.006 0.403 0.172
GD+
Hybrid-cvtol 0.001 0.004 0.005 0.009 0.002
Hybrid-ftol 0.030 0.055 0.037 0.175 0.048

Table 15.  Statistical results of different methods on the OSY problem. a “None” means no BU method or
hybrid is used and only feasibility rules as an explicit constraint handling method is worked alone. Significant
values are in bold.

Performance indicator p-value statistic value

IGD 0.0006 17.22
IGD+ 0.0003 18.44
GD 0.0005 17.38
GD+ 0.0006 17.26

Table 16.  Summary of the p-value of the Friedman rank test over all runs.

Hybrid-cvtol
Hybrid-cvtol vs. Without BU vs.
vs. BU BU Without BU
Performance indicator p-value h p-value h p-value h
IGD 0.03 + 0.005 + 0.90 ~
IGD+ 0.10 ~ 0.005 + 0.70 ~
GD 0.019 + 0.010 + 0.90 ~
GD+ 0.04 + 0.026 + 0.90 ~

Table 17.  Summary of the p-value of the posthoc_nemenyi_friedman over all runs.

Hybrid-ftol Hybrid-ftol
Hybrid-ftol vs. Without vs. Hybrid-
vs. BU BU cvtol
Performance indicator p-value h p-value h p-value h
IGD 0.06 ~ 0.01 + 0.90 ~
IGD+ 0.06 ~ 0.002 + 0.90 ~
GD 0.035 + 0.019 + 0.90 ~
GD+ 0.014 + 0.007 + 0.90 ~

Table 18.  Summary of the p-value of the posthoc_nemenyi_friedman over all runs (continue).

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 18

Vol:.(1234567890)
www.nature.com/scientificreports/

Figure 10.  Violin plot of implementation of using different indicators (OSY problem).

intricate solution space effectively. The objectives and constraints may exhibit intricate relationships, posing a
considerable challenge to traditional optimization approaches.
In addressing this nonlinear bi-objective optimization problem, the algorithmic focus will be on developing
robust strategies that balance the optimization of both objectives while adhering to the imposed constraints.
The inherent trade-offs between conflicting objectives necessitate a careful exploration of the Pareto front to
identify solutions that represent optimal compromises. This problem involves two continuous variables, x1 and
x2, which can be strategically chosen as repairing variables. The findings presented in Tables 19, 20, 21, 22, 23,
24 and 25 and Fig. 11 illuminate a substantial disparity between the hybrid method and alternative approaches.
Particularly noteworthy is the superior performance of the hybrid method, especially when objective function
tolerance is considered, showcasing noteworthy enhancements over other methods across various cases. The
results unequivocally demonstrate the hybrid approaches’ consistent outperformance in tackling this bi-objective
optimization problem characterized by nonlinear constraints and objective functions. Furthermore, the per-
formance indicator values associated with the hybrid approaches are significantly smaller compared to their
counterparts, underscoring the effectiveness of the proposed methodology in achieving superior convergence
and solution quality. These insights not only contribute to the optimization of the BNH problem but also offer
valuable considerations for addressing the nuances of nonlinear bi-objective optimization challenges across
diverse applications.

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 19

Vol.:(0123456789)
 www.nature.com/scientificreports/

Performance indicator BU step Best Mean Median Worst St. Dev

Nonea 0.0423 0.0803 0.0834 0.1281 0.0253
BU 0.0255 0.0535 0.0426 0.0896 0.0272
IGD
Hybrid-cvtol 0.0123 0.0289 0.0209 0.0660 0.0172
Hybrid-ftol 0.0110 0.0183 0.0200 0.0222 0.0038
None 0.0170 0.0268 0.0247 0.0433 0.0084
BU 0.0112 0.0222 0.0179 0.0419 0.0111
IGD+
Hybrid-cvtol 0.0024 0.0061 0.0048 0.0143 0.0040
Hybrid-ftol 0.0029 0.0035 0.0033 0.0043 0.0005
None 0.0045 0.0052 0.0050 0.0065 0.0006
BU 0.0061 0.0073 0.0066 0.0098 0.0013
GD
Hybrid-cvtol 0.0044 0.0046 0.0046 0.0049 0.0001
Hybrid-ftol 0.0040 0.0044 0.0046 0.0047 0.0002
None 0.0029 0.0034 0.0033 0.0048 0.0005
BU 0.0039 0.0053 0.0049 0.0079 0.0013
GD+
Hybrid-cvtol 0.0026 0.0027 0.0027 0.0029 0.00009
Hybrid-ftol 0.0020 0.0028 0.0030 0.0031 0.0003

Table 19.  Statistical results of different methods on the BNH problem. Significant values are in bold.

Performance indicator p-value Statistic value

IGD 0.0006 17.22
IGD+ 0.0003 18.44
GD 0.0005 17.38
GD+ 0.0006 17.26

Table 20.  Summary of the p-value of the Friedman rank test over all runs.

Hybrid-cvtol vs. Without

Hybrid-cvtol vs. BU BU BU vs. Without BU
Performance indicator p-value ha p-value h p-value h
IGD 0.244362 ~ 0.003837 + 0.244362 ~
IGD+ 0.244362 ~ 0.003837 + 0.244362 ~
GD 0.001488 + 0.376245 ~ 0.082380 ~
GD+ 0.001000 + 0.147296 ~ 0.147296 ~

Table 21.  Summary of the p-value of the posthoc_nemenyi_friedman over all runs. a Posthoc_nemenyi_
friedman: +, ~, and – presents the first method performs statistically significantly better, equal, and worse than
the second approach.

Method Hybrid-cvtol Hybrid-ftol BU Without BU

Hybrid-cvtol – 0.9000 0.0358 0.0051
Hybrid-ftol 0.9000 – 0.06243 0.01026
BU 0.0358 0.06243 – 0.9000
Without BU 0.0051 0.01026 0.9000 –

Table 22.  Summary of the p-value of the posthoc_nemenyi_friedman over all runs- IGD value comparison.

Welded beam design problem

The welded beam fabrication is a well-known engineering optimization benchmark ­problem48 that examines
the trade-off between the strength and cost of a beam. The problem includes four constraints (Eqs. 32–35) and
four design variables, which are:

• x1 = h, the thickness of the welds

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 20

Vol:.(1234567890)
www.nature.com/scientificreports/

Method Hybrid-cvtol Hybrid-ftol BU Without BU

Hybrid-cvtol – 0.900000 0.103359 0.005121
Hybrid-ftol 0.900000 – 0.062433 0.002444
BU 0.103359 0.062433 – 0.704147
Without BU 0.005121 0.002444 0.704147 -

Table 23.  Summary of the p-value of the posthoc_nemenyi_friedman over all runs-IGD+value comparison.

Method Hybrid-cvtol Hybrid-ftol BU Without BU

Hybrid-cvtol – 0.9000 0.019641 0.010266
Hybrid-ftol 0.9000 – 0.035858 0.019641
BU 0.019641 0.035858 – 0.9000
Without BU 0.010266 0.019641 0.9000 –

Table 24.  Summary of the p-value of the posthoc_nemenyi_friedman over all runs- GD value comparison.

Method Hybrid-cvtol Hybrid-ftol BU Without BU

Hybrid-cvtol – 0.9000 0.047574 0.026690
Hybrid-ftol 0.9000 – 0.014282 0.007293
BU 0.047574 0.014282 – 0.9000
Without BU 0.026690 0.007293 0.9000 –

Table 25.  Summary of the p-value of the posthoc_nemenyi_friedman over all runs- GD+value comparison.

• x2 = l, the length of the welds

• x3 = t, the height of the beam
• x4 = b, the width of the beam

The constraints are:

 
1
g1 = × (t − tmax ) ≤ 0 (33)
tmax
 
1
g2 = × (s − smax ) ≤ 0 (34)
smax
 
1
g3 = × (x1 − x4 ) ≤ 0 (35)
5 − 0.125
 
1
g4 = × (P − Pc ) ≤ 0 (36)
P
where tmax = 13600, smax = 30000, P = 6000, Pc = 64746.022 × (1 − 0.0282346 × x3 ) × x3 × x43 , s = 6 × P × 14/(x4 × x32 ). The
two objective functions are to minimize the fabrication cost of the beam and minimize the deflection of the end
of the beam:

f1 = 1.10471 × x12 × x2 + 0.04811 × x3 × x4 × (14.0 + x2 )) (37)

f2 = 2.1952/(x4 × x33 ) (38)

The four variables of this problem include: (1) thickness of the weld, (2) welded joint’s length, (3) beam’s
width, and (4) beam’s thickness. Using the three strategies, NSGA-II was applied to this problem, and fitness
values were evaluated.
The exploration of sampling and crossover operators in EAs stands as a critical endeavor, influencing the
convergence speed and efficacy of these algorithms, particularly in the context of noisy optimization. In the con-
text of the Welded Beam Design problem, where the impact of noise is significant, the experimentation involved

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 21

Vol.:(0123456789)
 www.nature.com/scientificreports/

Figure 11.  Violin plot of implementation of using different indicators (BNH problem).

diverse operators, including real random sampling, Latin hypercube sampling, and two distinct crossovers:
real-SBX crossover and uniform crossover. Four performance indicators, namely GD, GD+ , IGD, and IGD+,
were meticulously evaluated against these operators to glean insights into their comparative effectiveness. The
compelling findings, detailed in Tables 26, 27, 28, 29, 30, 31 and 32 and illustrated in Fig. 12, underscore the
superiority of the hybrid approach over the two alternative methods.
Examining the Pareto optimal solutions generated by the hybrid methods across different generations reveals
a consistent outperformance, particularly noteworthy when considering constraint violations as tolerance with
real random sampling. This suggests that the hybrid approach, when coupled with appropriate operators, signifi-
cantly improves the optimization process for the Welded Beam Design problem. Notably, the hybrid method’s
effectiveness is evident with Latin hypercube sampling, showcasing its compatibility and superior performance
within this optimization context. Furthermore, the hybrid method incorporating objective function tolerance
using uniform crossover emerges as a robust option, outperforming optimization without the BU approach.
In contrast to initial expectations regarding the BU approach as a constraint handling technique, the results
indicate that it may not be the optimal choice for this specific nonlinear bi-objective model. However, the hybrid
approaches, through strategic integration of various operators, demonstrate a substantial enhancement in per-
formance. This nuanced observation highlights the importance of selecting appropriate operators within hybrid
approaches, suggesting that the effectiveness of the BU method can be context-dependent.
The statistical analyses, including the Friedman rank test and the Posthoc_Nemenyi_Friedman, contribute to
the robustness of the findings. These tests not only affirm the significant differences between the hybrid approach

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 22

Vol:.(1234567890)
www.nature.com/scientificreports/

Sampling
Real_lhs Real_random
Performance indicator BU step Best Mean Median Worst St. Dev BU step Best Mean Median Worst St. Dev
Nonea 0.12 0.15 0.147 0.179 0.021 None 0.12 0.167 0.178 0.19 0.024
BU 0.05 0.15 0.161 0.239 0.051 BU 0.08 0.153 0.139 0.24 0.048
IGD
Hybrid-cvtol 0.01 0.03 0.021 0.089 0.028 Hybrid-cvtol 0.01 0.033 0.034 0.05 0.010
Hybrid-ftol 0.044 0.058 0.058 0.069 0.009 Hybrid-ftol 0.028 0.057 0.061 0.084 0.0190
None 0.05 0.08 0.071 0.119 0.023 None 0.08 0.111 0.116 0.14 0.019
BU 0.03 0.08 0.070 0.138 0.034 BU 0.05 0.090 0.079 0.16 0.034
IGD+
Hybrid-cvtol 0.00 0.00 0.003 0.016 0.004 Hybrid-cvtol 0.00 0.005 0.004 0.00 0.002
Hybrid-ftol 0.015 0.018 0.018 0.020 0.001 Hybrid-ftol 0.012 0.018 0.018 0.27 0.004
None 0.05 0.11 0.113 0.196 0.037 None 0.09 0.166 0.189 0.21 0.048
BU 0.02 0.13 0.128 0.282 0.071 BU 0.08 0.162 0.145 0.27 0.067
GD
Hybrid-cvtol 0.00 0.00 0.002 0.004 0.000 Hybrid-cvtol 0.00 0.002 0.002 0.00 0.001
Hybrid-ftol 0.004 0.008 0.008 0.011 0.002 Hybrid-ftol 0.004 0.008 0.005 0.030 0.008
None 0.05 0.11 0.113 0.196 0.037 None 0.09 0.166 0.189 0.21 0.048
BU 0.02 0.13 0.128 0.282 0.071 BU 0.08 0.162 0.145 0.27 0.067
GD+
Hybrid-cvtol 0.00 0.00 0.001 0.003 0.000 Hybrid-cvtol 0.00 0.001 0.001 0.00 0.000
Hybrid-ftol 0.001 0.007 0.007 0.012 0.004 Hybrid-ftol 0.003 0.008 0.005 0.030 0.009

Table 26.  Statistical results of different methods on the welded beam design problem (sampling). a “None”
means no BU method or hybrid is used and only feasibility rules as an explicit constraint handling method is
worked alone. Significant values are in bold.

Crossover
Uniform real_sbx
Performance indicator BU step Best Mean Median Worst St. Dev BU step Best Mean Median Worst St. Dev
None 0.140 0.166 0.167 0.191 0.017 None 0.080 0.144 0.151 0.197 0.035
BU 0.080 0.126 0.127 0.192 0.033 BU 0.136 0.167 0.167 0.186 0.015
IGD
Hybrid-cvtol 0.006 0.050 0.055 0.095 0.027 Hybrid-cvtol 0.009 0.032 0.028 0.074 0.022
Hybrid-ftol 0.032 0.052 0.051 0.067 0.012 Hybrid-ftol 0.017 0.033 0.033 0.046 0.010
None 0.072 0.092 0.096 0.116 0.015 None 0.036 0.071 0.069 0.108 0.025
BU 0.035 0.071 0.077 0.104 0.023 BU 0.058 0.090 0.088 0.135 0.028
IGD+
Hybrid-cvtol 0.002 0.008 0.009 0.017 0.004 Hybrid-cvtol 0.001 0.005 0.005 0.007 0.002
Hybrid-ftol 0.005 0.008 0.008 0.010 0.001 Hybrid-ftol 0.003 0.005 0.005 0.006 0.001
None 0.086 0.154 0.176 0.21 0.042 None 0.030 0.102 0.086 0.185 0.053
BU 0.042 0.146 0.123 0.311 0.081 BU 0.062 0.194 0.189 0.410 0.099
GD
Hybrid-cvtol 0.001 0.002 0.002 0.005 0.001 Hybrid-cvtol 0.001 0.002 0.002 0.003 0.000
Hybrid-ftol 0.001 0.002 0.002 0.002 0.000 Hybrid-ftol 0.001 0.002 0.002 0.002 0.000
None 0.086 0.154 0.176 0.21 0.042 None 0.030 0.102 0.086 0.185 0.053
BU 0.042 0.146 0.123 0.311 0.081 BU 0.062 0.194 0.189 0.410 0.099
GD+
Hybrid-cvtol 0.000 0.001 0.001 0.004 0.001 Hybrid-cvtol 0.000 0.001 0.001 0.002 0.000
Hybrid-ftol 0.000 0.001 0.001 0.001 0.0004 Hybrid-ftol 0.000 0.001 0.001 0.001 0.000

Table 27.  Statistical results of different methods on the welded beam design problem (crossover). Significant
values are in bold.

and the two other methods but also reinforce the superiority of the proposed hybrid method for solving the
Welded Beam Design problem. The identification of the switching point for the second phase of the optimization
process at evaluation 300 further exemplifies the efficiency of the hybrid approach compared to the without BU
approach, where the feasible region is reached at evaluation 1600 (Fig. 13).
In conclusion, the investigation into sampling and crossover operators, coupled with the nuanced insights
from performance indicators and statistical analyses, positions the hybrid approach as the preferred strategy for
solving the Welded Beam Design problem. The adaptability of the hybrid method, evident in its performance
across various operators, reinforces its efficacy in addressing the challenges posed by the specific characteristics
of this nonlinear bi-objective optimization scenario.

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 23

Vol.:(0123456789)
 www.nature.com/scientificreports/

Sampling Crossover
Real_lhs Real_random Uniform real_sbx
Perofrmance
indicator p-value Statistic value p-value STATISTIC value p-value statistic value p-value Statistic value
IGD 0.16 10.23 0.01 11.09 0.01 9.89 0.01 10.20
IGD+ 0.01 11.15 0.03 8.39 0.01 11.15 0.009 11.36
GD 0.01 11.09 0.01 11.09 0.01 10.78 0.008 11.75
GD+ 0.01 11.09 0.1 11.09 0.01 10.78 0.008 11.75

Table 28.  Summary of the p-value of the Friedman rank test over all runs.

Sampling
Real_lhs Real_random
Hybrid-cvtol Hybrid-cvtol
Hybrid-cvtol vs. Without BU vs. Hybrid-cvtol vs. Without BU vs.
vs. BU BU Without BU vs. BU BU Without BU
Performance indicator p-value ha p-value h p-value h p-value h p-value h p-value h
IGD 0.04 + 0.09 ~ 0.90 ~ 0.06 ~ 0.01 + 0.90 ~
IGD+ 0.04 + 0.02 + 0.90 ~ 0.12 ~ 0.03 + 0.90 ~
GD 0.01 + 0.06 ~ 0.90 ~ 0.06 ~ 0.01 + 0.90 ~
GD+ 0.01 + 0.06 ~ 0.90 ~ 0.06 ~ 0.01 + 0.90 ~

Table 29.  Summary of the p-value of the posthoc_nemenyi_friedman test evaluated over all runs (sampling).
a
Posthoc_nemenyi_friedman: +, ~, and – presents the first method performs statistically significantly better,
equal, and worse than the second approach.

Sampling
Real_lhs Real_random
Hybrid-ftol Hybrid-ftol Hybrid-ftol Hybrid-ftol
Hybrid-ftol vs. Without vs. Hybrid- Hybrid-ftol vs. Without vs. Hybrid-
vs. BU BU cvtol vs. BU BU cvtol
Performance indicator p-value ha p-value h p-value h p-value h p-value h p-value h
IGD 0.04 + 0.09 ~ 0.90 ~ 0.51 ~ 0.22 ~ 0.67 ~
IGD+ 0.43 ~ 0.28 ~ 0.67 ~ 0.90 ~ 0.67 ~ 0.35 ~
GD 0.22 ~ 0.51 ~ 0.67 ~ 0.51 ~ 0.22 ~ 0.67 ~
GD+ 0.22 ~ 0.51 ~ 0.67 ~ 0.51 ~ 0.22 ~ 0.67 ~

Table 30.  Summary of the p-value of the posthoc_nemenyi_friedman test evaluated over all runs (sampling-
Continue). a Posthoc_nemenyi_friedman: +, ~, and – presents the first method performs statistically
significantly better, equal, and worse than the second approach.

Crossover
Uniform Real_sbx
Hybrid-cvtol Hybrid-cvtol
Hybrid-cvtol vs. Without BU vs. Hybrid-cvtol vs. Without BU vs.
vs. BU BU Without BU vs. BU BU Without BU
Performance indicator p-value ha p-value h p-value h p-value h p-value h p-value h
IGD 0.22 ~ 0.06 ~ 0.90 ~ 0.03 + 0.12 ~ 0.90 ~
IGD+ 0.43 ~ 0.04 + 0.67 ~ 0.03 + 0.35 ~ 0.67 ~
GD 0.28 ~ 0.09 ~ 0.90 ~ 0.04 + 0.43 ~ 0.67 ~
GD+ 0.28 ~ 0.09 ~ 0.90 ~ 0.04 + 0.43 ~ 0.67 ~

Table 31.  Summary of p value of the posthoc_nemenyi_friedmantest evaluated over all runs (crossover).
a
Posthoc_nemenyi_friedman: +, ~, and – presents the first method performs statistically significantly better,
equal, and worse than the second approach.

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 24

Vol:.(1234567890)
www.nature.com/scientificreports/

Crossover
Uniform real_sbx
Hybrid-ftol Hybrid-ftol Hybrid-ftol Hybrid-ftol
Hybrid-ftol vs. Without vs. Hybrid- Hybrid-ftol vs. Without vs. Hybrid-
vs. BU BU cvtol vs. BU BU cvtol
Performance indicator p-value ha p-value h p-value h p-value h p-value h p-value h
IGD 0.22 ~ 0.06 ~ 0.90 ~ 0.12 ~ 0.35 ~ 0.90 ~
IGD+ 0.28 ~ 0.02 + 0.90 ~ 0.03 + 0.35 ~ 0.67 ~
GD 0.16 ~ 0.04 + 0.90 ~ 0.02 + 0.28 ~ 0.90 ~
GD+ 0.16 ~ 0.04 + 0.90 ~ 0.02 + 0.28 ~ 0.90 ~

Table 32.  Summary of p value of the posthoc_nemenyi_friedmantest evaluated over all runs (crossover-
Continue). a Posthoc_nemenyi_friedman: +, ~, and – presents the first method performs statistically
significantly better, equal, and worse than the second approach.

Figure 12.  Pareto-optimal fronts obtained for the welded beam design problem for both BU and without BU
compared with the true Pareto-optimal front (Generation 20).

Cantilevered stepped beam design problem

The cantilevered stepped design problem is an example of large-scale size problem. In the cantilevered beam
design problem, the stepped cantilever beam must be able to carry a prescribed end l­oad46. This problem is
originally a single-objective optimization problem considering minimizing the beam volume; however, this
study has extended the original problem and added one more objective function, which minimize end deflection
subject to various engineering design variables (see Appendix). The beam supports the given load, P, at a fixed

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 25

Vol.:(0123456789)
 www.nature.com/scientificreports/

Figure 13.  Constraint violation evaluated by NSGA-II (convergence of the whole population to feasible
solutions).

distance L from the support. Besides, designers of the beam can vary the width (bi ) and height (hi ) of each sec-
tion (Fig. 14). The problem is highly constraint and large-dimensional; however, for this example five segments
(N = 5) including 10 dimensions (10 constraints) are considered. In the pursuit of solving both the single- and
multi-objective versions of this intricate problem, repairing variables x2, x4 , x6, x8 , and x10 have been identified.
The application of the BU method to this constrained nonlinear optimization problem is showcased, provid-
ing a valuable demonstration of the BU method’s utility. Figure 15a–d visually presents the optimal solutions
discovered by NSGA-II employing different methods across various generations.
To assess the performance of the proposed approaches, the hypervolume (HV) indicator is utilized, offering
a comprehensive measure of the quality of solutions (Tables 33, 34). The results, detailed in Tables 35, 36, 37, 38
and 39 and visualized in Fig. 16, unequivocally demonstrate the efficacy of the hybrid approaches in enhancing
solution quality. This improvement is particularly noteworthy in the case of Latin hypercube sampling and simu-
lated binary crossover (SBX), where the hybrid approaches exhibit a significant enhancement in solution quality.
This example provides valuable insights into the application of the BU method in addressing complex, large-
scale, constrained optimization problems. The extension of the original single-objective problem to a multi-
objective setting showcases the flexibility of the proposed methodology. The utilization of hypervolume as a
performance indicator reaffirms the substantial improvements achieved by the hybrid approaches. Overall, this
study contributes not only to the optimization of the Cantilevered Stepped Design problem but also sheds light
on the broader potential of the BU method in tackling diverse engineering optimization challenges.

Multi‑objective car‑side impact problem formulation

In scenarios where the black-box constraints of an optimization problem exhibit highly complex behavior, the
direct evaluation of these constraints can be prohibitively time-consuming. In such instances, the use of surro-
gates becomes a valuable strategy to mitigate the complexity and expedite the optimization process. Surrogates,

Figure 14.  Schematic of the stepped beam design problem.

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 26

Vol:.(1234567890)
www.nature.com/scientificreports/

Figure 15.  Pareto solutions found by NSGA-II by different approaches and generations.

Sampling
Real_random Real_lhs
Algorithm BU step Best Mean Median Worst St. Dev BU step Best Mean Median Worst St. Dev
Nonea 0.6960 0.6344 0.6131 0.6118 0.0325 None 0.7324 0.6585 0.6396 0.6118 0.0478
BU 0.7415 0.7309 0.7355 0.7144 0.0102 BU 0.7428 0.7313 0.7329 0.7156 0.0103
NSGA-II
Hybrid-cvtol 0.7734 0.7731 0.7730 0.7729 0.0002 Hybrid-cvtol 0.7737 0.7731 0.7731 0.7727 0.0003
Hybrid-ftol 0.7734 0.7730 0.7729 0.7728 0.0002 Hybrid-ftol 0.7737 0.7729 0.7727 0.7723 0.0004

Table 33.  Results of different methods on the cantilevered stepped design problem (HV values-sampling).
a
“None” means no BU method or hybrid is used and only feasibility rules as an explicit constraint handling
method is worked alone. Significant values are in bold.

effectively approximating the actual constraints, enable the optimization algorithm to learn from previous evalu-
ations and make informed decisions. The surrogate model, if reducible to a single variable, provides an opportu-
nity to apply the BU approach to solve the constrained surrogate model efficiently, as detailed in the Appendix.
The car side impact problem stands out as a prime example of a complex and time-consuming optimization
challenge, focusing on the multi-objective optimization of a vehicle’s side impact crashworthiness 50. Repairing
variables X2 andX3 are identified for this problem. Various evolutionary algorithms, including NSGA-II, NSGA-
III, UNSGA-III, and AGEMOEA, are implemented to explore the solution space. Tables 40, 41, 42, 43, 44, 45
and 46 present the results of the non-dominated solutions obtained by these methods.
The results underscore the superiority of the hybrid methods, implemented with the proposed evolutionary
algorithms, in terms of diversity. The hybrid approaches exhibit a well-distributed population across the entire
search space, leading to enhanced performance. While the application of the BU approach alone may face

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 27

Vol.:(0123456789)
 www.nature.com/scientificreports/

Crossover
Real_sbx Uniform
Algorithm BU step Best Mean Median Worst St. Dev BU step Best Mean Median Worst St. Dev
Nonea 0.6544 0.6243 0.6396 0.5762 0.0277 None 0.6950 0.6849 0.6941 0.6539 0.0157
BU 0.7428 0.7315 0.7347 0.7156 0.0089 BU 0.7428 0.7285 0.7347 0.7106 0.0129
NSGA-II
Hybrid-cvtol 0.7737 0.7733 0.7733 0.7727 0.0003 Hybrid-cvtol 0.7733 0.7728 0.7730 0.7721 0.0004
Hybrid-ftol 0.7737 0.7732 0.7733 0.7729 0.0002 Hybrid-ftol 0.7734 0.7732 0.7732 0.7730 0.0001

Table 34.  Results of different methods on the the cantilevered stepped design problem (HV values-crossover).
a
“None” means no BU method or hybrid is used and only feasibility rules as an explicit constraint handling
method is worked alone. Significant values are in bold.

Sampling Crossover
Real_random Real_lhs real_sbx Uniform
Algorithm p-value Statistic value p-value Statistic value p-value Statistic value p-value Statistic value
NSGA-II 0.002 14.61 0.003 13.77 0.002 14.03 0.002 14.03

Table 35.  Summary of the p-value of the Friedman rank test over all runs.

Sampling
Real_lhs Real_random
Hybrid-cvtol Hybrid-cvtol
Hybrid-cvtol vs. Without BU vs. Hybrid-cvtol vs. Without BU vs.
vs. BU BU Without BU vs. BU BU Without BU
Algorithm p-value ha p-value h p-value h p-value h p-value h p-value h
NSGA-II 0.16 ~ 0.008 + 0.67 ~ 0.16 ~ 0.008 + 0.67 ~

Table 36.  Summary of the p-value of the posthoc_nemenyi_friedman test evaluated over all runs (sampling).
a
Posthoc_nemenyi_friedman: +, ~, and – presents the first method performs statistically significantly better,
equal, and worse than the second approach.

Sampling
Real_lhs Real_random
Hybrid-ftol Hybrid-ftol Hybrid-ftol Hybrid-ftol
Hybrid-ftol vs. Without vs. Hybrid- Hybrid-ftol vs. Without vs. Hybrid-
vs. BU BU cvtol vs. BU BU cvtol
Algorithm p-value ha p-value h p-value h p-value h p-value h p-value h
NSGA-II 0.59 ~ 0.09 ~ 0.82 ~ 0.59 ~ 0.09 ~ 0.82 ~

Table 37.  Summary of the p-value of the posthoc_nemenyi_friedman test evaluated over all runs (sampling-
Continue). a Posthoc_nemenyi_friedman: +, ~, and – presents the first method performs statistically
significantly better, equal, and worse than the second approach.

Crossover
Uniform real_sbx
Hybrid-cvtol Hybrid-cvtol
Hybrid-cvtol vs. Without BU vs. Without Hybrid-cvtol vs. Without BU vs.
vs. BU BU BU vs. BU BU Without BU
Algorithm p-value ha p-value h p-value h p-value h p-value h p-value h
NSGA-II 0.67 ~ 0.12 ~ 0.67 0.35 ~ 0.03 + 0.67 ~

Table 38.  Summary of p value of the posthoc_nemenyi_friedmantest evaluated over all runs (crossover).
a
Posthoc_nemenyi_friedman: +, ~, and – presents the first method performs statistically significantly better,
equal, and worse than the second approach.

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 28

Vol:.(1234567890)
www.nature.com/scientificreports/

Crossover
Uniform real_sbx
Hybrid-ftol Hybrid-ftol Hybrid-ftol Hybrid-ftol
Hybrid-ftol vs. Without vs. Hybrid- Hybrid-ftol vs. Without vs. Hybrid-
vs. BU BU cvtol vs. BU BU cvtol
Algorithm p-value ha p-value h p-value h p-value h p-value h p-value h
NSGA-II 0.12 ~ 0.005 + 0.67 ~ 0.35 ~ 0.03 + 0.90 ~

Table 39.  Summary of p value of the posthoc_nemenyi_friedmantest evaluated over all runs (crossover-
Continue). a Posthoc_nemenyi_friedman: +, ~, and – presents the first method performs statistically
significantly better, equal, and worse than the second approach.

Figure 16.  Violin plot of implementation of using different methods (Stepped beam design problem problem).

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 29

Vol.:(0123456789)
 www.nature.com/scientificreports/

efficiency challenges, particularly in the context of a three-objective optimization problem, the hybrid approach
proves adept at overcoming these difficulties. Statistical analyses of the results reveal that, in some cases and with
different algorithms, the BU method may not be as efficient, yet the hybrid approach consistently outperforms
it, signifying the added benefits of the hybridization strategy.
Comparative statistical tests emphasize the significant advantage of the hybrid approaches over the BU
method. Notably, although some statistical tests indicate no significant difference between the hybrid and BU
methods, the hybrid method consistently achieves higher hypervolume (HV) values. This suggests that, even
when statistical significance is not apparent, the hybrid approach provides a qualitative improvement in solu-
tion quality.
In conclusion, the exploration of the car side impact problem demonstrates the efficacy of hybrid approaches
in addressing complex, multi-objective optimization challenges with intricate black-box constraints. The incor-
poration of surrogates and the application of the BU method within a hybrid framework contribute to improved
diversity and solution quality. These findings not only advance the optimization of the car side impact problem
but also offer valuable insights into the broader applicability of hybrid evolutionary algorithms in addressing
real-world, complex optimization scenarios.

Discussion
The BU method is an implicit constraint-handling technique coupled with an explicit constraint-handling tech-
nique that aims to cut the infeasible search space over iterations to find the feasible region faster. Since optimi-
zation is an iterative process whose boundaries can be changed during the process, the i-th decision variable
boundaries in each iteration exhibit a dynamic nature. By applying this philosophy to boost the optimization
process in the BU method, the bounds of selected variable (s) are repaired (repairing variables) to satisfy the
constraints. As a result, new solutions are created within the updated boundary of the latest search space. The
above-mentioned process helps the algorithm search for a solution within the feasible region instead of the whole
area. Although the BU method directs the search operator to the feasible space and reaches the feasible region
in a reasonable time, it twists the search space, making the optimization problem more challenging. As such,
the results are not always accurate, and the algorithm may result in premature convergence. Therefore, in this
paper, two “hybrid” switching mechanisms are proposed. In the proposed approach, the BU method is applied
to the problem after initializing the algorithm, and in this phase, a threshold tolerance is set. The BU method

Sampling
real_random Real_lhs
Algorithm BU step Best Mean Median Worst St. Dev BU step Best Mean Median Worst St. Dev
Nonea 0.249 0.244 0.2466 0.2398 0.0038 None 0.260 0.217 0.213 0.180 0.0283
BU 0.770 0.756 0.76 0.733 0.0132 BU 0.767 0.743 0.750 0.700 0.0228
NSGA-II
Hybrid-cvtol 0.818 0.800 0.7985 0.790 0.0094 Hybrid-cvtol 0.814 0.795 0.799 0.768 0.0151
Hybrid-ftol 0.803 0.803 0.803 0.802 0.0004 Hybrid-ftol 0.806 0.797 0.797 0.785 0.0075
None 0.259 0.165 0.1333 0.1225 0.0520 None 0.263 0.186 0.154 0.136 0.0505
BU 0.818 0.803 0.805 0.778 0.0141 BU 0.824 0.809 0.813 0.785 0.0132
NSGA-III
Hybrid-cvtol 0.850 0.826 0.826 0.810 0.0141 Hybrid-cvtol 0.853 0.829 0.824 0.810 0.0162
Hybrid-ftol 0.858 0.852 0.852 0.845 0.004 Hybrid-ftol 0.841 0.833 0.837 0.819 0.008
None 0.214 0.191 0.1996 0.1575 0.0195 None 0.182 0.141 0.149 0.074 0.0361
BU 0.815 0.802 0.80 0.788 0.0090 BU 0.833 0.805 0.803 0.790 0.0146
UNSGA-III
Hybrid-cvtol 0.823 0.811 0.816 0.791 0.0110 Hybrid-cvtol 0.830 0.819 0.827 0.796 0.0133
Hybrid-ftol 0.816 0.810 0.810 0.803 0.004 Hybrid-ftol 0.829 0.822 0.826 0.811 0.006
None 0.272 0.226 0.222 0.1937 0.0265 None 0.276 0.224 0.251 0.163 0.0481
BU 0.849 0.801 0.794 0.771 0.0257 BU 0.804 0.777 0.773 0.763 0.0140
AGEMOEA
Hybrid-cvtol 0.849 0.828 0.819 0.812 0.0145 Hybrid-cvtol 0.853 0.833 0.833 0.816 0.0147
Hybrid-ftol 0.835 0.827 0.826 0.816 0.006 Hybrid-ftol 0.841 0.834 0.838 0.823 0.006
None 0.825 0.803 0.811 0.770 0.018 None 0.830 0.773 0.782 0.707 0.045
BU 0.812 0.780 0.771 0.747 0.026 BU 0.777 0.757 0.767 0.734 0.018
NSDE
Hybrid-cvtol 0.818 0.810 0.816 0.794 0.008 Hybrid-cvtol 0.827 0.813 0.811 0.803 0.008
Hybrid-ftol 0.808 0.798 0.793 0.791 0.007 Hybrid-ftol 0.823 0.809 0.809 0.791 0.011
None 0.782 0.737 0.737 0.704 0.029 None 0.816 0.805 0.812 0.774 0.015
BU 0.790 0.759 0.767 0.734 0.021 BU 0.794 0.782 0.781 0.776 0.006
NSDER
Hybrid-cvtol 0.829 0.816 0.817 0.809 0.006 Hybrid-cvtol 0.840 0.809 0.805 0.791 0.016
Hybrid-ftol 0.823 0.806 0.809 0.791 0.104 Hybrid-ftol 0.808 0.798 0.793 0.791 0.007

Table 40.  Results of different methods on the multi-objective car-side impact problem (HV values-sampling).
“None” means no BU method or hybrid is used and only feasibility rules as an explicit constraint handling
method is worked alone. Significant values are in bold.

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 30

Vol:.(1234567890)
www.nature.com/scientificreports/

Crossover
real_sbx Uniform
Algorithm BU step Best Mean Median Worst St. Dev BU step Best Mean Median Worst St. Dev
None 0.237 0.189 0.180 0.170 0.024 None 0.309 0.264 0.271 0.166 0.051
BU 0.765 0.741 0.750 0.697 0.025 BU 0.773 0.756 0.761 0.732 0.013
NSGA-II
Hybrid-cvtol 0.810 0.798 0.798 0.776 0.012 Hybrid-cvtol 0.806 0.798 0.799 0.786 0.007
Hybrid-ftol 0.808 0.801 0.805 0.790 0.006 Hybrid-ftol 0.801 0.799 0.800 0.793 0.003
None 0.254 0.187 0.186 0.120 0.054 None 0.237 0.2064 0.212 0.164 0.025
BU 0.815 0.806 0.811 0.793 0.008 BU 0.816 0.785 0.772 0.766 0.021
NSGA-III
Hybrid-cvtol 0.853 0.824 0.829 0.781 0.023 Hybrid-cvtol 0.844 0.828 0.828 0.815 0.008
Hybrid-ftol 0.833 0.826 0.830 0.814 0.007 Hybrid-ftol 0.833 0.829 0.831 0.822 0.004
None 0.188 0.181 0.182 0.174 0.004 None 0.255 0.226 0.218 0.198 0.020
BU 0.797 0.783 0.788 0.767 0.011 BU 0.828 0.801 0.802 0.775 0.016
UNSGA-III
Hybrid-cvtol 0.851 0.834 0.829 0.816 0.013 Hybrid-cvtol 0.834 0.824 0.828 0.808 0.008
Hybrid-ftol 0.842 0.835 0.839 0.824 0.006 Hybrid-ftol 0.830 0.825 0.825 0.819 0.003
None 0.267 0.225 0.217 0.193 0.030 None 0.278 0.252 0.270 0.192 0.032
BU 0.810 0.781 0.791 0.729 0.027 BU 0.799 0.788 0.791 0.777 0.008
AGEMOEA
Hybrid-cvtol 0.856 0.839 0.843 0.826 0.011 Hybrid-cvtol 0.853 0.839 0.851 0.817 0.016
Hybrid-ftol 0.845 0.838 0.838 0.830 0.005 Hybrid-ftol 0.848 0.839 0.838 0.827 0.007
None 0.814 0.804 0.804 0.791 0.008 None 0.810 0.805 0.805 0.799 0.003
BU 0.796 0.782 0.781 0.763 0.012 BU 0.789 0.783 0.782 0.774 0.005
NSDE
Hybrid-cvtol 0.815 0.810 0.810 0.804 0.003 Hybrid-cvtol 0.812 0.810 0.811 0.808 0.001
Hybrid-ftol 0.802 0.798 0.798 0.793 0.003 Hybrid-ftol 0.799 0.798 0.798 0.796 0.001
None 0.755 0.739 0.738 0.718 0.013 None 0.747 0.740 0.739 0.730 0.0060
BU 0.772 0.760 0.760 0.745 0.009 BU 0.765 0.760 0.760 0.754 0.0037
NSDER
Hybrid-cvtol 0.819 0.816 0.816 0.812 0.002 Hybrid-cvtol 0.816 0.816 0.816 0.815 0.0005
Hybrid-ftol 0.872 0.813 0.811 0.737 0.048 Hybrid-ftol 0.846 0.823 0.836 0.781 0.0251

Table 41.  Results of different methods on the multi-objective car side design problem (HV values-crossover).

Sampling Crossover
Real_lhs Real_random Uniform real_sbx
Algorithm p-value Statistic value p-value Statistic value p-value Statistic value p-value Statistic value
NSGA-II 0.0067 10.0 0.0067 10.0 0.0067 10.0 0.0067 10.0
NSGA-III 0.0149 8.40 0.0149 8.40 0.0067 10.0 0.0149 8.40
UNSGA-III 0.0223 7.60 0.02237 7.60 0.0067 10.0 0.0067 10.0
AGEMOEA 0.0223 7.60 0.0149 8.40 0.0067 10.0 0.0067 10.0
NSDE 0.032 8.75 0.050 7.73 0.0018 15.0 0.002 14.03
NSDER 0.054 7.62 0.001 15.0 0.003 13.56 0.006 12.11

Table 42.  Summary of the p-value of the Friedman rank test over all runs.

optimizes the problem based on the threshold tolerances so that the population has zero constraint violations or
there is no more change in the objective space. Then, the optimization problem is continued with the BU method.
The results suggest that:

• Advantageously, the hybrid method can be used with and without the BU method. In the first phase of the
hybrid method, the BU method helps the optimization algorithm to avoid searching inside the infeasible
region and boosts it to explore the feasible area more efficiently. This way, the optimization algorithm finds
the feasible area with fewer FEs. However, because the BU brings about changes to the landscape of the
search space, it may result in premature convergence. Hence, in the second phase of the hybrid method, the
optimization algorithm searches the feasible area without using the BU approach.
• GA has the least standard deviation among the other algorithms for the single-objective function optimiza-
tion problem.
• The statistical tests for single-objective optimization problems show there is a significant difference between
hybrid methods and two other methods (with BU and without BU) using GA. Moreover, GA works well for
this method and boosts optimization problems. Due to the fact that GA can benefit from different operator

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 31

Vol.:(0123456789)
 www.nature.com/scientificreports/

Sampling
Real_lhs Real_random
Hybrid-cvtol Hybrid-cvtol
Hybrid-cvtol vs. Without BU vs. Hybrid-cvtol vs. Without BU vs.
vs. BU BU Without BU vs. BU BU Without BU
Algorithm p-value ha p-value h p-value h p-value h p-value h p-value h
NSGA-II 0.31 ~ 0.01 + 0.59 ~ 0.51 ~ 0.06 ~ 0.67 ~
NSGA-III 0.51 ~ 0.06 ~ 0.67 ~ 0.67 ~ 0.12 ~ 0.67 ~
UNSGA-III 0.82 ~ 0.06 ~ 0.35 ~ 0.22 ~ 0.01 + 0.67 ~
AGEMOEA 0.51 ~ 0.06 ~ 0.67 ~ 0.67 ~ 0.04 + 0.43 ~
NSDE 0.03 + 0.35 ~ 0.67 ~ 0.03 + 0.82 ~ 0.22 ~
NSDER 0.04 + 0.74 ~ 0.35 ~ 0.12 ~ 0.005 + 0.67 ~

Table 43.  Summary of the p-value of the posthoc_nemenyi_friedman test evaluated over all runs (sampling).
a
Posthoc_nemenyi_friedman: +, ~, and – presents the first method performs statistically significantly better,
equal, and worse than the second approach.

Sampling
Real_lhs Real_random
Hybrid-ftol Hybrid-cvtol Hybridfvtol Hybrid-cvtol
Hybrid-ftol vs. Without vs Hybrid- Hybrid-ftol vs. Without vs Hybrid-
vs. BU BU ftol vs. BU BU ftol
Algorithm p-value ha p-value h p-value h p-value h p-value h p-value h
NSGA-II 0.20 ~ 0.007 + 0.90 ~ 0.22 ~ 0.01 + 0.90 ~
NSGA-III 0.22 ~ 0.01 + 0.90 ~ 0.12 ~ 0.005 + 0.67 ~
UNSGA-III 0.82 ~ 0.06 ~ 0.90 ~ 0.51 ~ 0.90 ~ 0.06 ~
AGEMOEA 0.22 ~ 0.01 + 0.90 ~ 0.74 ~ 0.06 ~ 0.90 ~
NSDE 0.35 ~ 0.90 ~ 0.67 ~ 0.03 + 0.82 ~ 0.22 ~
NSDER 0.59 ~ 0.90 ~ 0.51 ~ 0.67 ~ 0.12 ~ 0.67 ~

Table 44.  Summary of the p-value of the posthoc_nemenyi_friedman test evaluated over all runs (sampling).
a
Posthoc_nemenyi_friedman: +, ~, and – presents the first method performs statistically significantly better,
equal, and worse than the second approach.

Crossover
Uniform real_sbx
Hybrid-cvtol
Hybrid-cvtol vs. Without BU vs. Hybrid-cvtol Hybrid-cvtol vs. BU vs.
vs. BU BU Without BU vs. BU Without BU Without BU
Algorithm p-value ha p-value h p-value h p-value h p-value h p-value h
NSGA-II 0.45 ~ 0.03 + 0.59 ~ 0.45 ~ 0.03 + 0.59 ~
NSGA-III 0.51 ~ 0.06 ~ 0.67 ~ 0.82 ~ 0.12 ~ 0.51 ~
UNSGA-III 0.67 ~ 0.12 ~ 0.67 ~ 0.82 ~ 0.06 ~ 0.35 ~
AGEMOEA 0.28 ~ 0.02 + 0.67 ~ 0.22 ~ 0.01 + 0.67 ~
NSDE 0.005 + 0.67 ~ 0.12 ~ 0.005 0.51 ~ 0.22 ~
NSDER 0.51 ~ 0.06 ~ 0.67 ~ 0.35 ~ 0.013 0.51 ~

Table 45.  Summary of the p-value of the posthoc_nemenyi_friedman test evaluated over all runs (Crossover).
a
Posthoc_nemenyi_friedman: +, ~, and – presents the first method performs statistically significantly better,
equal, and worse than the second approach.

settings, such as crossover and sampling, better results by choosing the appropriate operator setting may be
achieved.
• Regarding the multi-objective optimization problems, the hybrid method could find more non-dominated
solutions with better distribution compared to with and without the BU methods. The BNH test problems
indicated that the hybrid method produced promising results for IGD and HV values compared to with BU
and without BU approaches.
• Although the BU and hybrid methods outperformed the other algorithms in single- and multi-objective
optimization, they are still case-dependent, and there are EAs that are not suitable in some cases.

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 32

Vol:.(1234567890)
www.nature.com/scientificreports/

Crossover
Uniform real_sbx
Hybrid-ftol Hybrid-cvtol Hybridfvtol Hybrid-cvtol
Hybrid-ftol vs. Without vs Hybrid- Hybrid-ftol vs. Without vs Hybrid-
vs. BU BU ftol vs. BU BU ftol
Algorithm p-value ha p-value h p-value h p-value h p-value h p-value h
NSGA-II 0.12 ~ 0.003 + 0.87 ~ 0.12 ~ 0.003 + 0.87 ~
NSGA-III 0.22 ~ 0.01 + 0.90 ~ 0.35 ~ 0.01 + 0.82 ~
UNSGA-III 0.12 ~ 0.005 + 0.67 ~ 0.82 ~ 0.06 ~ 0.90 ~
AGEMOEA 0.43 ~ 0.04 + 0.90 ~ 0.51 ~ 0.06 ~ 0.90 ~
NSDE 0.67 ~ 0.67 ~ 0.12 ~ 0.51 ~ 0.90 ~ 0.22 ~
NSDER 0.22 ~ 0.013 + 0.90 ~ 0.82 ~ 0.12 ~ 0.82 ~

Table 46.  Summary of the p-value of the posthoc_nemenyi_friedman test evaluated over all runs (Crossover).
a
Posthoc_nemenyi_friedman: +, ~, and – presents the first method performs statistically significantly better,
equal, and worse than the second approach.

• By boosting the optimization process, the hybrid method finds the first and whole populations much faster
than the classic BU and without BU methods for most problems (i.e., the single- or multi-objective optimi-
zation problems of this paper). Moreover, regarding the car side impact design problem, the hybrid method
obtained good diversity, whereas without the BU method searched for solutions in a specific feasible space.
• For some case studies, although the hybrid method resulted in the same (or close-to-final) solution as the
two other methods, it requires fewer FEs to converge to a solution.
• Compared to unconstrained optimization problems, constrained optimization problems are more challenging
since many infeasible regions appear in the search space (meaning the hit ratio is low). This makes solving a
constrained problem more challenging, especially highly constrained problems that can lead to difficulties
such as convergence-, diversity-, and feasibility-hardness. Therefore, finding just one feasible solution is a
significant achievement. The primary goal of the BU method is to minimize the search space by narrowing
down the variable range and eliminating infeasible options. To compare the effectiveness of the proposed
approach, the first feasible solution found by different methods is compared to one another, and the whole
population is tracked against generations. The hybrid method uses this philosophy to switch the optimiza-
tion process and avoid reaching premature convergence. On the other hand, once the whole feasible area is
found by the BU method, the final solution (global) could be achieved with fewer FEs in a reasonable time.
• Regarding the multi-objective car side impact design problem, the hybrid approach works significantly bet-
ter than without the BU method when incorporated in the employed multi-objective algorithms. It is worth
mentioning that although there is no significant difference in the final HV values between the hybrid and
BU methods, as mentioned earlier, the hybrid method could achieve the final solution in fewer FEs.
• In scenarios where explicit formulation of constraints is challenging, the use of surrogate models, also known
as metamodels, proves to be a valuable solution. Surrogate models have demonstrated effectiveness in situ-
ations where deriving explicit formulations for constraints is impractical. These models serve as approxi-
mations of intricate simulations, providing an efficient means to evaluate constraints during optimization
without the computational overhead of running resource-intensive simulations. For instance, consider opti-
mization problems associated with finite element models, such as those involving nodal displacements and
bar stresses in truss optimization. In these cases, surrogate models can be strategically trained to emulate the
intricate behavior observed in simulation results. This approach becomes particularly advantageous when
the direct formulation of these constraints proves to be impractical due to their complexity or reliance on
computationally expensive procedures.

In EAs, sampling and crossover are critical factors that can strongly affect convergence speed and are often
used in noisy optimization problems where reducing the negative impact of noise is crucial. Moreover, sampling
is a popular strategy for dealing with noise. Therefore, experiments with different sampling operators, includ-
ing real random sampling, Latin hypercube sampling, and two different real-SBX and uniform crossovers, were
conducted to reduce the noise impact.

Limitations
Although this study provides valuable insights into optimization problems, there are certain limitations to this
work. One significant limitation of this work lies in the relatively small size of variables in the problems. It should
be noted that this study aims to contribute by addressing a specific aspect of the optimization problem. In this
study, the benchmark libraries have not been used and some specific problems with solvable constraints with
respect to at least one variable have been used. While the limited number of variables in our optimization prob-
lems may restrict the immediate applicability of our findings to large-scale size problems, it is clear the insights
gained from this research serve as a valuable foundation for further research in the field.
Another limitation of this study is related to the experimental section, where the algorithms compared with
the proposed algorithm are not the most recent. It is important to note that the aim of this work is not focused on
proposing and developing state-of-the-art algorithms; rather, the advancement of constraint handling techniques

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 33

Vol.:(0123456789)
 www.nature.com/scientificreports/

and the development of the BU approach are being concentrated on. Consequently, well-known algorithms
such as NSGA-II in the analysis have been included. However, as a future study, it would be valuable to consider
incorporating state-of-the-art algorithms to provide a more comprehensive evaluation of our proposed approach
in comparison to the latest advancements in the field.
Moreover, while the current selection of repairing variables is designed to be applicable even in scenarios
with multiple disjoint feasible regions, the evaluation of constraints can pose challenges in situations where the
behavior is non-trivial. In such cases, alternative approaches to repairing variable selection, involving book-
keeping or linking of these disjoint feasible ranges, may be necessary. This intriguing aspect has been highlighted
and it is proposed as a subject for future investigation, intending to delve deeper into the nuanced evaluation of
constraints in conditions assessment.

Conclusion
This study presents an extension of the work proposed b ­ y6 for solving constrained optimization problems effi-
ciently. The proposed methods use an implicit constraint handling approach called boundary updating (BU),
which focuses on directly handling constraints and is coupled with other explicit constraint-handling techniques
(here, feasibility rules).
In this study, two augmented versions of the BU method were applied to single-objective, multi-objective,
and many-objective constrained optimization problems, including benchmarks and real-world problems. To
accomplish this goal, two thresholds are considered, each representing different switching method. In the initial
method, the optimization process shifts to a state that does not involve the BU approach once constraint viola-
tions reach zero. In the second approach, the optimization process enters a phase without BU methods when
there is no longer any observed change in the objective space. For single-objective constrained optimization
problems, popular algorithms like GA, DE, ES, SRES, and BRKGA were employed, and NSGA-II, a renowned
algorithm, was used for bi-objective problems. For the multi-objective constrained optimization problems,
NSGA-III, UNSGA-III, and AGEMOEA were applied. The results show that the hybrid approach performs
significantly better than with- and without the BU methods, and can find better solutions with fewer iterations.
As a future study, it is suggested that the proposed BU method be coupled with other constraint-handling
techniques, for example different types of penalty methods with different objective functions. Also, additional
test problems in the literature, such as many-objective and dynamic functions, could be considered for further
validating the proposed hybrid approach. After BU generalization, it can also be applied to benchmark libraries
such as CEC, NEVERGRAD, BBOB, GNBG, and DIRECTGOLib. Although BU is not defined for any specific
optimization algorithms, implementing BU to a wider range of optimization algorithms can be investigated in
another future study. Moreover, as a prospective avenue for future research, exploring the applicability of the
proposed approach to constrained problems featuring multiple disconnected regions is a promising direction.
Many constrained optimization scenarios exhibit this characteristic, with only certain regions potentially close
to the true optimum. Investigating the effectiveness and adaptability of our approach in addressing such com-
plex problem structures holds significant potential for enhancing the robustness and versatility of the proposed
methodology. This important consideration will be a key focus in our forthcoming studies.

Data availability
All data used in this study are simulated and not publicly available. The datasets were generated solely for the
purpose of conducting the research outlined in this manuscript. Simulated data files are not deposited in any
public repository or third-party platform. However, upon reasonable request, and subject to any applicable
restrictions, the corresponding author can provide the simulated data used in this study. Also, our code has been
executed using the open-source Python library, Pymoo (https://​pymoo.​org/​index.​html) and Pymoode (https://​
pymoo​de.​readt​hedocs.​io/​en/​latest/).

Appendix
G01 formulation:

f1 (x) = 5 × (x1 + x2 + x3 + x4 ) − 5 × x12 + x22 + x32 + x42

 

−(x5 + x6 + x7 + x8 + x9 + x10 + x11 + x12 + x13 ) (39)

S.t.
g1 (x) = 2 × x1 + 2 × x2 + x10 + x11 ≤ 10 (40)

g2 (x) = 2 × x1 + 2 × x3 + x10 + x12 ≤ 10 (41)

g3 (x) = 2 × x2 + 2 × x3 + x11 + x12 ≤ 10 (42)

g4 (x) = −8 × x1 + x10 ≤ 0 (43)

g5 (x) = −8 × x2 + x11 ≤ 0 (44)

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 34

Vol:.(1234567890)
www.nature.com/scientificreports/

g6 (x) = −8 × x3 + x12 ≤ 0 (45)

g7 (x) = −2 × x4 − x5 + x10 ≤ 0 (46)

g8 (x) = −2 × x6 − x7 + x11 ≤ 0 (47)

g9 (x) = −2 × x8 − x9 + x12 ≤ 0 (48)

0 ≤ x1 · · · x9 ≤ 1, 0 ≤ x10 , x11 , x12 ≤ 100, 0 ≤ x13 ≤ 1 (49)

Car-side problem formulation:
f1 (x) = 1.98 + 4.90 × x1 + 6.67 × x2 + 6.98 × x3 + 4.01 × x4 + 1.78 × x5 + 2.73 × x7 (50)

g1 (x) = 1.16 − 0.3717 × x2 × x4 − 0.00931×x 2 × x10 − 0.484 × x3 × x9 + 0.01343 × x6 × x10 (51)

g2 (x) = 46.36 − 9.9 × x2 − 12.9 × x1 × x8 + 0.1107 × x3 × x10 (52)

g3 (x) = 33.86 + 2.95 × x3 + 0.1792 × x10 − 5.057 × x1 × x2 − 11.0 × x2 × x8 − 0.021 × x5

(53)
× x10 − 9.98 × x7 × x8 + 22.0 × x8 × x9

g4 (x) = 28.98 + 3.818 × x3 − 4.2 × x1 × x2 − 5.0207 × x5 × x10 + 6.63 × x6 × x9 − 7.7 × x7 × x8 + 0.32 × x9 × x10
(54)
g5 (x) = 0.261 − 0.0159 × x1 × x2 − 0.188 × x1 × x8 − 0.019 × x2 × x7 + 0.0144 × x3 × x5 + 0.0008757
× x5 × x10 + 0.08045 × x6 × x9 + 0.00139 × x8 × x11 + 0.00001575 × x10 × x11 (55)
g6 (x) = 0.214 + 0.00817 × x5 − 0.0131 × x1 × x8 − 0.0704 × x1 × x9 + 0.03099 × x2
× x6 − 0.018 × x2 × x7 + 0.0208 × x3 × x8 + 0.121 × x3 × x9 − 0.00364 × x5
× x6 + 0.0007715 × x5 × x10 − 0.0005354 × x6 × x10 + 0.00121 × x8 (56)

× x11 + 0.00184 × x9 × x10 − 0.02 × x22

g7 (x) = 0.74 − 0.61 × x2 − 0.163 × x3 × x8 + 0.001232 × x3 × x10 − 0.166 × x7 × x9 + 0.227 × x22

(57)
2
g8 (x) = 4.72 − 0.5 × x4 − 0.19 × x2 × x3 − 0.0122 × x4 × x10 + 0.009325 × x6 × x10 + 0.000191 × x11
(58)
g9 (x) = 10.58 − 0.674 × x1 × x2 − 1.95 × x2 × x8 + 0.02054 × x3 × x10 − 0.0198 × x4 × x10 + 0.028 × x6 × x10
(59)
2
g10 = 16.45 − 0.489 × x3 × x7 − 0.843 × x5 × x6 + 0.0432 × x9 × x10 − 0.0556 × x9 × x11 − 0.000786 × x11
(60)
0.5 ≤ x1 , x3 , x4 ≤ 1.5, 0.45 ≤ x2 ≤ 1.35, 0.875 ≤ x5 ≤ 2.625, 0.4 ≤ x6 , x7 ≤ 1.2, 0.19 ≤ x8 , x9 ≤ 0.35, −30 ≤ x10 , x11 ≤ 30
(61)
Speed reducer formulation:

minz = 0.785 × x1 × x22 × (3.33 × x32 + 14.9334 × x3 − 43.0934) − 1.508 × x1

(62)
× (x62 + x72 ) + 7.4777 × (x63 + x73 ) + 0.7854 × (x4 × x62 + x5 × x72 )

S.t. (63)

27/(x1 × x22 × x3 ) ≤ 1 (64)

397.5/x1 × x22 × x32 ≤ 1 (65)

1.93 × x43 /x2 × x3 ×x 46 ≤ 1 (66)

1.93 × x53 /(x 2 × x3 × x74 ) ≤ 1 (67)

 2
x4
745 × + 1.69 × 107 /(110 × x63 ) ≤ 1 (68)
x2 × x3

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 35

Vol.:(0123456789)
 www.nature.com/scientificreports/

 2
x5
745 × + 1.575 × 108 /(85 × x73 ) ≤ 1 (69)
x2 × x3

(x2 × x3 )/40 ≤ 1 (70)

(5 × x2 )/x1 ≤ 1 (71)

x1 /(12 × x2 ) ≤ 1 (72)

(1.5 × x6 + 1.9)/x4 ≤ 1 (73)

(1.1 × x7 + 1.9)/x5 ≤ 1 (74)

2.6 ≤ x1 ≤ 3.6, 0.7 ≤ x2 ≤ 0.8, 17 ≤ x3 ≤ 28,

7.3 ≤ x4 ≤ 8.3, 7.8 ≤ x5 ≤ 8.3, 2.9 ≤ x6 ≤ 3.9,

5.0 ≤ x7 ≤ 5.5 (75)

OSY formulation:

minf1 (x) = −[25 × (x1 − 2)2 + (x2 − 2)2 + (x3 − 1)2 + (x4 − 4)2 + (x5 − 1)2 ] (76)

minf2 (x) = x12 + x22 + x32 + x42 + x52 + x62

S.t. (77)

C1 (x) = x1 + x2 − 2 ≥ 0 (78)

C2 (x) = 6 − x 1 − x2 ≥ 0 (79)

C3 (x) = 2 − x2 + x 1 ≥ 0 (80)

C4 (x) = 2 − x1 + 3 × x 2 ≥ 0 (81)

C5 (x) = 4 − (x 3 − 3)2 − x 4 ≥ 0 (82)

C6 (x) = (x 5 − 3)2 + x 6 − 4 ≥ 0 (83)

0 ≤ x1 , x2 , x6 ≤ 10 · · · 1 ≤ x3 , x5 ≤ 5 · · · 1 ≤ x4 ≤ 6 (84)
BNH formulation:

minf1 (x) = 4 × x12 + 4 × x22 (85)

minf2 (x) = (x1 − 5)2 + (x2 − 5)2 (86)

S.t. (87)

C1 (x) = (x1 − 5)2 + x22 ≤ 25 (88)

C2 (x) = (x1 − 8)2 + (x2 + 3)2 ≥ 7.7 (89)

0 ≤ x1 ≤ 5 · · · 0 ≤ x2 ≤ 3 (90)
Cantilevered beam design problem:
Minf1 (x) = 100 × (x5 × x6 + x7 × x8 + x9 × x10 + x1 × x2 + x3 × x4 ) (91)

244 148 76 28 4
Minf2 (x) = 10000 × ( 3
+ 3
+ 3
+ 3
+ ) (92)
x5 × x6 x7 × x8 x9 × x10 x1 × x2 x3 × x34

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 36

Vol:.(1234567890)
www.nature.com/scientificreports/

Subject to
x5 × x26
g1 (x) = 10.7143 − ≤0 (93)
1000

x7 × x28
g2 (x) = 8.5714 − ≤0 (94)
1000

x9 × x210
g3 (x) = 6.4286 − ≤0 (95)
1000

x1 × x22
g4 (x) = 4.2857 − ≤0 (96)
1000

x3 × x24
g5 (x) = 2.1429 − ≤0 (97)
1000

g6 (x) = x6 − 20 × x5 ≤ 0 (98)

g7 (x) = x8 − 20 × x7 ≤ 0 (99)

g8 (x) = x10 − 20 × x9 ≤ 0 (100)

g9 (x) = x2 − 20 × x1 ≤ 0 (101)

g10 (x) = x4 − 20 × x3 ≤ 0 (102)

xl = [1, 30, 2.4, 45, 2.4, 45, 1, 30, 1, 30] (103)

xu = [5, 65, 3.1, 60, 3.1, 60, 5, 65, 5, 65] (104)

Multi-objective car side formulation:
f1 (x) = 1.98 + 4.9 × x1 + 6.67 × x2 + 6.98 × x3 + 4.01 × x4 + 1.78 × x5 + 0.00001 × x6 + 2.73×x 7
(105)
f2 (x) = 4.72 − 0.5 × x4 − 0.19 × x2 × x3 (106)

f3 (x) = 0.5 × (10.58 − 0.674 × x1 × x2 − 0.67275 × x2 + 16.45 − 0.489 × x3 × x7 − 0.843 × x5 × x6

(107)
g1 (x) = 1.16 − 0.3717 × x2 × x4 − 0.0092928 × x3 ≤ 1 (108)

g2 (x) = 0.261 − 0.0159 × x1 × x2 − 0.06486 × x1 − 0.019 × x2 × x7 + 0.0144 × x3 × x5 + 0.0154464 × x6 ≤ 0.32

(109)
g3 (x) = 0.214 + 0.00817 × x5 − 0.045195 × x1 − 0.0135168 × x1 + 0.03099 × x2 × x6 − 0.018
(110)
× x2 × x7 + 0.007176 × x3 + 0.023232 × x3 − 0.00364 × x5 × x6 − 0.018 × x22 ≤ 0.32

g4 (x) = 0.74 − 0.61 × x2 − 0.31296 × x3 − 0.031872 × x7 + 0.227 × x22 ≤ 0.32 (111)

g5 (x) = 28.98 + 3.818 × x3 − 4.2 × x1 × x2 + 1.2729 × x6 − 2.68065 × x7 ≤ 32 (112)

g6 (x) = 33.86 + 2.95 × x3 − 5.057 × x1 × x2 − 3.795 × x2 − 3.4431 × x7 + 1.45728 ≤ 32 (113)

g7 (x) = 46.36 − 9.9 × x2 − 4.4505 × x1 ≤ 32 (114)

g8 (x) = 4.72 − 0.5 × x4 − 0.19 × x2 × x3 ≤ 4 (115)

g9 (x) = 10.58 − 0.674 × x1 × x2 − 0.672 × x2 ≤ 9.9 (116)

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 37

Vol.:(0123456789)
 www.nature.com/scientificreports/

g10 (x) = 16.45 − 0.489 × x3 × x7 − 0.843 × x5 × x6 ≤ 15.7 (117)

0.5 ≤ x1 , x3 , x4 ≤ 1.5, 0.45 ≤ x2 ≤ 1.35, 0.875 ≤ x5 ≤ 2.625, 0.4 ≤ x6 , x7 ≤ 1.2, (118)

Received: 23 June 2023; Accepted: 17 February 2024

References
1. Toloo, M., Talatahari, S., Gandomi, A. H. & Rahimi, I. Multiobjective combinatorial optimization problems: Social, keywords, and
journal maps. in Multi-Objective Combinatorial Optimization Problems and Solution Methods 1–9 (Elsevier, 2022).
2. Rahimi, I., Gandomi, A. H., Chen, F. & Mezura-Montes, E. A review on constraint handling techniques for population-based
algorithms: From single-objective to multi-objective optimization. Arch. Comput. Methods Eng. 30, 2181–2209 (2023).
3. Rahimi, I., Gandomi, A. H., Nikoo, M. R. & Chen, F. Extending boundary updating approach for constrained multi-objective
optimization problems. in International Conference on the Applications of Evolutionary Computation (Part of EvoStar) 102–117
(2023).
4. Deb, K., Pratap, A., Agarwal, S. & Meyarivan, T. A fast and elitist multiobjective genetic algorithm: NSGA-II. IEEE Trans. Evol.
Comput. 6, 182–197 (2002).
5. Binh, T. T. & Korn, U. MOBES: A multiobjective evolution strategy for constrained optimization problems. in The third international
conference on genetic algorithms (Mendel 97) vol. 25–27 (1997).
6. Gandomi, A. H. & Deb, K. Implicit constraints handling for efficient search of feasible solutions. Comput. Methods Appl. Mech.
Eng. 363, 112917 (2020).
7. Jiménez, F., Verdegay, J. L. et al. Evolutionary techniques for constrained multiobjective optimization problems. (1999).
8. Jordehi, A. R. A review on constraint handling strategies in particle swarm optimisation. Neural Comput. Appl. 26, 1265–1275
(2015).
9. Kaul, S. & Kumar, Y. Nature-inspired metaheuristic algorithms for constraint handling: Challenges, issues, and research perspec-
tive. Constraint Handling Metaheuristics Appl. 55–80 (2021).
10. Mezura-Montes, E. & Coello, C. A. C. Constraint-handling in nature-inspired numerical optimization: Past, present and future.
Swarm Evol. Comput. 1, 173–194 (2011).
11. Salcedo-Sanz, S. A survey of repair methods used as constraint handling techniques in evolutionary algorithms. Comput. Sci. Rev.
3, 175–192 (2009).
12. Diwekar, U. M. & Rubin, E. S. Efficient handling of the implicit constraints problem for the ASPEN MINLP synthesizer. Ind. Eng.
Chem. Res. 32, 2006–2011 (1993).
13. Raghavan, B., Xiao, M., Breitkopf, P. & Villon, P. Implicit constraint handling for shape optimisation with pod-morphing. Eur. J.
Comput. Mech. 21, 325–336 (2012).
14. Uemura, K., Nakashima, N., Nagata, Y. & Ono, I. A new real-coded genetic algorithm for implicit constrained black-box function
optimization. in 2013 IEEE Congress on Evolutionary Computation 2887–2894 (2013).
15. Mirabel, J. & Lamiraux, F. Handling implicit and explicit constraints in manipulation planning. Robot. Sci. Syst. 9 (2018).
16. Nomura, M., Sakai, N., Fukushima, N. & Ono, I. Distance-weighted exponential natural evolution strategy for implicitly constrained
black-box function optimization. in 2021 IEEE Congress on Evolutionary Computation (CEC) 1099–1106 (2021).
17. Deb, K., Pratap, A., Agarwal, S. & Meyarivan, T. A fast and elitist multiobjective genetic algorithm: NSGA-II. IEEE Trans. Evolut.
Comput. 6, 182–197 (2002).
18. Deb, K. Multi-objective optimisation using evolutionary algorithms: an introduction. in Multi-objective Evolutionary Optimisation
for Product Design and Manufacturing 3–34 (Springer, 2011).
19. Blank, J. & Deb, K. Pymoo: Multi-objective optimization in python. IEEE Access 8, 89497–89509 (2020).
20. Runarsson, T. P. & Yao, X. Stochastic ranking for constrained evolutionary optimization. IEEE Trans. Evolut. Comput. 4, 284–294
(2000).
21. Gonçalves, J. F. & Resende, M. G. C. Biased random-key genetic algorithms for combinatorial optimization. J. Heuristics 17, 487–525
(2011).
22. Michalewicz, Z. Genetic algorithms, numerical optimization, and constraints. Proc. Sixth Int. Conf. Genet. Algorithms 195, 151–158
(1995).
23. Deb, K. & Jain, H. An evolutionary many-objective optimization algorithm using reference-point-based nondominated sorting
approach, part I: Solving problems with box constraints. IEEE Trans. Evolut. Comput. 18, 577–601 (2013).
24. Seada, H. & Deb, K. A unified evolutionary optimization procedure for single, multiple, and many objectives. IEEE Trans. Evolut.
Comput. 20, 358–369 (2015).
25. Panichella, A. An adaptive evolutionary algorithm based on non-Euclidean geometry for many-objective optimization. in Proceed‑
ings of the Genetic and Evolutionary Computation Conference 595–603 (2019).
26. Angira, R. & Babu, B. V. Non-dominated sorting differential evolution (NSDE): An extension of differential evolution for multi-
objective optimization. in IICAI 1428–1443 (2005).
27. Reddy, S. R. & Dulikravich, G. S. Many-objective differential evolution optimization based on reference points: NSDE-R. Struct.
Multidis. Optimization 60, 1455–1473 (2019).
28. van Veldhuizen, D. A. Multiobjective Evolutionary Algorithms: Classifications, Analyses, and New Innovations (Air Force Institute
of Technology, 1999).
29. Ishibuchi, H., Masuda, H., Tanigaki, Y. & Nojima, Y. Modified distance calculation in generational distance and inverted genera-
tional distance. in International Conference on Evolutionary Multi-criterion Optimization 110–125 (2015).
30. Coello Coello, C. A. & Reyes Sierra, M. A study of the parallelization of a coevolutionary multi-objective evolutionary algorithm.
in MICAI 2004: Advances in Artificial Intelligence: Third Mexican International Conference on Artificial Intelligence, Mexico City,
Mexico, April 26–30, 2004. Proceedings 3 688–697 (2004).
31. Blank, J. & Deb, K. A running performance metric and termination criterion for evaluating evolutionary multi-and many-objective
optimization algorithms. in 2020 IEEE Congress on Evolutionary Computation (CEC) 1–8 (2020).
32. Behmanesh, R., Rahimi, I. & Gandomi, A. H. Evolutionary many-objective algorithms for combinatorial optimization problems:
A comparative study. Arch. Comput. Methods Eng. 28, 673–688 (2021).
33. Deb, K. Multi-objective Optimization (Springer, 2014).
34. Coello, C. A. C. Evolutionary Algorithms for Solving Multi-objective Problems (Springer, 2007).
35. Deb, K. Multi-objective Optimisation Using Evolutionary Algorithms: an Introduction (Springer, 2011).
36. Corne, D. W., Knowles, J. D. & Oates, M. J. The Pareto envelope-based selection algorithm for multiobjective optimization. in
International Conference on Parallel Problem Solving from Nature 839–848 (2000).

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 38

Vol:.(1234567890)
www.nature.com/scientificreports/

37. Horn, J., Nafpliotis, N. & Goldberg, D. E. A niched Pareto genetic algorithm for multiobjective optimization. in Proceedings of the
first IEEE Conference on Evolutionary Computation. IEEE World Congress on Computational Intelligence 82–87 (1994).
38. Price, K., Storn, R. M. & Lampinen, J. A. Differential Evolution: A Practical Approach to Global Optimization (Springer Science &
Business Media, 2006).
39. Rechenberg, I. Evolution strategy: Nature’s way of optimization. in Optimization: Methods and Applications, Possibilities and
Limitations 106–126 (Springer, 1989).
40. Gonçalves, J. F. & Resende, M. G. C. Biased random-key genetic algorithms for combinatorial optimization. Journal of Heuristics
17, 487–525 (2011).
41. Deb, K. An efficient constraint handling method for genetic algorithms. Comput. Methods Appl. Mech. Eng. 186, 311–338 (2000).
42. Youn, B. D., Choi, K. K., Yang, R.-J. & Gu, L. Reliability-based design optimization for crashworthiness of vehicle side impact.
Struct. Multidisc. Optim. 26, 272–283 (2004).
43. Golinski, J. An adaptive optimization system applied to machine synthesis. Mech. Mach. Theory 8, 419–436 (1973).
44. Osyczka, A. & Kundu, S. A new method to solve generalized multicriteria optimization problems using the simple genetic algo-
rithm. Struct. Optimization 10, 94–99 (1995).
45. Rao, S. S. Engineering Optimization: Theory and Practice (Wiley, 2019).
46. Vanderplaats, G. Very large scale optimization. in 8th Symposium on Multidisciplinary Analysis and Optimization 4809 (2002).
47. Jain, H. & Deb, K. An evolutionary many-objective optimization algorithm using reference-point based nondominated sorting
approach, part II: Handling constraints and extending to an adaptive approach. IEEE Trans. Evolut. Comput. 18, 602–622 (2013).
48. Deb, K. & Sundar, J. Reference point based multi-objective optimization using evolutionary algorithms. in Proceedings of the 8th
Annual Conference on Genetic and Evolutionary Computation 635–642 (2006).

Author contributions
I.R.: Conceptualization, Methodology, Formal analysis, Software, Writing—Original draft preparation, Data
curation, Investigation. A.H.G.: Conceptualization, Supervision, Methodology, Validation, Writing—review &
editing, M.R.N. and M.M.: Revising, Review & Editing, Visualization. F.C.: Co-supervisory.

Funding
Open access funding provided by Óbuda University.

Competing interests
The authors declare no competing interests.

Additional information
Correspondence and requests for materials should be addressed to A.H.G.
Reprints and permissions information is available at www.nature.com/reprints.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International
License, which permits use, sharing, adaptation, distribution and reproduction in any medium or
format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the
Creative Commons licence, and indicate if changes were made. The images or other third party material in this
article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the
material. If material is not included in the article’s Creative Commons licence and your intended use is not
permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder. To view a copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/.

© The Author(s) 2024

Scientific Reports | (2024) 14:4816 | https://doi.org/10.1038/s41598-024-54841-z 39

Vol.:(0123456789)