Received 8 July 2023, accepted 31 July 2023, date of publication 7 August 2023, date of current version 16 August 2023.

Digital Object Identifier 10.1109/ACCESS.2023.3303204

Efficient Energy Management for Household:

Optimization-Based Integration of Distributed
Energy Resources in Smart Grid
QAMAR AKHTER 1 , ABUBAKAR SIDDIQUE2 , SALMAN A. ALQAHTANI 3 ,(Member, IEEE),
ANZAR MAHMOOD 4 , (Senior Member, IEEE), MEHBOOB ALAM 1 , (Member, IEEE),
ZOHAIB MUSHTAQ 5 , MUHAMMAD FARRUKH QURESHI 6 , WASEEM ASLAM5 ,
AND PARANAV KUMAR PATHAK7
1 Department of Electrical Engineering, University of Poonch Rawalakot, Rawalakot 12350, Pakistan
2 Department of Electrical Engineering, Khwaja Fareed University of Engineering and Information Technology (KFUEIT), Rahim Yar Khan 64200, Pakistan
3 Department of Computer Engineering, College of Computer and Information Sciences, King Saud University, Riyadh 11543, Saudi Arabia
4 Department of Electrical Engineering, Mirpur University of Science and Technology (MUST), Mirpur 10250, Pakistan
5 Department of Electrical Engineering, College of Engineering and Technology, University of Sargodha, Sargodha 40100, Pakistan
6 Department of Biomedical Engineering, Riphah International University, Islamabad 44000, Pakistan
7 School of Continuing, McGill University, Montreal, QC H3A 3R1, Canada

Corresponding author: Abubakar Siddique ([email protected])

This work was supported by the Research Supporting, King Saud University, Riyadh, Saudi Arabia, under Project RSPD2023R585.

ABSTRACT Energy demand is increasing globally due to the growing human population and progressive
lifestyle. The adequate use of available energy resources, including renewable, contributes to a country’s
economic sustainability and future development. Optimization–based energy management and cost min-
imization plays a significant role in overcoming energy crises in less developed countries. In this paper,
an optimization–based dynamic energy management technique for smart grids is developed based on the
integration of available renewable resources and variable consumer demand, distinctive to underdeveloped
countries. Consumer demand is classified into fixed, flexible, and highly variable based on population
characteristics. In this work, we developed a Dynamic Multiple Knapsack DMKNS algorithm, which
automatically schedules energy provision to various users by optimally accounting for the available resources
(Grid and Renewable). The proposed method provides a low-cost solution by maintaining a constant energy
supply while preserving consumer comfort and grid stability. The simulation results with various intermittent
availability of resources using MKNS show a saving up to 50% for a variable energy demand user. The
proposed method is general and can also be applied to various underdeveloped regions with similar consumer
demand and statistics.

INDEX TERMS Optimization-based energy management, dynamic multiple knapsack, energy storage,
renewable energy system, smart grid.

I. INTRODUCTION environment. Smart grid provides energy resource integration

Over the last few decades, energy crisis and global warm- and management with optimal distribution and monitoring
ing have motivated the use and development of alterna- using two-way information exchange [2], [3]. In many under-
tive, sustainable, and clean energy sources [1]. The use of developed countries, the use of renewable energy sources
these resources reduces the emission of carbon monoxide and their management and integration with the smart grid
and other poisonous gases, which significantly improves the is still in the early stages of its development [4]. The
lack of technology and planning in these countries causes
The associate editor coordinating the review of this manuscript and under-utilization and sometimes wastage of these resources,
approving it for publication was Wanqing Zhao . leading to an energy crisis and excessive usage of fos-
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.
85716 For more information, see https://creativecommons.org/licenses/by-nc-nd/4.0/ VOLUME 11, 2023
Q. Akhter et al.: Efficient Energy Management for Household: Optimization-Based Integration

TABLE 1. User classification. TABLE 3. Comparison of proposed study with existing studies.

TABLE 2. Comparison of Traditional Grid and Smart Grid.

This highlights the importance of a dynamic pricing scheme

to address the continuous growth of electricity demand. Fol-
sil fuels. In addition, the exponential population growth in lowing subsections elaborate on real-time pricing algorithms
these underdeveloped countries also significantly increases and energy optimization techniques in smart grid.
the demand-supply gap. Recent research of forecasting con- The paper is divided into five Sections. Section II reviews
sumer’s electricity also predicts energy growth across dif- the literature review. Materials and methods are presented
ferent sectors with significant contributions coming from in Section III. Results are discussed in Section IV, a brief
renewable energy sources [5]. Contrary to the traditional discussion is made in Section V and the paper is finally
power grid, the smart grid allows demand-supply interaction concluded in Section VI.
by sharing utility and consumer data, thereby providing opti-
mal cost-effective solutions [6]. A comparison of inherent II. LITERATURE REVIEW
features between traditional and smart grids indicating overall Optimization techniques in energy management aim to the
operational efficiency and optimization is given in Table 2. analysis of available consumer and utility data to make an
Globally, the increasing energy demand and excessive use informed decision, providing cost-effective solutions. Sev-
of non-renewable sources are of serious concern for under- eral optimization techniques are developed over a period
developed countries [7]. The main sources of energy are of time to maximize usage of power assets with necessary
dominated by traditional power systems, which are central- renewable resources integration [12], [13], [14], [15]. Differ-
ized and mainly dependent on petroleum, natural gas, and ent techniques have already been implemented with various
coal-based power plants [8]. This results in extensive use advantages and limitations.
of fossil fuels for power generation, leading to a significant In [16] a novel energy management scheme is presented
increase in CO2 emission. Renewable resources are widely in which an efficient model for smart home infrastructure is
used across the world to overcome energy demand and the developed to minimize user electricity bills, optimize energy
use of fossil fuels. However, the intermittent nature of these usage with the integration of renewable energy resources, and
resources in the form of biomass, waste, geothermal, hydro, maximize user comfort. Simulation results showed that elec-
solar, and wind often present integration challenges lead- tricity bill is reduced with peak power management. A genetic
ing to severe under-utilization [9]. Consequently, this leads algorithm was used and to validate results PSO and ASO
to reliability issues, which affects the grid’s sustainability can also be used. In [17] a fast randomized first-randomized
and resilience. In such cases, the smart grid responds by first-order optimization system is offered for dynamic energy
demand side management (DSM), which includes action management. An ideal forecast problem was formulated by
taken to plan, implement, and monitor predefined activities reducing the compeer’s costs of the allied power utility, as a
that affect the consumer’s electricity utilization patterns [10]. huge numeral of user’s manageable devices arena compassed
The adopted strategy in response to consumer data by DSM in the system. This model validated that existing energy
is usually a dynamic pricing pattern adapted by the smart grid resources can provide benefits for the economic operation of
to overcome and reduce the effect of intermittent demand for the power grid.
energy. Recent research suggested various pricing schemes In [18] a single-leader-multiple-follower Stackelberg game
to serve demand-side management expectations to deliver (SLMFSG) among the auctioneer and residential users.
consumer-based cost-effective solutions [11]. The notable An improved auction-based dual energy storage scheme is
pricing schemes based on usage patterns and estimations are modeled among a number of domestic units and share facility
Real Time Pricing (RTP), Time of Use (TOU), and Critical controllers in a smart grid. In [19] power management and
Peak Pricing (CPP). These pricing schemes are real-time mechanism algorithms of microgrid by energy storage are
and based on demand response, with usually an estimated conferred. In [20] a price-based demand response scheme
increase of 1.2% in electricity supply at 1% GDP growth. was designed and implemented. Sequences of undeviating

VOLUME 11, 2023 85717

 Q. Akhter et al.: Efficient Energy Management for Household: Optimization-Based Integration

estimate models for the load forecast were presented, such as a storage system in overcoming the intermittent nature of
standard autoregressive (AR) and time-varying AR (TVAR). renewable energy resources adds much-needed stability to the
A mutual energy storage and load scheduling with renewable grid. A study shows that a grid, with a storage capability of
incorporation for real-time residential side energy manage- 10% to 1% of its generating capacity is considered to be more
ment using the Lyapunov optimization technique is pre- reliable and stable [28]. In [31] a critical look at dynamic
sented [21]. multi-dimensional knapsack problem generation is presented.
Similarly, a home energy management scheme using a Similarly, the dynamic programming approach to solve
genetic algorithm is presented in [22]. Particle Swarm Opti- real-world applications of the multi-Objective unbounded
mization (PSO) is a cloud-built optimization method, which knapsack problem is presented in [32].
is used for minimizing energy cost with the contribution of
consumer’s demand [23]. Contrary to the genetic algorithm, III. MATERIALS AND METHODS
the PSO technique contains personal best as well as global The algorithm is based on solving Multiple Knapsack Prob-
best to get an overall finest solution. However, it primarily lem (MKP), which dynamically minimizes the piece-wise
works with load-shifting techniques, so cannot be imple- cost function. DMKNS adds an additional constraint to the
mented without constraints. Recently, the Knapsack method Knapsack problem, which automatically regulates energy
of optimization is used to solve combinatorial problems [24]. consumption and cost variations. The MKP is the generaliza-
It is applied, where an economical and productive solution tion of a single Knapsack Problem, where more constraints
to the scheduling problems is required. Experimental results are added in the form of multiple objects to find the optimal
demonstrated that the Knapsack method is effective in mini- solution.
mizing and bringing regularity to the users fluctuating energy In this work, we have developed a Dynamic Multiple
pattern [26]. However, the method is cost-centered and often KnapSack (DMKNS) algorithm. The major contributions in
ignores user satisfaction. The optimization techniques alone this work are as follows:
can have limited effect, however, an effective smart grid sys- • Developed a Dynamic Multiple KnapSack algorithm,
tem deployment needs resource integration and their optimal which automatically schedules the provision of energy
utilization. The need for integration of an energy storage sys- to various users by optimally accounting for the avail-
tem adds flexibility to the system, however, it also introduces able energy resources (Grid and Renewable).
new challenges by adding complexity of management and • The proposed DMKNS algorithm provides a low-cost
control [25]. solution by maintaining a constant energy supply and
A closed-loop pricing algorithm addressing the random at the same time preserving consumer comfort and grid
demand of customers uses the deviation in prices caused by stability.
the variation between actual and desirable load [23]. The • The intermittent availability of resources and energy
algorithm gains the advantage of both open and closed-loop demand is met by introducing piece-wise pricing, which
schemes, however, it heavily relies on constant consumer dynamically provides a low-cost demand-specific solu-
feedback to adjust real-time pricing response. A recent energy tion.
management scheme based on an efficient model for smart • Simulation results of dynamic energy management
home infrastructure minimizes the user bill and optimizes scheduling using DMKNS shows a saving up to 50%
usage with the integration of renewable energy resources [6]. for a variable energy demand user.
However, the scheme is primarily residential-based, with very • The proposed method is general and can also be applied
restricted control and management. A new concept of micro- to various underdeveloped regions with similar con-
grid with energy storage is introduced to prioritize renewable sumer demand and statistics.
energy resources by providing power to the load and effi-
ciently managing battery discharging and overcharging [27]. The next sub-sections, briefly discuss different classes of
However, excessive use of battery power and storage renders users. Note that in DMKNS, the users are classified based
it unfeasible for large-scale commercial use. on their energy consumption, ranging from fixed to variable
The energy generation in most of underdeveloped demand.
countries predominantly uses fossil fuel, which is not
environment-friendly [29], [30]. In addition, the increase A. CLASS A (FIXED DEMAND)
in population in some of these countries has compelled the The Class A consists of two users U1 and U2 . They have fixed
energy sector to make use of alternate sources. These sources energy demand and may have the same or different power
of energy are mostly pollution-free and environment-friendly, ratings. The energy consumption is fixed and a day-a-head
however, due to their intermittent nature and lack of energy demand needs to be provided to the grid. In return, the utility
storage technologies, their efficiency is quite limited. The provides energy at the lowest possible rate using piece-wise
installation of any adopted Energy storage system (ESS) is pricing.
usually very costly, requiring regular scheduled maintenance. Some of the important abbreviations used in this Section
However, despite the overwhelming cost, the advantage of are listed below.

85718 VOLUME 11, 2023

Q. Akhter et al.: Efficient Energy Management for Household: Optimization-Based Integration

FIGURE 1. Schematic of the proposed DMKs scheme.

TABLE 4. Demand of fixed users U1 ,U2 ,Uf3 and Uf4 . demand. The utility will give an option in case of variable
energy demand to switch to fixed energy consumption to
obtain the lowest rate. However, in case the users disagree,
the energy will be provided at higher rates.

C. CLASS C (VARIABLE DEMAND)

The Class C also consists of two users Uv5 and Uv6 . Unlike
ClassA and ClassB, these users have unpredictable and time-
varying demand. As long as the demand is less than the
supply, the utility will continue to provide energy from the
grid at a slightly higher rate due to variable demand. However,
when the demand is about to exceed the maximum capacity
of the grid, the energy supply from the grid is automatically
cut off. The energy is then provided by the renewable energy
source integrated with the storage system. In this case, it is
assumed that renewable sources attached to the battery stor-
age system have enough energy to meet the excessive demand
of the Class C. The Schematic of the proposed scheme is
shown in Figure 1.
Figure 1 provides a visual representation of the proposed
energy management scheme with a brief representation of
multiple energy users, energy storage systems, renewable
energy systems, and mainly the control unit which is automat-
B. CLASS B (FLEXIBLE DEMAND) ically scheduling the overall system.Consumers are divided
The Class B again consists of two users U3 and U4 . They into three different classes, Class A,Class B and Class C
f f
are further classified according to their demand as U3 , U4 along with multiple users. Class A has two users U1 and
and Uv3 , Uv4 . Note that the demand in this class is flexible, U2 and their energy consumption is fixed (i.e.) the utility
f f
with U3 and U4 showing the behavior of U3 and U4 in has the knowledge of their energy demand day ahead. The
fixed demand. Similarly, the Uv3 and Uv4 show their variable demand of these users can vary in magnitude but will remain
demand behavior. The utility will provide energy at the lowest fixed in a specific time slot. The utility will provide energy
possible rate, once responding to U3 and U4 fixed energy to these users at the lowest possible rate as a reward. Class

VOLUME 11, 2023 85719

 Q. Akhter et al.: Efficient Energy Management for Household: Optimization-Based Integration

B has two users U3 and U4 and their energy consumption is Note that ‘‘m′′ defines the maximum number of time slots,
flexible. The utility will provide energy at the lowest possible which is incremented hourly. The total energy consumption in
rate, once responding to U3 and U4 fixed energy demand. The the proposed method is constrained by the maximum capacity
utility will give an option in case of variable energy demand (Cj ) of the grid and can be mathematically expressed as:
to switch to fixed energy consumption to obtain the lowest m
rate. However, in case the users disagree, the energy will be
X
Ei Yi − j ≤ Cj (4)
provided at higher rates. Similarly Class C also consists of j=1
two users Uv5 and Uv6 . Unlike ClassA and ClassB, these users
have unpredictable and time-varying demand. As long as the The main focus of the DMKs algorithm is to meet the
demand is less than the supply, the utility will continue to unpredictable and time-varying demand of Class C users.
provide energy from the grid at a slightly higher rate due In order to accommodate these demands, let Ejt be the total
to variable demand. However, when the demand is about energy consumed by Uv5 and Uv6 users, when their total
to exceed the maximum capacity of the grid, the energy demand becomes greater than the maximum capacity (Cj ).
supply from the grid is automatically cut off. A renewable In this case, the capacity constraints change to:
energy storage system is shown in Figure 1. The energy is m
X
then provided by this renewable energy source. Similarly, Ei Yi − j ≤ Cj − Ejt (5)
in case of unavailability of energy from renewable energy j=1
systems, an energy storage system is integrated with the In order to account for the excess energy generated by
renewable energy system. A control center is also present to renewable energy, let Ejr represent the excess renewable
automatically schedule the variable energy demand by using energy. Note that the constraint equation is further modified
an evolutionary algorithm. to account for these changes and can be written as:
In the mathematical preliminaries, we will first define
m
various variables used in user scheduling in DMKs: X
Ei Yi − j ≤ Cj − Ejt + Ejr (6)
• The number of users is defined by ‘‘n’’ which corre-
j=1
sponds to the number of objects in DMKs.
• The time slots variable ‘‘j’’ relates to the Knapsacks, It is observed that (6) account for the case when the Ejr
with ‘‘m’’ as its maximum value. The prices of electricity exceeds the demand and is returned back to the grid to impact
in ‘‘j’’ slots are user dependent and can be fixed and the overall cost of the scheme. The parameters used in the
variable. mathematical modeling and analysis of the proposed DMKs
• The total time required by any of the users to complete scheme are summarized in Table 1.
req The derived mathematical expressions defining the scheme
its task (i.e. 24 hours) is given by Tj .
• The cost function ‘‘Costj ’’ is a piece-wise function that along with these parameters are used in the next section for
shows the cost of energy consumed in a given time slot the simulation and results.
‘‘j’’. The system total capacity is represented by Cj .
• The amount of energy consumed by any user is denoted IV. RESULTS
by Ei , which corresponds to the weight of the Knapsack. In this section, we present simulations and performance eval-
Whereas, ‘‘i’’ accounts for the time slots of energy con- uation results of the proposed DMKs scheme with additional
sumed. constraints of energy consumption and average cost under
various availability conditions of energy sources and time
In order to represent the energy usage in a given time slot,
slots. We are using the pie-wise concept of pricing in mul-
let us define a Boolean integer variable Yij as:
( tiple time slots. It is important to note that the need and
1, Energy consumption in time slot j selection of multiple sacks are to account for variable prices
Yij = (1) in different time slots, which adds much-needed flexibility
0, No energy consumption in time slot j
to a single knapsack and accommodates user comfort. The
The linear objective function for cost minimization can proposed scheme is valid for large-scale integration, however,
now be defined as without loss of generality, we consider only three classes with
Xn Xm a maximum of two users in each class.
min Ei Yij Costj (2)
i=1 j=1 A. DEMAND OF Class A (FIXED), B (FLEXIBLE) AND C
(VARIABLE)
The integer variable Yij defining the energy usage is The first classification in the simulation and analysis is the
related to the amount of total energy consumed by each user simple case of energy consumption by the fixed demand
as user i.e. Class A (U1 and U2 ). The grid provides them with
m energy at the lowest possible rate due to fixed demand. The
req
X
Yij = Tj (3) demand data spread over the entire day, charged at a fixed
j=1 rate of$2/kW is shown in 4. Note that the fixed demand of the

85720 VOLUME 11, 2023

Q. Akhter et al.: Efficient Energy Management for Household: Optimization-Based Integration

FIGURE 2. Energy fluctuation of U1 and U2 in terms of units consumed (KWh).

FIGURE 3. Energy fluctuation of Uf3 and Uf4 in terms of units consumed (KWh).

users follows a random demand model, where each temporal to their energy usage. In this case, the utility will provide
slot is uniformly distributed from 1:00 am to 12:00 am all- energy comparatively at higher rates (i.e. 5/KW ) than the
inclusive. fixed energy users. The data of units consumed is plotted
The energy usage of U1 and U2 giving details of random in Figure 4.
fluctuation in consumption is shown in Figure 2.
The user data along with its consumption shows an ideal
case, where fixed demand places a predictable load on the B. DEMAND OF Class C (VARIABLE USERS)
grid allowing low prices at rates for the users. The next is the variable class users Uv5 and Uv6 . Note that
Next are the Class B users, which are also known as flex- the maximum energy allowed to all users in this class from
f f
ible users U3 /Uv3 and U4 /Uv4 . They can switch their demand the grid is 15 kW. In case of excessive energy consumption
between fixed and variable. Note that the behavior of a fixed (> 15 kW) by the users, the same can be drawn from the
demand user is already discussed in the preceding paragraph renewable energy source and energy storage system. In addi-
in detail and also shown in Figure 3. tion, note that the excess energy during the day is provided
So without loss of generality, we will discuss variable by renewable sources, and the same is drawn from the storage
users behaviour only i.e. Uv3 and Uv4 . The demand of system at night. The demand of variable users (Uv5 and Uv6 ) is
these users is shown in Table 5. These users have variable shown in Table 6.
energy demands throughout the day with added flexibility Energy consumption of of Uv5 and Uv6 is shown in Figure 5.

VOLUME 11, 2023 85721

 Q. Akhter et al.: Efficient Energy Management for Household: Optimization-Based Integration

FIGURE 4. Energy fluctuation of Uv3 and Uv4 in terms of units consumed (KWh).

FIGURE 5. Energy fluctuation of Uv5 and Uv6 in terms of units consumed (KWh).

Similarly, the total energy consumption of all the users is as the demand of any of the variable users is beyond the
shown in Figure 6. maximum capacity of the grid, RES, and ESS will schedule
The plotted results show that the majority of the variable themselves to provide excess energy. A brief overview of
users have an energy requirement greater than the total capac- energy consumption of Uv5 and Uv6 from renewable and energy
ity of the grid. However, in some cases, the demand of these storage systems is given in Table 7.
users is almost double the grid capacity (i.e. 30 kw for Uv6
).Therefore, an alternative is needed to overcome this problem
in the form of RES and EES. D. BILL CALCULATION FOR Uv5 AND Uv6 WITH AND
WITHOUT RES AND ESS
C. ENERGY CONSUMPTION OF Uv5 AND Uv6 FROM In the bill calculations, the DMKs consider both of these
RENEWABLE AND ENERGY STORAGE SYSTEM variable users (Uv5 and Uv6 ), and demand for electricity
In this sub-section role of RES and ESS is briefly discussed. may exceed the maximum capacity of the grid (i.e. 15kW).
As long as the demand of variable users is within range of If demand exceeds the limit, then the extra energy is provided
the maximum capacity of the grid (i.e.15 kw), the grid will by the RES and ESS using automated scheduling. In order
continue to provide energy at the prescribed rates. As soon to observe the significance of RES and ESS integration,

85722 VOLUME 11, 2023

Q. Akhter et al.: Efficient Energy Management for Household: Optimization-Based Integration

FIGURE 6. Total energy consumption of all users in terms of units consumed (KWh).

FIGURE 7. Comparison of bill for Uv5 with and without RES and ESS.

a comparison of Bill for Uv5 with and without RES and ESS availability of resources and demand. In Figure 7 and Figure 8
is shown in Figure 7. it can be clearly observed that in a few time slots, the value of
The results clearly show that for twenty-four hours, the bill the bill becomes negative which means that at that time user
for Uv5 without using RES and ESS is US$ 34.6. However, is not consuming energy from the grid, renewable and energy
when RES and ESS are integrated, the bill is reduced to storage system rather it supplies surplus energy back to the
US$ 22.7. Similarly, the comparison of Bill for Uv6 with and grid.
without RES and ESS is shown in Figure 8. A comparative analysis for both criteria (with and without
Note that over a period of twenty-four hours, the bill of RES and ESS) is illustrated in Table 8, which shows a saving
Uv6 without RES and ESS is US$ 43.2. However, when RES of up to 50% for a variable user. In Table 8 Bill/day (US
and ESS are integrated with the system, the bill reduces to dollars) is presented for all the users. For U1 and U2 , the cost
US$ 15.9. The comparison of the two users i.e. Uv5 and Uv6 of electricity over a period of 24 hours is 25.5 (US dollars)
f f
in Figure 7 and Figure 8 respectively, clearly shows dynamic and 30.47 (US dollars) respectively. Similarly, U3 and U4
energy management scheduling using DMKs results in pro- have 12.6 (US dollars) and 12.3 (US dollars) respectively.
viding low-cost solution even in the presence of intermittent In order to account for the effect of renewable energy storage

VOLUME 11, 2023 85723

 Q. Akhter et al.: Efficient Energy Management for Household: Optimization-Based Integration

FIGURE 8. Comparison of bill for Uv6 with and without RES and ESS.

TABLE 5. Demand of variable users Uv3 and Uv4 . TABLE 6. Demand of variable users Uv5 and Uv6 .

and energy storage system on the unit price of electricity for and energy storage system on the unit price of electricity for
the highly variable users, it is evident from the results that the highly variable users, it is evident from the results that for
for Uv5 (without RES and ESS) the Bill/day (US dollars) is Uv5 (without RES and ESS) the Bill/day (US dollars) is 34.6
34.6 (US dollars), which is reduced to 22.7 (US dollars) by (US dollars), which is reduced to 22.7 (US dollars) by the
the integrating of RES and ESS. Similarly for Uv6 (without integrating of RES and ESS. Similarly for Uv6 (without RES
RES and ESS) the Bill/day is reduced to 15.9 (US dollars) and ESS) the Bill/day is reduced to 15.9 (with RES and ESS)
from 43.2(without RES and ESS). In Table 8 Bill/day (US (US dollars) from 43.2(without RES and ESS). The impact
dollars) is presented for all the users. For U1 and U2 , the cost of the decrease in per unit price of electricity is reduced to
of electricity over a period of 24 hours is 25.5 (US dollars) 50% by the integration of RES and ESS and managing the
f f
and 30.47 (US dollars) respectively. Similarly, U3 and U4 available resources by optimal scheduling.
have 12.6 (US dollars) and 12.3 (US dollars) respectively. From the above table is clear that adding RES and ESS has
In order to account for the effect of renewable energy storage decreased the per unit price of electricity up to 50%.

85724 VOLUME 11, 2023

Q. Akhter et al.: Efficient Energy Management for Household: Optimization-Based Integration

TABLE 7. Demand of variable users (Uv5 and Uv6 ) from grid,RES and ESS in the maximum demand remains within the maximum grid
kW.
capacity.

VI. CONCLUSION
In this paper, a dynamic energy management technique
is developed based on the integration of available renew-
able resources and variable consumer demand, distinctive
to underdeveloped countries. The work classifies consumer
demand into fixed, flexible, and highly variable based on
population characteristics. The developed optimization-based
Dynamic Multiple KnapSack algorithm automatically sched-
ules the provision of energy to various users by optimally
accounting for the available energy resources. The pro-
posed method provides a low-cost solution by maintaining
a constant energy supply and at the same time preserves
consumer comfort and grid stability. The final simulation
results of dynamic energy management scheduling using the
optimization-based Dynamic Multiple KnapSack algorithm
shows a saving of up to 50% for a variable energy demand
user. The proposed approach uses a method of integration,
which is general and can also be applied to regions with
similar consumer demand and statistics.

A. LIMITATIONS AND FUTURE WORK

There are a few limitations, which may generate future ideas
and work. The optimization and practical implementation of
TABLE 8. Comparative analysis (with and without RES and ESS).
DMKNS algorithm may require significant initial resources
and infrastructure upgrades to support the integration of
DERs and AI-based energy management. Similarly, the effec-
tiveness of the system to some extent relies on weather con-
ditions and other environmental factors, which may degrade
the overall performance and affect the system’s stability and
reliability.
In future work, the following is suggested.
• The system can be further improved by integrating
advanced AI techniques, such as deep learning or rein-
forcement learning, to enhance the accuracy and effec-
V. DISCUSSION
tiveness of energy management.
The integration of RES and ESS has a dominant effect on
• The system can be extended to support more complex
the price of electricity consumed as well as the availability
energy management scenarios, such as multi-household
of electricity during peak hours. Similarly, the comfort of
or community-level energy management, to achieve
variable users is also maintained without imposing a penalty
greater energy efficiency and cost savings.
during peak hours. Thus DMKNs proved to be an efficient
algorithm with multiple constraints.
Another consequence of using this technique is that the NOMENCLATURE
Cj Total capacity of grid.
grid stability will not be disturbed in case of variation in Costj Cost function.
the load on a large scale as the load will be automatically DMKNS Dynamic Multiple Knapsack.
scheduled using maximum grid capacity as well as RES Ei Total Energy consumed by any user from grid.
and ESS.It can be observed from the results that the per Ejr Excess renewable energy.
unit price of electricity has reduced to a great extent with
Ejt Total energy consumption by Uv5 and Uv6 .
the integration of RES and ESS.However, the constraints
related to ESS are not discussed in detail, especially the EES Energy storage system.
capital cost and storage area for the ESS, which can be easily j Time slot.
managed when this algorithm is implemented on a large kW Kilowatt.
commercial scale. The grid will continue to supply energy RES Renewable energy storage.
req
to various users on a piece-wise pricing concept as long as Tj Total time.

VOLUME 11, 2023 85725

 Q. Akhter et al.: Efficient Energy Management for Household: Optimization-Based Integration

REFERENCES [22] K. E. Nygard, S. Bou Ghosn, Md. M. Chowdhury, D. Loegering,

R. McCulloch, and P. Ranganathan, ‘‘Optimization models for energy
[1] M. Umar, X. Ji, D. Kirikkaleli, and A. A. Alola, ‘‘The imperative-
reallocation in a smart grid,’’ in Proc. IEEE Conf. Comput. Commun.
ness of environmental quality in the United States transportation sector
Workshops (INFOCOM WKSHPS), Apr. 2011, pp. 186–190.
amidst biomass-fossil energy consumption and growth,’’ J. Cleaner Prod.,
vol. 285, Feb. 2021, Art. no. 124863. [23] T. Li and M. Dong, ‘‘Real-time residential-side joint energy storage man-
agement and load scheduling with renewable integration,’’ IEEE Trans.
[2] Z. Zhu, S. Lambotharan, W. H. Chin, and Z. Fan, ‘‘Overview of demand
Smart Grid, vol. 9, no. 1, pp. 283–298, Jan. 2018.
management in smart grid and enabling wireless communication technolo-
[24] S. Rahim and R. Maqsood, ‘‘A novel energy management approach in
gies,’’ IEEE Wireless Commun., vol. 19, no. 3, pp. 48–56, Jun. 2012.
smart grid challenges and opportunities,’’ in Proc. 4th Int. Conf. Energy,
[3] J. N. Bharothu, M. Sridhar, and R. S. Rao, ‘‘A literature survey report on
Environ. Sustain. Develop., 2016.
smart grid technologies,’’ in Proc. Int. Conf. Smart Electric Grid (ISEG),
[25] N. Vaughn, Renewable Energy and The Environment. Boca Raton, FL,
Sep. 2014, pp. 1–8.
USA: CRC Press, 2009, p. 63.
[4] M. S. Hossain, N. A. Madlool, N. A. Rahim, J. Selvaraj, A. K. [26] U. Shahzad, ‘‘The need for renewable energy sources,’’ Energy, vol. 2,
Pandey, and A. F. Khan, ‘‘Role of smart grid in renewable energy: pp. 16–18, Jan. 2012.
An overview,’’ Renew. Sustain. Energy Rev., vol. 60, pp. 1168–1184,
[27] R. Zamora and A. K. Srivastava, ‘‘Energy management and control algo-
Jul. 2016.
rithms for integration of energy storage within microgrid,’’ in Proc. IEEE
[5] U. Perwez, A. Sohail, S. F. Hassan, and U. Zia, ‘‘The long-term forecast 23rd Int. Symp. Ind. Electron. (ISIE), Jun. 2014, pp. 1805–1810.
of Pakistan’s electricity supply and demand: An application of long range [28] P. Breeze, Power Generation Technologies. Newnes, 2019.
energy alternatives planning,’’ Energy, vol. 93, pp. 2423–2435, Dec. 2015. [29] I. Hanif, B. Aziz, and I. S. Chaudhry, ‘‘Carbon emissions across the spec-
[6] J. He, C. Zhao, L. Cai, P. Cheng, and L. Shi, ‘‘Practical closed-loop trum of renewable and nonrenewable energy use in developing economies
dynamic pricing in smart grid for supply and demand balancing,’’ Auto- of Asia,’’ Renew. Energy, vol. 143, pp. 586–595, Dec. 2019.
matica, vol. 89, pp. 92–102, Mar. 2018. [30] N. Singh, R. Nyuur, and B. Richmond, ‘‘Renewable energy development
[7] M. Aydin, ‘‘Renewable and non-renewable electricity consumption- as a driver of economic growth: Evidence from multivariate panel data
economic growth nexus: Evidence from OECD countries,’’ Renew. Energy, analysis,’’ Sustainability, vol. 11, no. 8, p. 2418, Apr. 2019.
vol. 136, pp. 599–606, Jun. 2019. [31] S. Uyar and H. T. Uyar, ‘‘A critical look at dynamic multi-dimensional
[8] Y. S. Mohammed, M. W. Mustafa, N. Bashir, and A. S. Mokhtar, ‘‘Renew- knapsack problem generation,’’ in Applications of Evolutionary Com-
able energy resources for distributed power generation in Nigeria: A review puting (Applications of Evolutionary Computing), Tübingen, Germany:
of the potential,’’ Renew. Sustain. Energy Rev., vol. 22, pp. 257–268, Springer, 2009, pp. 762–767.
Jun. 2013. [32] A. P. Khandekar and A. Nargundkar, ‘‘Dynamic programming approach to
[9] T. T. D. Tran and A. D. Smith, ‘‘fEvaluation of renewable energy tech- solve real-world application of multi-objective unbounded knapsack prob-
nologies and their potential for technical integration and cost-effective lem,’’ in Intelligent Systems and Applications. Berlin, Germany: Springer,
use within the U.S. energy sector,’’ Renew. Sustain. Energy Rev., vol. 80, 2023, pp. 417–422.
pp. 1372–1388, Dec. 2017.
[10] E. Sarker, P. Halder, M. Seyedmahmoudian, E. Jamei, B. Horan,
S. Mekhilef, and A. Stojcevski, ‘‘Progress on the demand side management
in smart grid and optimization approaches,’’ Int. J. Energy Res., vol. 45,
no. 1, pp. 36–64, Jan. 2021.
[11] A. Aoun, M. Ghandour, A. Ilinca, and H. Ibrahim, ‘‘Demand-side manage-
ment,’’ in Hybrid Renewable Energy Systems and Microgrids. Amsterdam,
The Netherlands: Elsevier, 2021, pp. 463–490. QAMAR AKHTER received the B.S. degree
[12] C. Bharathi, D. Rekha, and V. Vijayakumar, ‘‘Genetic algorithm based (Hons.) from the Mirpur University of Science and
demand side management for smart grid,’’ Wireless Pers. Commun., Technology (MUST), Mirpur, Pakistan, and the
vol. 93, no. 2, pp. 481–502, Mar. 2017. M.Sc. degree in electrical engineering with a spe-
[13] B. Hegerty, C.-C. Hung, and K. Kasprak, ‘‘A comparative study on differ- cialization in power systems. He is currently pur-
ential evolution and genetic algorithms for some combinatorial problems,’’ suing the Ph.D. degree in biomedical engineering
in Proc. Mexican Int. Conf. Artif. Intell., vol. 9, 2009, p. 13. with Riphah International University, Islamabad,
[14] N. Kumaraguruparan, H. Sivaramakrishnan, and S. S. Sapatnekar, ‘‘Resi- Pakistan. He is also a Lab Engineer with the
dential task scheduling under dynamic pricing using the multiple knapsack Department of Electrical Engineering, University
method,’’ in Proc. IEEE PES Innov. Smart Grid Technol. (ISGT), Jan. 2012, of Poonch Rawalakot, Pakistan. He was awarded
pp. 1–6. the Gold Medal in F.Sc (Pre-Engineering) from the Board of Intermediate
[15] N. Qayyum, A. Amin, U. Jamil, and A. Mahmood, ‘‘Optimization tech- and Secondary Education Mirpur (BISE) Azad Kashmir, Pakistan, securing
niques for home energy management: A review,’’ in Proc. 2nd Int. Conf. first position.
Comput., Math. Eng. Technol. (iCoMET), Jan. 2019, pp. 1–7.
[16] D. Li and S. K. Jayaweera, ‘‘Uncertainty modeling and price-based demand
response scheme design in smart grid,’’ IEEE Syst. J., vol. 11, no. 3,
pp. 1743–1754, Sep. 2017.
[17] D. Han, W. Sun, and X. Fan, ‘‘Dynamic energy management in smart grid:
A fast randomized first-order optimization algorithm,’’ Int. J. Electr. Power
Energy Syst., vol. 94, pp. 179–187, Jan. 2018.
ABUBAKAR SIDDIQUE was born in Lodhran,
[18] O. Elma, A. Tascşkaraoglu, A. T. Ince, and U. S. Selamogullari, ‘‘Imple-
Pakistan. He received the bachelor’s and mas-
mentation of a dynamic energy management system using real time
ter’s degrees from The Islamia University of
pricing and local renewable energy generation forecasts,’’ Energy, vol. 134,
pp. 206–220, Sep. 2017. Bahawalpur (IUB), Pakistan, and the Ph.D. degree
[19] M. Cirrincione, M. Cossentino, S. Gaglio, V. Hilaire, A. Koukam,
from North China Electric Power University
M. Pucci, L. Sabatucci, and G. Vitale, ‘‘Intelligent energy manage- (NCEPU), Beijing, China, in 2019. He was with
ment system,’’ in Proc. 7th IEEE Int. Conf. Ind. Inform., Jun. 2009, IUB. He is currently an Assistant Professor with
pp. 232–237. the Electrical Engineering Department, Khwaja
[20] N. Batra and D. H. Naidu, ‘‘Research paper on home energy management Fareed University of Engineering and Information
system using genetic algorithm,’’ Int. J. Res. Appl. Sci. Eng. Technol. Technology (KFUEIT), Punjab, Pakistan. He is the
(IJRASET), vol. 5. author and coauthor of more than 20 publications in international journals
[21] M. Kamran, M. Mudassar, I. Abid, M. R. Fazal, S. R. Ahmed, M. I. Abid, and proceedings in the area of power systems and power electronics. His
R. Khalid, and S. H. Anjum, ‘‘Reconsidering the power structure of research interests include power flow control, power electronics, high volt-
Pakistan,’’ Int. J. Renew. Energy Res., vol. 9, pp. 480–492, Mar. 2019. age, and power system stability.

85726 VOLUME 11, 2023

Q. Akhter et al.: Efficient Energy Management for Household: Optimization-Based Integration

SALMAN A. ALQAHTANI (Member, IEEE) is MUHAMMAD FARRUKH QURESHI received

currently a Full Professor with the Department of the B.S. degree from International Islamic Uni-
Computer Engineering, College of Computer and versity, Islamabad, in 2012, and the M.S. degree
Information Sciences, King Saud University. His from Riphah International University, Islamabad,
research interests include radio resource manage- in 2014, where he is currently pursuing the Ph.D.
ment for wireless and cellular networks (4G, 5G, degree. He has been a Senior Lecturer with Riphah
the IoT, industry 4.0, LTE, LTE-advanced, femto- International University, since May 2015. He has
cell, cognitive radio, and cyber sovereignty) with a published numerous works in reputable confer-
focus on call admission control, packet scheduling, ences and journals on signal processing, machine
radio resource sharing, and the quality of service learning, and deep learning. His research interests
guarantees for data services. include applications of machine learning and deep learning techniques in
rehabilitation engineering.

ANZAR MAHMOOD (Senior Member, IEEE)

received the B.Sc. degree in electrical engineer-
ing from The University of Azad Jammu and
Kashmir, in 2005, the M.Eng. degree in nuclear
power from NED University, Karachi, in 2007,
and the Ph.D. degree in electrical engineering from
COMSATS University Islamabad, in 2016. He was
an Assistant Professor with COMSATS Univer-
sity Islamabad and a Senior Design Engineer with
Pakistan Atomic Energy Commission. He is cur-
rently an Associate Professor with the Department of Electrical Engineering, WASEEM ASLAM was born in Muzaffargarh,
Mirpur University of Science and Technology (MUST), Mirpur, Pakistan. Pakistan. He received the B.Sc. degree in elec-
He has published numerous research articles and international conference trical engineering (power) from UCET, in 2011,
proceedings. His research interests include smart grids, optimization and the M.Sc. degree in electrical engineering (power)
machine learning, energy management and load forecasting, renewables, and from The Islamia University of Bahawalpur,
prosumer communities. Punjab, Pakistan, in 2013, and the Ph.D. degree in
electrical engineering (power) from North China
Electric Power University (NCEPU), Beijing,
China, in 2019. He is currently an Assistant Profes-
sor with the Department of Electrical Engineering,
MEHBOOB ALAM (Member, IEEE) received the University of Sargodha (UOS), Sargodha, a major public sector university
B.S. degree (Hons.) from the University of Engi- in Punjab, Pakistan. He is the author and coauthor of many publications
neering and Technology, Lahore, Pakistan, the in international journals and conference proceedings in the area of power
M.Sc. degree in electrical and computer engineer- systems and power electronics. His research interests include power quality
ing from the University of Calgary, Calgary, AB, improvement, renewable energy, smart grids, high voltage, reliability, and
Canada, in 2003, and the Ph.D. degree in electrical power electronics.
and computer engineering from Rice University,
Houston, TX, USA, in 2007. He was a NSERC
Postdoctoral Fellow with the Department of Elec-
trical and Computer Engineering, University of
Toronto, Canada. He is currently an Associate Professor with the Department
of Electrical Engineering, University of Poonch Rawalakot, Pakistan. He was
a recipient of the Informatics Circle of Research Excellence (iCORE) Canada
Postgraduate Scholarship, in 2002 and 2003. He received the Canadian
Strategic Microelectronics Council of ITAC Industrial Collaboration Award
(SMC Award), in 2003.

PARANAV KUMAR PATHAK received the Ph.D.

ZOHAIB MUSHTAQ received the B.Sc. degree degree in AI. He is currently a Renowned
from Islamia University, the M.S. degree from Researcher and an Academician in Montreal,
Government College University, Lahore, and the Canada. With extensive experience in teaching,
Ph.D. degree in electrical engineering from the research, and academia, he has showcased his
National Taiwan University of Science and Tech- expertise across diverse computer science courses.
nology, in 2020. He was with Riphah Inter- He remains dedicated to professional develop-
national University, as an Assistant Professor. ment, staying updated with the latest advance-
He is currently an Assistant Professor of electri- ments in the field. He holds certifications in AI,
cal engineering with the University of Sargodha, IT, and networking, including Cisco and cloud
Sargodha, Pakistan. He has published research in computing. His research interests include artificial intelligence, big data, and
IEEE and other reputable journals. His research interests include neural net- networking.
works, machine learning, deep learning, computer vision, and data science.

VOLUME 11, 2023 85727