Vol. 13(1), pp.

39-50, January-June 2020

DOI: 10.5897/AJMCSR2019.0804
Article Number: 8AB802563098
ISSN: 2006-9731
Copyright©2020 African Journal of Mathematics and Computer
Author(s) retain the copyright of this article
http://www.academicjournals.org/AJMCSR Science Research

Full Length Research Paper

A mathematical model for solving integer linear

programming problems
Ammar E. E. and Emsimir A. A*.
Department of Mathematics, Faculty of Science, Tanta University, Egypt.
Received 2 July 2019; Accepted 30 January 2020

A suggested algorithm to solve fully rough integer linear programming (FRILP) problems is introduced
in this paper in order to find rough value optimal solutions and decision rough integer variables, where
all parameters and decision variables in the constraints and the objective function are rough intervals
(RIs). In real-life situations, the parameters of linear programming problem model may not be defined
precisely, because of globalization of the market, uncontrollable factors, etc., hence for that the FRILP
problem solving methodology is presented using the slice-sum method with the branch and bound
technique, where we will construct two integers linear programming (ILP) problems with interval
coefficients and variables. One of these problems is an ILP problem, where all of its coefficients are
upper approximations interval (UAI) of rough intervals and represents rather satisfactory solutions, the
other is an ILP problem where all of its coefficients are lower approximations interval (LAI) of rough
intervals and represents complete solutions. Thereafter, the two ILP problems are sliced into four crisp
problems. Integer programming is used because many linear programming (LP) problems require that
the decision variables should be integers. In addition, rough intervals are very important to tackle the
uncertainty and imprecise data in decision making problems. Furthermore, the proposed algorithm
enables us to search for the optimal solution in the largest range of possible solutions. A flowchart is
also provided to illustrate the problem-solving steps. Finally, some examples are given to demonstrate
the results.

Key words: Integer linear programming, rough set theory, full rough interval coefficients and variables, upper
approximation, lower approximation, optimal solution, crisp coefficients.

INTRODUCTION

Linear programming (LP) is one of the most popular and integer variables. More widely, the applications of
models used in decision making and optimization integer programming can be used to appropriately
problems. Lots of research, studies and applications of describe the decision problems concerning the effective
LP models have been reported in numerous books, use of resources in engineering technology, business
monographs, articles and chapters in books like management and other numerous fields. According to
Chinneck and Ramadan (2000). Taha (1997) introduced Pawlak and Skowron (2007), for a vague concept rough
Integer programming (IP) problems as optimization (R), a lower approximation is contained for all objects
problems that minimize or maximize the objective which surely belong to the concept R and an upper
function, taking into consideration the limits of constraints approximation is contained for all objects which possibly

*Corresponding author. E-mail: [email protected].

Author(s) agree that this article remain permanently open access under the terms of the Creative Commons Attribution
License 4.0 International License
40 Afr. J. Math. Comput. Sci. Res.

belong to the concept R. In other words, the lower coefficients to find rough value optimal solutions and
approximation of the concept is the union of all the decision rough integer variables in order to reach
elementary concepts which are included in it, whereas solutions, such as completely satisfactory solutions (sure
the upper approximation is the union of all the elementary solutions) and rather satisfactory solutions (possible
concepts that have non-empty intersection with the solutions) by lower approximation interval and upper
concept. The concept of rough variable, which is a approximations interval respectively. All parameters and
measurable function from rough space to the set of real decision variables in the constraints and the objective
numbers, was proposed by Liu (2012). Garg (2015) also function are rough intervals (RIs). The linear programming
presented an alternative approach for solving the multi- problems with integer restrictions on the decision
objective reliability optimization problem by utilizing the variables are called integer programming problems which
uncertain, vague and imprecise data. Pandian et al. form a special class of the linear programming. This type
(2016) believed that transportation problem has all or of problems is of particular importance in business and
some parameters as rough integer intervals. They also industry where quite often the discrete nature of variables
proposed a new method named, a slice-sum method to is involved in many decision-making situations. For
solve Rough Integer Interval Transportation Problem example, in the manufacturing field, the problem is that
(RIITP), where transportation cost, supply and demand the frequently scheduled interims of batches, lots,
are rough integer intervals. Rani et al. (2016) presented distribution and shipment must involve a discrete number
an algorithm for solving multi-objective optimization of trucks, aircrafts, or freight cars; hence, integer
problem under the optimistic and pessimistic viewpoint. programming problem has been applied to solve many
Garg and Arora (2018) developed a methodology for real-world problems. Nonetheless, this still fails to deal
solving the decision-making problems under the with the imprecise and uncertain data. Many researchers
environment in which the ratings of all alternatives on have succeeded in capturing imprecise information by
attributes and preference information on attributes are fuzzy linear programming problem (Bellman and Zadeh,
incomplete and expressed with the sets. Pandian et al. 1970). For more clarification, the coefficients of integer
(2018) introduced a new method namely; level-bound linear programming problems (some or all) in the
method that was proposed to solve fuzzy interval integers objective function and constraints are assumed to be
transportation problems. Jamkhaneh and Garg (2018) known and fixed during the model application period,
presented generalized intuitionistic fuzzy sets and their which in practice is not possible due to some
corresponding operations, and some new arithmetic and measurement errors or instability of the market
geometric mean operations are defined to aggregate the conditions… etc. In this study, these situations can be
different preferences of the decision makers during the modelled efficiently through rough intervals integer linear
process. Garg (2018) presented an alternative method, programming. The motivation behind the study is to
under the fuzzy environment, for computing the various enable the decision maker to make the right decision in
arithmetic operations of a system, using the sigmoidal the field of proposed solutions, while dealing with the
number and used a-cut approach for finding the uncertain and imprecise data. The elements of the paper
membership function of the system. More widely applied are organized as follows. First, some bases of the
integer programming can be used to appropriately preliminaries of RIs are presented, followed by suggestion
describe the decision problems with the effective use of of a solution method for ILP with interval coefficients and
resources in engineering technology, business variable. Thereafter, a case study to illustrate the general
management and other numerous fields (Shaocheng, formulation of the problem under consideration and an
1994; Taha, 1997). Osman et al. (2016) introduced a integer linear programming with fully rough intervals is
solution approach for rough interval multi-objective presented, which comprises the use of a slice-sum
transportation problem (RIMOTP). The concept of solving method (Pandian et al., 2016) for solving FRILP problems
conventional interval programming combined with fuzzy as well as numerical examples for demonstrating the
programming is used to build the solution approach for solution procedure of the proposed method and finally,
RIMOTP. (De 2017) deals with a triangular dense fuzzy concluding remarks are outlined.
set having special property on Cauchy sequence. (De
and Mahata 2019a) deals with a classical economic order
quantity (EOQ) model under monsoon type fuzzy BASIC PRELIMINARIES
demand rate. Also, De and Mahata (2019b) introduced a
comprehensive study of an economic order quantity Here some definitions and properties of rough intervals
model under fuzzy monsoon demand and the solution are given. An RI can be considered as a qualitative value
was obtained with the help of a nonlinear optimization from uncertainty and vague concept defined on a variable
technique that requires maximum aspiration level of the x in real numbers . In addition, the transformation of
fuzzy membership of the objective function. In this paper, the linear programming problem decision parameters and
the focus of our study is to develop a method for solving decision variables into the upper and lower approximation
integer linear programming problems with rough interval intervals is usually a hard work for many cases, and
 Ammar and Emsimir 41

conversion process needs the following definitions to be Let [[ ] [ ]] and

known. Further details are found in Hamazehee et al. [[ ] [ ]] be two RIs if , ≥ 0.
(2014) and Rebolledo (2006): Then.

[Addition:] = [[ ] [
Definition 1 ]]

''The qualitative value is called a rough interval (RI) [Subtraction:] = [[ ] [

when two closed intervals are assigned:
]]
and on to where .Moreover,
[Negation:] =[[ ] [ ]].
1. If then A surely takes [ { } { }]
2. If [Intersection:] ( )
[ { } { }]
3. If
[ { } { }]
are called the lower approximation [Union:] =( ) ''
[ { } { }]
interval (LAI) and the upper approximation interval (UAI)
of , respectively. Moreover, is denoted by = ( ).
Also, the intervals and are not the complement of INTEGER LINEAR PROGRAMMING WITH INTERVAL
each other" (Hamazehee et al., 2014). COEFFICIENTS
Note: We can symbolize [ ] and The more relevant method of solving integer linear
[ ] where LL, UL, LU, and UU represent lower programming with interval coefficients (ILPIC) was first
lower, upper lower, lower upper, upper upper proposed by Shaocheng in 1994. After that, Chinneck
respectively. and Ramadan completed Shaocheng (1994) studies in
2000. They transformed the original interval LP into two
LPs with crisp coefficients. Thereafter, linear
Example 1 programming with rough interval coefficients (LPRIC) was
discussed by Hamzehee et al. (2014). Here, we will
If A is the concept of “Hot” as the qualitative value consider all parameters and decision variables in the
defined on a temperature variable x in R, one may constraints and the objective function to be rough
consider: intervals (RIs). Also, the decision variables are
considered integers, and the problem is introduced as
[[ ] [ ]] follows:

Then, by Definition 1, the temperature variable x surely ∑ [ ]

takes a value between 25 and 45°C. Similarly, x possibly ∑ [ ] [ ] } (1)
takes a value between 15 and 65°C. Note that, common
sense and general knowledge can help define and .
For instance:
Where for all i, j : [ ][ ] [ ] are closed
– The temperature of the human body is about 37°C; any intervals on real numbers and where lower (L), upper
warmer temperature would not be considered “Cold”. (U) refer to the minimum and the maximum in the interval
– On average, a human hand cannot hold an object, respectively.
whose temperature is over 65°C, because it is “Hot”. Now, from a previous study by Chinneck et al. (2000),
– If something is colder than the environment (let us say we can use some definitions and theorems for
15°C, it cannot be considered “Warm” anymore. maximization of problem (1) for the inequality constraints
as follows:

∑ [ ] [ ]
Definition 2
The above inequality constraints have p interval
''The arithmetic operations on RIs are depending on coefficients in the right and/or left-hand sides. Thereafter,
interval arithmetic, so we will state some of these it can be transformed into different extreme inequalities
arithmetic operations as follows'' (Rebolledo, 2006): by setting the interval coefficients at appropriate
42 Afr. J. Math. Comput. Sci. Res.

Table 1. Production data for an example 2.

Required production time in department (h/unit)

Product Seasonal profit ($/unit)
1 2 3
A [3,5] [2,4] [1,2] [1,3]
B [2,4] [1,3] [3,5] [2,4]
Available time [550,700] [380,500] [200,300] -

combinations of the bounding values on the coefficient ∑

ranges", such as in inequality constraints where P =2, ∑ } (2)
. Let be the set of all solutions for the extreme
inequality. Then:
(2) The worst optimal solution is found by solving:
⋃ and ⋂
∑
∑ } (3)
Definition 3

''For inequality constraints, if there exists one extreme Briefly, and used the greatest and the smallest
inequality that its solution set is the same as ( ), then suitable value of the objective function and the maximum
it is called the maximum value range (minimum value (minimum) value range inequalities of problem (1).
range) inequality'' (Hamazehee et al., 2014). Note that, there are 3 possible outcomes for ILPIC
problem (1) as follows:
Theorem 3a - If and have optimal solutions, then ILPIC
problem (1) has a finite bounded optimal range.
Consider inequality constraints, where - If is unbounded then ILPIC problem (1) is
Then, ∑ and ∑ be the unbounded.
minimum value range and maximum value range - If is infeasible then ILPIC problem (1) is infeasible.
inequalities, respectively (Hamazehee et al., 2014).
More clarification and proof can be found in Hamazehee
et al. (2014).
Theorem 3b
Example 2. A company produces two sorts of products,
Consider ILPIC problem (1) (Hamazehee et al., 2014). A and B, and three departments 1, 2 and 3 are assigned
Then, for any given feasible solution , we have: to produce them. Each product takes a time to be
produced by the three departments; the following Table 1
∑ ∑ shows the details.
Thereafter, the model of this optimization problem can
for all be formulated as follows:

ILPIC = [ ] [ ]
Definition 4
[ ] [ ] [ ]
''For a given feasible solution of ILPIC Problem
(1), the value ∑ ∑ is called the most [ ] [ ] [ ]
suitable value (the least suitable value) of the objective
function'' (Hamazehee et al., 2014). [ ] [ ] [ ]
From the above Theorems 1 and 2 and the Definitions
3 and 4, we can find the best and worst optimal solutions
of the ILPIC Problem (1). By transforming the original
ILPIC problem (1) into two classical ILP problems, we Here, the optimal range of ILPI for Example 2 can be
can call them and . obtained by solving two classical ILPs as follows:

(1) The best optimal solution is found by solving: The best optimal solution
 Ammar and Emsimir 43

Table 2. The first plan.

Required production time in department (h/unit) Seasonal profit

Product
1 2 3 ($/unit)
A [3,4] [2,4] [1,2] [2,3]
B [2,4] [3,4] [3,5] [2,3]
Available time [600,650] [400,450] [200,300] -

Table 3. The second plan.

Required production time in department (h/unit)

Product Seasonal profit ($/unit)
1 2 3
A [2,4] [2,3] [1,3] [2,4]
B [2,3] [2,4] [3,4] [2,3]
Available time [550,700] [380,500] [230,260] -

Formulating linear programming model requires that

specific values be chosen for the model coefficients.
However, the values of many of these coefficients are
only approximately known. For instance, the Director of
the company wishes to know the range of optimal
solutions that could be returned by an ILP model with
uncertain coefficients. On the other hand, the manager
may put some plans or ask the opinions of some experts
to determine the uncertain coefficients. Now, we will give
a case study to clarifying the idea.

and Case study

The worst optimal solution
A company produces two sorts of products, A and B, and three
departments 1, 2 and 3 are required to produce them. Each product
takes a time to be produced by the three departments. Production
experts put three plans for production due to the uncertainty of
parameters. Each plan gives the approximate production time of
each product in the 3 departments and seasonal profit of each
product as shown in Tables 2, 3 and 4, e.g. the required production
time for product A in Department 1 due to the first plan is between 3
and 4 h per unit and its seasonal profit is between $2 and $3 per
unit with hours of production time available in the three
departments.
The production experts would like to allocate the production
capacity so that it not only finds the optimal product mixture but also
use the three plans together. In order to solve this problem, the
parameters should be determined. Since the parameters are given
by the three plans, one way to involve all plans is representing the
We then apply the branch and bound to Example 2 to parameters by rough interval. In Table 5, the upper and lower
approximations of rough interval coefficients are computed based
find integer optimal solutions. on the union and intersection of the plans.
Now, in the case study, let and denote the respective
[ ] [ ] [ ] amounts of products 1 and 2 which should be produced. Then, the
model of this optimization problem can be formulated in Example 3
as follows:
INTEGER LINEAR PROGRAMMING WITH FULLY
𝑈 𝑈 𝑈𝑈
[[2,3]: [1,4]] ⨂[[ 1 , 1 ]: [ 1 , 1 ]]⨁
ROUGH INTERVALS 𝑳𝑷𝑭𝑹𝑰 = ([ ], [ 𝑈
]) = Max 𝑈 ]: [ 𝑈 𝑈𝑈
[[2,3]: [2,4]] ⊗ [[ 2 , 2 2 , 2 ]]
[[ 3,4]: [2,5]] ⨂[[ 1 , 1𝑈 ]: [ 1 𝑈 , 1𝑈𝑈 ]]⨁[[2,3]: [2,4]]
In this part, we consider the linear programming problems S.t
⨂[[ 2 , 2𝑈 ]: [ 2 𝑈 , 2𝑈𝑈 ]] [[600,650]: [550,700]]
with rough interval coefficients and variables (ILPFRI).
44 Afr. J. Math. Comput. Sci. Res.

Table 4. The third plan.

Required production time in department (h/unit) Seasonal profit

Product
1 2 3 ($/unit)
A [3,5] [2,4] [1,2] [1,3]
B [2,4] [1,3] [3,5] [2,4]
Available time [550,700] [380,500] [200,300] -

Table 5. The coefficients of the problem.

Required production time in department (h/unit) Seasonal profit

Product
1 2 3 ($/unit)
A ([3,4]: [2,5]) ([2,3]: [2,4]) ([1,2]: [1,3]) ([2,3]: [1,4])
B ([2,3]: [2,4]) ([3,3]: [1,4]) ([3,4]: [3,5]) ([2,3]: [2,4])
Available time ([600,650]: [550,700]) ([400,450]: [380,500]) ([230,260]: [200,300]) -

[[ ] [ ]] ⨂[[ ] [ ]]⨁[[ ] [ ]] [ ] [ ] ,
⨂[[ ] [ ]] [[ ] [ ]] [ ] [ ]
[ ] [ ]
[[ ] [ ]] ⨂[[ ] [ ]]⨁[[ ] [ ]] [ ] [ ] ,
⨂[[ ] [ ]] [[ ] [ ]]

{ } Definition 5

Example 3 may involve other constraints and variables, in other In problem (4), we can define the following sets:
words, the constraints can be of the form ≤, ≥ or =. Also, variables
can be sign-restricted (x ≤ 0 or x ≥ 0) or unrestricted in sign. |∑
This study will be limited to the variables which are sign-restricted
as .
|∑
Problem formulation

The general formula of the integer linear programming problem with |∑

fully rough interval for coefficients and variables (ILPFRI) may be
presented as:
|∑
]: [ 𝑈 ]] = 𝑴𝒂𝒙 ∑ 𝑈 𝑈 𝑈𝑈 𝑈 𝑈 𝑈𝑈
[[ =1 [[ , ]: [ , ]] ⨂ [[ , ]: [ , ]]
𝑈 𝑈 𝑈𝑈 𝑈 𝑈 𝑈𝑈
∑𝒎
𝒊=𝟏 ∑ =1 [[ , ]: [ , ]] ⨂ [[ , ]: [ , ]]
} Where
𝑈 𝑈 𝑈𝑈 (4)
[[ , ]: [ , ]] ( .
𝑈 𝑈 𝑈𝑈
, , , 0 , ,
Note: for the sets we have:
(14) (Hamazehee et al., 2014).
Where

[[ ] [ ]] , [[ ] [ ]] Definition 6

[[ ] [ ]] [[ ] [ ]] Consider all of the corresponding ILPFRI problems and LP of

problem (4) (Atteya, 2016; Bazaraa, 2010).

are rough interval coefficients and variables of the objective a) The interval [ ] [ ] is called the surely
function and the constraints. Also, let denote the vector of all (possibly) optimal range symbolized [ ] [ ] of problem
decision variables. (4), if the optimal range of each (ILPFRI) is a superset (subset) of
[ ] [ ]
Remark 1: According to rough interval properties introduced earlier, b) Let [ ] [ ] be surely (possibly) optimal
we have range of the problem (4). Then the rough interval
 Ammar and Emsimir 45

[[ ] [ ]] is called the rough optimal range of By Definition 5, the feasible solution set of problems (7) and (8) is
problem (4); also any point, optimal value belongs to equivalent to . Thus, the interval [ ] is the
[ ] [ ] and is called a completely (rather) optimal possible solution range of problem (5).
satisfactory solution of the problem (4). We can see that the interval [ ] is equivalent to the
c) A solution is surely-feasible, iff it belongs to the lower possibly optimal range of problem (4). Toward this end, since
approximation of the feasible set. problem (1) is an arbitrary corresponding LPIC problem of (4), we
d) A solution is possibly -feasible, iff it belongs to the upper have
approximation of the feasible set.
e) A solution is surely-not feasible, iff it does not belong to the [ ] [ ] [ ] [ ]
upper approximation of the feasible set. [ ] [ ] [ ] [ ]
[ ] [ ] [ ] [ ]
Solution procedures [ ] [ ] [ ] [ ]

In order to solve problem (4), we will construct two integers linear

programming (ILP) problems with interval coefficients and variables.
One of these problems is an ILP problem where all of its Step 4: The surly optimal range solution of problem (6) can
coefficients are upper approximations interval (UAI) of rough be obtained by solving two classical LPs as:
intervals and represents rather satisfactory solutions, while the
other is an ILP problem where all of its coefficients are lower
approximations interval (LAI) of rough intervals and represents ∑
complete solutions. Then the two ILP problems are sliced into four
crisp problems through the following steps. ∑

Step 1: Find the possibly optimal range solution

[ ] by solving the upper approximation interval of
problem (4).
and
∑ [ ]⨂[ ]
∑
∑ ∑ [ ]⨂[ ] [ ] (5) ∑
(10)

By Definition 5, the feasible solution set of problems (9) and (10) is

Step 2: Find the surly optimal range solution equivalent to . Thus, the interval [ ] is the
[ ] by solving the lower approximation of optimal solution range of problem (6). Also, the
problem (4). interval [ ] is the optimal possible solution range for
problem (5) and the interval [ ] is the sure solution range
∑ [ ]⨂[ ] (6).
Where the rough value optimal solutions and decision
∑ ∑ [ ]⨂[ ] [ ] rough integer variables in problem (4) will be as:
(6)

[[ ] [ ]],

Step 3: According to Definition 6, the possibly optimal range [[ ] [ ]],

solution of problem (5) can be obtained by slicing it into two
classical ILPs as follows: .

∑ Where the possibly optimal values range solutions for

[ ].
∑ Also, the surely optimal values range solutions for
[ ] In addition, [ ] is the integer completely
satisfactory solutions, and [ ] is integer rather satisfactory
solutions.
Now, a relation has been established between optimal solutions
and of the integer linear programming problem with fully rough interval
in problem (4) and four problems ,
∑ problems (7), (8), (9) and (10) respectively. The established relation
is used in the proposed method, slice-sum method.
∑
Theorem 4

If the set { } is an optimal solution for the or

46 Afr. J. Math. Comput. Sci. Res.

(7) problem of the problem (ILPFRI) or (4) with the maximum To solve the problem indicated in the case study, we should solve
optimal value for , the set { } is an optimal two integer linear programming fully interval problems ILPFI =
solution for the or (10) problem of the problem (ILPFRI) or [ ] as follows:
(4) with the maximum optimal value for , the set
{ } is an optimal solution for the or (9) problem [ ]
of the problem (ILPFRI) or (4) with the maximum optimal value for
, and the set { } is an optimal solution for the [ ]⨂[ ]⨁ [ ]⨂[ ]
or (8) problem of the problem (ILPFRI) or (4) with the
maximum optimal value for , then the set of rough integer Subject to
intervals {( [ ] [ ]) } is an optimal
[ ]⨂[ ]⨁[ ]⨂ [ ]
solution for the problem (4) with maximum optimal values
([ ] [ ]) provided [ ]

Now, since { }{ } { [ ] ⨂[ ] ⨁[ ]⨂[ ]

}{ } are
optimal solutions for the problems
,respectively and [ ]
then we can conclude that the set of rough integer
[ ]⨂[ ] ⨁ [ ]⨂[ ]
intervals {( [ ] [ ]) } is a feasible solution
to the problem (4). [ ]
Let {( [ ] [ ]) } be a feasible
solution to the problem (4).
Therefore, { }{ } {
} { } are feasible solutions to the problems [ ]
respectively.
Since { }{ } { [ ]⨂[ ]⨁ [ ]⨂ [ ]
}{ } are feasible solutions to the problems
Subject to
respectively. We have:
[ ]⨂ [ ]⨁ [ ]⨂ [ ]
∑ ∑ [ ]

[ ]⨂ [ ]⨁[ ]⨂[ ]
∑ ∑
[ ]

[ ]⨂ [ ]⨁ [ ] ⨂[ ]
∑ ∑
[ ]

∑ ∑
- In the ILPFI Problem, ( ) is transformed to ILP problems
and where their feasible sets are ,
This implies that respectively.
- In the ILPFI Problem, ( ) is transformed to ILP problems
[ ] [ ] and , where their feasible sets are
∑ [ ] [ ] ⨂ [ ] [ ] respectively.

max
𝒎
S.t
∑([ ] [ ]) ⊗ ([ ] [ ])
𝟏
𝒎

∑([ ] [ ])
𝟏
⊗ ([ ] [ ])
max
Therefore, the set of rough integer intervals S.t
{( [ ] [ ]) } is an optimal solution for the
problem (4) with maximum optimal values ([ ] [ ]).
Hence, the theorem is proved (Pandian et al., 2016). Figure 1
shows flowchart of the solution steps, which was provided for more
clarification.
 Ammar and Emsimir 47

Figure 1. Flowchart of the proposed approach for solving fully rough integer
linear programming problems.

Table 6. Optimal values and solutions of ILP programs in the case study.

Problem 𝑰𝑳𝑷𝑳 𝑰𝑳𝑷𝑳𝑳 𝑰𝑳𝑷 𝑳 𝑰𝑳𝑷

Optimal values 66 230 660 1040
𝒙𝟏𝑹 66 115 210 240
𝒙𝑹 0 0 10 20

max 6. Optimal values and solutions of ILP programs in the case study
S.t are presented in Table 6. Here the integer rough optimal solutions
is

[[ ] [ ]]

[[ ] [ ]].

max The rough optimal values range solutions for

S.t
[[ ] [ ]] [[ ] [ ]].

Where the possibly optimal values range solutions for

[ ] [ ]
Also, the surely optimal values range solutions for
[ ] [ ]. In addition,
We used ''WinQSB'' program (Olga et al., 2009) to find rough value [ ] [ ][ ] [ ] is integer completely
optimal solutions and decision rough integer variables satisfactory solutions and [ ] [ ][ ]
and apply branch and bound algorithm (Gupta and Mohan, 2006) to [ ] is integer rather satisfactory solutions.
find integer optimal solutions for the case study as shown in Table We think that this is the way to the solution, where the director of
48 Afr. J. Math. Comput. Sci. Res.

the company has put all the possible solutions to avoid the
economic crises and market fluctuations (the uncertainty of { }
parameters).
and

Example 3

Consider the following ILPFRI problem as follows: [ ]⨂[ ]⨁[ ]⨂[ ]

⨁[ ]⨂[ ]⨁[ ]⨂[ ] }
[[ ] [ ]] ⨁[ ]⨂[ ]⨁[ ]⨂[ ]

[ ] [ ] ⊗ ⨁ [ ] [ ] ⊗ Subject to
⨁ [ ] [ ] ⊗ ⨁ [ ] [ ] ⊗ }
[ ] [ ] ⊗ ⨁ [ ] [ ] ⊗ [ ]⨂[ ]⨁[ ]⨂[ ]⨁[ ]
⨂[ ]⨁[ ]⨂ [ ]⨁[ ]⨂ }
Subject to [ ]⨁[ ]⨂[ ] [ ]
[ ]⨂[ ]⨁[ ] ⨂ [ ]
[
] [ ]⨂ ⨁ [ ] [ ] ⨂ ⨁ { }
[ ]
[
] [ ] ⨂ ⨁ [ ] [ ] ⨂ [ ]⨂[ ]⨁[ ]⨂[ ]
{ }
[ ] [ ] ⨂ ⨁ [ ] [ ] ⨂ [ ]
[ ] [ ] [ ]⨂[ ]⨁[ ]⨂[ ]
{ { }
[ ] [ ] ⨂ ⨁ [ ] [ ] ⊗ [ ]
{ }
[ ] [ ] { }
[ ] [ ] ⨂ ⨁ [ ] [ ] ⨂
{ }
[ ] [ ]
[ ] [ ] ⨂ ⨁ [ ] [ ] ⨂ According to Solution Procedures, the optimal range of ILPIC
{ } Problems and can be obtained by solving four
[ ] [ ]
classical LPs as follows:
{ }

Where

[ ] [ ] , Subject to
[ ] [ ]

[ ] [ ] ,

[ ] [ ]

[ ] [ ] and
And
[ ] [ ]

In order to solve problem (Example 3), we will solve two ILPFI

problems as follows:
Subject to

[ ]⨂[ ]⨁ [ ]⨂[ ]
⨁[ ] ⨂[ ]⨁[ ]⨂[ ]}
⨁[ ]⨂[ ]⨁[ ]⨂[ ]

Subject to

[ ]⨂[ ]⨁[ ]⨂[ ]⨁[ ]

⨂[ ]⨁[ ]⨂[ ]⨁[ ] ⊗ }
[ ]⨁[ ]⨂[ ] [ ]
[ ]⨂[ ]⨁[ ]⨂[ ]
{ }
[ ] Subject to
[ ]⨂[ ]⨁[ ] ⨂[ ]
{ }
[ ]
[ ]⨂[ ]⨁[ ]⨂[ ]
{ }
[ ]
 Ammar and Emsimir 49

Table 7. Optimal values and solutions of Example 3. intervals. Thus, we have completely satisfactory solutions and
rather satisfactory solutions. Then, we give the decision maker
Problem 𝑰𝑳𝑷𝑳 𝑰𝑳𝑷𝑳𝑳 𝑰𝑳𝑷 𝑳 𝑰𝑳𝑷 more freedom to choose.
Optimal values 14 46 328 1065
𝒙𝟏𝑹 1 2 8 30 DISCUSSION
𝒙𝑹 0 0 0 0
𝒙𝑹 0 0 0 0 Comparing the linear programming with rough Interval
𝒙𝑹 1 4 10 15 coefficients by Hamazehee et al. (2014), who only used
𝒙𝑹 1 2 8 15 the rough interval with coefficients in linear programming
problems, and with fuzzy interval integer transportation
𝒙𝑹 0 0 0 0
problems by Pandian et al. (2018), who used a new
method namely; level-bound method that was proposed
to solve fuzzy interval integers transportation problems,
this study used fully rough intervals integer linear
programming problems, where all parameters and
decision variables in the constraints and the objective
function are rough intervals, since many linear
programming problems in our real life require that the
decision variables be integers. In addition, rough intervals
are very important to tackle the uncertainty and imprecise
Subject to data in decision making problems. Moreover, the
proposed algorithm enables us to search for the optimal
solution in the largest range of possible solutions.
Furthermore, N suggested solutions are obtained to
enable the decision maker to choose the best decisions.
On the other hand, some solutions, such as completely
satisfactory solutions (surely solutions) as in problem (6)
To find rough value optimal solutions decision rough integer (step 2), are successfully reached, which lets us be sure
variables and apply branch and bound algorithm (Gupta and that the optimal solution is in the lower approximation
Mohan, 2006) to find integer optimal solutions of Example 3 as interval. While the rather satisfactory solutions (possible
shown in Table 7.
The possibly optimal values range solutions for solutions), as in problem (5) (step 1), makes it possible
that the optimal solution is in the upper approximation
[ ] [ ] interval. The slice-sum method can be served as an
important tool for the decision makers, when they are
the surely optimal values range solutions for handling various types of logistic problems with rough
[ ] [ ] in addition, the rough optimal
values range solutions ILPRI
variable parameters. Furthermore, the branch and bound
technique is used to reach the integer programming. The
[ ][ ] [[ ][ ]] results, in the form of rough intervals method, do not
ignore any part of the solution area. The motivation
Where behind the study is to enable the decision maker to make
[ ] [ ] [ ] [ ], the right decision in the field of proposed solutions, in
[ ] [ ] [ ] [ ]
[ ] [ ][
case of having to deal with the uncertainty and imprecise
] [ ].
data. Finally, to clarify the idea and support this paper,
Are the integers completely satisfactory solutions. And some examples are solved by "WinQSB" program (Olga
et al., 2009) in order to illustrate the new concepts.
[ ] [ ][ ] [ ]
[ ] [ ][ ] [ ]
[ ] [ ][ ] [ ]
Conclusion
are the integers’ rather satisfactory solutions. Also,
Some basic concepts of rough intervals are reviewed in
[ ] [ ] [ ] [ ] this research paper. Then we presented a methodology
for solving fully rough integer linear programming
[ ] [ ] [ ] [ ] problems and found rough value optimal solutions and
[ ] [ [ ] [ ] .
decision rough integer variables, where all parameters
]
and decision variables in the constraints and the objective
are the integers rough optimal solutions. functions are rough intervals. The proposed model
Note that, decision variables and the optimal values are rough depends on slice-sum method, branch and bound
50 Afr. J. Math. Comput. Sci. Res.

method and integer programming, which is a good Garg H, Arora R (2018). A nonlinear-programming methodology for
multi-attribute decision-making problem with interval-valued
technique for many LP problems which require that the
intuitionistic fuzzy soft sets information. Applied Intelligence
decision variables are integers. Also, the used rough 48(8):2031-2046.
intervals are very important to tackle the uncertainty in Gupta PK, Mohan M (2006). Problems in Operations Research, Sultan
decision making problems. In addition, we obtained N Chand and Sons, New Delhi.
Hamazehee A, Yaghoobi M, Mashinchi M (2014). "Linear Programming
suggested solutions in order to enable the decision
with Rough Interval Coefficients". Journal of Intelligent and Fuzzy
maker to take the best decision. Furthermore, we got on Systems 26(3):1179-1189.
solutions such as completely satisfactory solutions (surely Jamkhaneh B, Garg H (2018). Some new operations over the
solutions) and rather satisfactory solutions (possibly generalized intuitionistic fuzzy sets and their application to decision-
making process. Granular computing 3(2):111-122.
solutions) by lower approximation interval and upper
Liu B (2012). Theory and Practice of Uncertain Programming (Vol. 239).
approximations interval respectively. The results are in Berlin: Springer.
the form of intervals and the interval method does not Olga IA, Doina F, Gheorghe P, Codruta OH (2009). ''WinQSB''
ignore any part of solution area. It is thought that the simulation software - a tool for professional development. Science
Direct 1(4):2786-2790.
rough intervals are useful new tools to tackle the
Osman S, El-Sherbiny M, Khalifa A, Farag H (2016). A Fuzzy Technique
uncertainty, vague and imprecise data in decision making for Solving Rough Interval Multiobjective Transportation Problem.
problems. Also, a flowchart of the steps to solve the International Journal of Computer Applications 147(10):49-57.
problem is provided for more clarification. Pandian P, Natarajan G, Akilbasha A (2016) "Fully Rough Integer
Interval Transportation Problems". International Journal of Pharmacy
and Technology 8(2):13866-13876.
Pandian P, Natarajan G, Akilbasha A (2018) Fuzzy interval integer
CONFLICT OF INTERESTS transportation problems. International Journal of Pure and Applied
Mathematics 119 (9):133-142.
Pawlak Z, Skowron A (2007). Rudiment of rough sets, Information
The authors have not declared any conflict of interests.
Sciences 177(1):3-27.
Rani D, Gulati T, Garg H (2016) Multi-objective non-linear programming
problem in intuitionistic fuzzy environment: Optimistic and pessimistic
REFERENCES view point. Expert Systems with Applications 64(1):228-238.
Rebolledo M (2006). Rough intervals-enhancing intervals for qualitative
Atteya T (2016) Rough multiple objective programming. European modeling of technical systems Artificial Intelligence 170(8-9):667-685.
Journal of the Operational Research 248(1):204-210. Shaocheng T (1994). Interval number and fuzzy number linear
Bazaraa M, Jarvis J, Sherali H (2010). Linear programming and network programming, Fuzzy Sets and Systems 66(3):301-306.
flows. John Wiley and Son, New York. Taha H (1997). "Operation Research-An Introduction", 6th Edition. Mac
Bellman R, Zadeh L (1970). Decision making in fuzzy environment. Milan Publishing Co, New York.
Management science 17(4):B-141.
Chinneck J, Ramadan K (2000). Linear programming with interval
coefficients. Journal of the Operational Research Society 51(2):209-
220.
De SK (2019). Triangular dense fuzzy lock sets. Soft Computing
22(21):7243-7254. DOI 10.1007/s00500-017-2726-0.
De SK, Mahata G (2019a). A cloudy fuzzy economic order quantity
model for imperfect‑quality items with allowable proportionate
discounts. Journal of Industrial Engineering International 15(4):571-
583. https://doi.org/10.1007/s40092-019-0310-1.
De SK, Mahata G (2019b). A comprehensive study of an economic
order quantity model under fuzzy monsoon demand. Sadhana
44(4):89. https://doi.org/10.1007/s12046-019-1059-3.
Garg H (2015). Multi-objective optimization problem of system reliability
under intuitionistic fuzzy set environment using Cuckoo Search
algorithm. Journal of intelligent and fuzzy systems 29(4):1653-1669.
Garg H (2018). Some arithmetic operations on the generalized
sigmoidal fuzzy numbers and its application. Granular computing
3(1):9-25.