Content uploaded by Sajjad Khan
Author content
All content in this area was uploaded by Sajjad Khan on Jan 26, 2023
Content may be subject to copyright.
COMSATS University Islamabad
Multi-Objective Home Energy Management
System with Multi-Class Appliances using
Meta-Heuristic Techniques
A Thesis Presented to
COMSATS University Islamabad
In partial fulfillment
of the requirement for the degree of
MS (Computer Science)
By
Sajjad Khan
CIIT/SP16-RCS-010/ISB
Fall, 2018
ii
Multi-Objective Home Energy Management
System with Multi-Class Appliances using
Meta-Heuristic Techniques
A Post Graduate Thesis submitted to the Department of Computer Science as
partial fulfilment of the requirement for the award of Degree of MS (Computer
Science).
Name Registration Number
Sajjad Khan CIIT/SP16-RCS-010/ISB
Supervisor:
Dr. Nadeem Javaid,
Associate Professor, Department of Computer Science,
COMSATS University Islamabad,
Islamabad, Pakistan
Co-Supervisor:
Dr. Safdar Hussain Bouk,
Research Professor, Department of Information and Communication Engineering,
Daegu Gyeongbuk Institute of Science and Technology, South Korea
iii
Final Approval
This thesis titled
Multi-Objective Home Energy Management System with
Multi-Class Appliances using Meta-Heuristic Techniques
By
Sajjad Khan
CIIT/SP16-RCS-010/ISB
has been approved
For the COMSATS University Islamabad, Islamabad
External Examiner:
Dr. Adnan Sohail
Assistant Professor, Department of Computing and Technology
Iqra University Islamabad
Supervisor:
Dr. Nadeem Javaid
Associate Professor, Department of Computer Science,
COMSATS University Islamabad, Islamabad
Co-Supervisor:
Dr. Safdar Hussain Bouk
Research Professor, Department of Information and Communication Engineering,
Daegu Gyeongbuk Institute of Science and Technology, South Korea
HoD:
Dr. Majid Iqbal Khan
Associate Professor, Department of Computer Science,
COMSATS University Islamabad, Islamabad
iv
Declaration
I Sajjad Khan (Registration No. CIIT/SP16-RCS-010/ISB) hereby declare that I
have produced the work presented in this thesis, during the scheduled period of
study. I also declare that I have not taken any material from any source except
referorred to wherever due that amount of plagiarism is within acceptable range.
If a violation of HEC rules on research has occurred in this thesis, I shall be liable
to punishable action under the plagiarism rules of the HEC.
Date: January 14, 2019
Sajjad Khan
CIIT/SP16-RCS-010/ISB
v
Certificate
It is certified that Sajjad Khan (Registration No. CIIT/SP16-RCS-010/ISB) has
carried out all the work related to this thesis under my supervision at the Depart-
ment of Computer Science, COMSATS University, Islamabad and the work fulfils
the requirement for award of MS degree.
Date: December 28, 2018
Supervisor:
Dr. Nadeem Javaid
Associate Professor, Department of Computer Science
Co-Supervisor:
Dr. Safdar Hussain Bouk
Research Professor, Department of Information
and Communication Engineering
Head of Department:
Dr. Majid Iqbal Khan
Department of Computer Science
vi
DEDICATION
Dedicated
to my loving parents.
vii
ACKNOWLEDGEMENT
First and foremost, I thank Allah (subhana wa taala), the Almighty, for endowing
me the knowledge to complete this thesis. After that, I would like to express my
special thanks of gratitude to my supervisor Dr. Nadeem Javaid for his never
ending support and valuable suggestions in completing this thesis. I would like to
express a special appreciation to my co-supervisor Dr. Safdar Hussain Bouk for
his support and guidance whenever needed.
I would like to express my sincere gratitude to my brother Ayaz Ahmad for funding
my studies at COMSATS University Islamabad. I am truly indebted to him.
I would like to thank my parents for their continuous support throughout my MS
studies. It would be difficult to find adequate words to convey how much I owe
them. I am always grateful to them for their encouragement and support.
Last but not the least, I am profoundly grateful to all the fellow researchers at
ComSens research lab for their endless support.
viii
ABSTRACT
Multi-Objective Home Energy Management System with
Multi-Class Appliances using Meta-Heuristic Techniques
The day to day increase in world’s population is producing a gap between the
demand and supply of electricity. Traditional Grid (TG) with the aging infras-
tructure is unable to address the increasing electricity demand. Installation of new
generation systems is not a good solution to tackle the high demand of electricity.
Smart Grid (SG) enhanced the TG by adopting information and communication
based technological solutions to address the increasing electricity demand. Smart
Home Energy Management System (SHEMS) plays an important role in the effi-
cacy of SG. To get the most out of the existing system, several demand response
schemes have been presented by researchers. These schemes try to schedule the
appliances in such a way that electricity consumption cost and peak-to-average
ratio are minimized along with maximum User Comfort (UC). However, there
exists a trade-off between UC and electricity consumption cost. In this thesis, a
SHEMS is developed to minimize the appliances waiting time and Peak to Av-
erage Ratio (PAR). For appliances waiting time minimization a novel population
based scheme namely UC Maximization (UCM) is developed. UCM schedule the
home appliances in such a way that appliances waiting time is minimized econom-
ically. Furthermore, we developed an Improved Algorithm for PAR Reduction
(IAPR) to enhance the reliability of power grid. To evaluate the effectiveness of
UCM in terms of appliances waiting time reduction, comparison is made with two
well known meta-heuristic techniques namely Flower Pollination Algorithm (FPA)
and Jaya Optimization Algorithm (JOA). Experimental results show that UCM
scheme outperforms FPA by 5.97% and JOA by 53.9%. Moreover, UCM scheme
reduced the electricity consumption cost and PAR by 58% and 56% as compared
to unscheduled scenario. To validate the effectiveness of IAPR in terms of PAR
minimization, comparison is made with the renowned meta-heuristic optimization
schemes namely Strawberry Algorithm (SA) and Salp Swarms Algorithm (SSA).
Experimental results show that IAPR outperforms SSA by 69.4% and SA by 42.7%
in terms of PAR reduction using the critical peak pricing scheme. Moreover, using
the real time pricing scheme IAPR exceeds SSA and SA by 61.67% and 37.77%
respectively .
ix
Journal publications
1 Mahnoor Khan, Nadeem Javaid, Sajjad Khan, Abdullah, Adnan Naseem,
Salman Ahmed, Muhammad Sajid Riaz, Mariam Akbar and Manzoor Ilahi,
”Game Theoretical Demand Response Management and Short-Term Load
Forecasting by Knowledge Based Systems on the basis of Priority Index”,
Electronics 2018, 7(12), 431
x
Conference proceedings
1Sajjad Khan, Zahoor Ali Khan, Nadeem Javaid, Waleed Ahmad, Raza
Abid Abbasi and Hafiz Muhammad Faisal, “On Maximizing User Comfort
using a Novel Meta-Heuristic Technique in Smart Home”, in 33rd Interna-
tional Conference on Advanced Information Networking and Applications
(AINA).
2Sajjad Khan, Nadeem Javaid, Zahoor Ali khan, Sahibzada Muhammad
Shuja, Muhammad Abdullah and Annas Chand, “Energy Efficient Schedul-
ing of Smart Homes”, in 33rd International Conference on Advanced Infor-
mation Networking and Applications (AINA), 2019
3Sajjad Khan, Nadeem Javaid, Annas Chand, Abdul Basit Majeed Khan,
Fahad Rashid and Imran Uddin Afridi, “Electricity load forecasting for each
day of week using deep CNN”, in 33rd International Conference on Advanced
Information Networking and Applications (AINA), 2019.
4 Sahibzada Muhammad Shuja, Nadeem Javaid, Sajjad Khan, Umair Sar-
fraz, Syed Hamza ALi, Muhammad Taha and Tahir Mehmood,“Electricity
Price Prediction by Enhanced Combination of Autoregression Moving Aver-
age and Kernal Extreme Learing Machine”, in 33rd International Conference
on Advanced Information Networking and Applications (AINA), 2019.
5 Sahibzada Muhammad Shuja, Nadeem Javaid, Sajjad Khan, Hina Akmal,
Murtaza Hanif, Qazi Fazalullah and Zain Ahmad Khan, “Efficient Schedul-
ing of Smart Home Appliances for Energy Management by Cost and PAR
Optimization Algorithm in Smart Grid”, in 33rd International Conference
on Advanced Information Networking and Applications (AINA), 2019.
6 Waleed Ahmad, Nadeem Javaid, Sajjad Khan, Maria Zuraiz, Tayyab Awan,
Muhammad Amir and Raza Abid Abbasi, “A New Memory Updation Heuris-
tic Scheme for Energy Management System in Smart Grid”, in 33rd Inter-
national Conference on Advanced Information Networking and Applications
(AINA), 2019.
7 Waleed Ahmad, Nadeem Javaid, Basit Karim, Syed Qasim Jan, Muham-
mad Ali, Raza Abid Abbasi and Sajjad Khan, “Pro Utility Pro Consumer
Comfort Demand Side Management in Smart Grid”, in 33rd International
Conference on Advanced Information Networking and Applications (AINA),
2019.
xi
8 Raza Abid Abbasi, Nadeem Javaid, Amanullah, Sajjad Khan, Hafiz Muham-
mad Faisal and Sajawal Ur Rehman Khan, “Minimizing Daily Electricity
Cost using Bird Chase Scheme with Electricity Management Controller in
a Smart Home”, in 33rd International Conference on Advanced Information
Networking and Applications (AINA), 2019.
9 Raza Abid Abbasi, Nadeem Javaid, Sajjad Khan, Shujat ur Rehman, Aman-
ullah, Rana Muhammad Asif and Waleed Ahmad, “Minimizing Daily Cost
and Maximizing User Comfort using a New Metaheuristic Technique”, in
33rd International Conference on Advanced Information Networking and
Applications (AINA), 2019.
10 Irfan Azam, Abdul Majid, Tanveer Khan, Sajjad Khan, Zahoor Ali Khan,
Umar Qasim, and Nadeem Javaid, “Avoiding Energy Holes in Underwater
Wireless Sensor Networks with Balanced Load Distribution”, 10th IEEE
International Conference on Complex, Intelligent, and Software Intensive
Systems (CISIS-2016), Japan, pp: 341-350.
xii
TABLE OF CONTENTS
Dedication vii
Acknowledgements viii
Abstract ix
Journal publications 92
Conference proceedings 93
List of figures xv
List of tables xvi
List of algorithms xvii
List of symbols xviii
11
1.1 Introduction ............................... 2
1.1.1 Thesis contributions ...................... 4
1.1.2 Thesis organization ....................... 4
25
2.1 Literature review ............................ 6
2.2 Problem statement ........................... 12
313
3.1 Proposed system model: ........................ 14
3.1.1 Appliances categories ...................... 14
3.1.2 Electricity price tariff ...................... 15
3.2 Waiting time objective ......................... 16
3.3 PAR objective ............................. 17
419
4.1 Optimization schemes .......................... 20
4.1.1 Existing schemes ........................ 20
4.1.1.1 FPA .......................... 20
4.1.1.2 Jaya Optimization Algorithm (JOA) ........ 21
4.1.1.3 SA ........................... 21
4.1.1.4 Salp Swarms Algorithm (SSA) ........... 22
4.1.2 Proposed schemes ....................... 22
4.1.2.1 UCM ......................... 22
4.1.2.2 IAPR ......................... 24
xiii
526
5.1 Simulation results and discussions ................... 27
636
6.1 Conclusion ................................ 37
6.2 Future work ............................... 37
738
Appendices 43
.A Detail of appendices .......................... 44
.B Implementation of waiting time objective ............... 45
.C Implementation of PAR reduction objective ............. 66
Journal publications 92
Conference proceedings 93
xiv
LIST OF FIGURES
3.1 System model .............................. 14
3.2 CPP signal ............................... 16
3.3 RTP signal ............................... 16
3.4 Execution pattern of an appliance ................... 17
5.1 Appliances waiting time of UCM ................... 27
5.2 Per hour cost consumption of UCM scheme .............. 28
5.3 Total cost of UCM scheme ....................... 28
5.4 Per hour load consumption of UCM scheme ............. 29
5.5 PAR of UCM scheme .......................... 29
5.6 Total load consumed by all schemes for UCM scheme ........ 30
5.7 PAR reduction using CPP scheme in IAPR .............. 31
5.8 PAR reduction using RTP scheme in IAPR .............. 31
5.9 Total cost using CPP scheme in IAPR ................ 32
5.10 Total cost using RTP scheme in IAPR ................ 32
5.11 Per hour cost using CPP scheme in IAPR .............. 33
5.12 Per hour cost using RTP scheme in IAPR .............. 33
5.13 Per hour load using CPP scheme in IAPR .............. 34
5.14 Per hour load using RTP scheme in IAPR .............. 34
5.15 Waiting time using CPP scheme in IAPR ............... 34
5.16 Waiting time using RTP scheme in IAPR ............... 35
5.17 Total load consumed using CPP and RTP in IAPR ......... 35
xv
List of symbols
tSingle time slot
TTotal time slots in a day
ADI Deferrable Interruptible Appliances
ADN Deferrable Non-interruptible Appliances
AN D Non-Deferrable Appliances
ωStarting time of an appliance by the consumer
σStarting time of an appliance by the scheduler
Ω Finishing time of an appliance
W T Waiting time of appliances
CostUnscheduled Electricity utilization cost before scheduling
CostScheduled Electricity utilization cost after scheduling
CostHElectricity utilization cost of a single time slot
CostTElectricity utilization cost of a single day
ρPower rating of an appliance
ηElectricity pricing signal announced by utility
αON or OFF status of an appliance
P AR Peak to Average Ration
LP eak Maximum load consumed in a single time slot
LAverage Average load consumed in a single day
LoadHLoad consumed in a single time slot
LoadTTotal load consumed in a single day
xviii
Chapter 1
Introduction
1
CHAPTER 1. 1.1. INTRODUCTION
1.1 Introduction
The gap between supply and demand is increased due to the increasing electricity
demand in industrial, residential, and commercial sectors [1]. To cope with the
increasing electricity demand problem, the power generation companies focused
on the installation of new energy generators. However, installing new generation
sources is neither an economical nor environment friendly option. It is due to the
fact that these sources mostly operate by fossil fuels like natural gas, oil and coal.
It is believed that these sources emit carbon dioxide into the Earth’s atmosphere
which causes global warming [2]. An alternate solution to the increasing electricity
demand is the efficacious utilization of the existing power generation resources [3].
This approach is economical and highly recommended, because it avoids the need
to install new generation plants.
Traditional Grid (TG) is unable to solve the increasing electricity demand problem.
To efficaciously utilize the existing resources TG is experiencing numerous chal-
lenges such as centralized generation system, high power losses, aged infrastructure
and wired technology. One of the most important challenge is the reliability and
sustainability of power grid. The increasing energy demand has a high influence
on the grid reliability, as a result power failures or blackout occurs frequently. The
major factors that affect the reliability of TG are the lack of communication and
inefficacious load management capability. Therefore, to make the power system
more reliable and sustainable, the infrastructure of Smart Grid (SG) is introduced.
SG upgrades the in efficient TG to a quick responsive electricity network by in-
corporating Information and Communication Technology (ICT) based solutions
[4]. These ICT based solutions not only enhances the reliability and stability of
power system but also enables the SG to amalgamate Distributed Energy Sources
(DESs). DESs tackle the transmission losses in the SG by producing energy lo-
cally. Moreover, SG amalgamates Advanced Metering Infrastructure (AMI) and
smart appliances in the existing power system. AMI enables the consumers to
learn their real time energy usage along with its consumption cost. The key ad-
vantages of SG over TG are: reliable grid operations, DESs, Demand Response
(DR) strategies, Smart Meter (SM), Demand Side Management (DSM) and Smart
Scheduler (SS), etc. In SG, DSM plays a key role in enhancing the efficacy and
reliability of the power system.
DSM refers to the strategies adopted by the utility companies to persuade con-
sumers to actively participate in efficacious energy utilization. In DSM, consumers
2Thesis by: Sajjad Khan
CHAPTER 1. 1.1. INTRODUCTION
minimize their electricity consumption cost and Peak to Average Ratio (PAR) by
altering the operational pattern of their home appliances from high to low price in-
tervals. For this purpose, incentive is given to the consumers by the utility so that
energy utilization is efficacious. To persuade the consumers for modifying their
energy utilization pattern in distinct hours (i.e., On-peak hours), DSM introduced
two types of DR schemes. These schemes are: i) price base DR schemes: ii) in-
centive base DR schemes. In price based DR schemes, numerous pricing schemes
are used across the world to encourage the consumers to alter their energy utiliza-
tion pattern. The well known pricing schemes are Critical Peak Pricing (CPP),
Inclined Block Rate Pricing (IBRP), day ahead Time of Use Pricing (ToUP) and
Real Time Pricing (RTP) [5]. In the incentive based DR schemes, consumers agree
to minimize their energy utilization in order to get some incentives from the utility
[6]. Incentive to the consumers is given in the form of minimized electricity bill.
In this scheme, home appliances are wirelessly turned-OFF by the utility due to
peak creation.
DR strategies not only solve the high electricity demand problem but also helps the
consumers to lessen their electricity consumption cost by maximizing User Com-
fort (UC). However, these schemes cannot solve the demand and supply problem
without involvement of the consumers. Whereas, the consumers are neither power
analysts nor economists to schedule as well as monitor their energy utilization.
Furthermore, due to lack of information about electricity tariffs, electricity de-
mand may increased. Therefore, the need of a Smart Home Energy Management
System (SHEMS) to intelligently switch the home appliances ON or OFF with
varying electricity demand in SG is inevitable.
Home appliances scheduling problem in SHEMS can be solved using two ap-
proaches: i) predictive energy management, ii) real-time energy management [7].
The former approach requires SHEMS to schedule the home appliances by analyz-
ing the historical energy utilization pattern [8]. The later approach uses real-time
algorithms to shift the operational time electrical appliance to other times in or-
der to reduce their electricity consumption during peak hours. Furthermore, the
real time approach to schedule home appliance is divided into two subcategories:
i) analytic approach and: 2) meta-heuristic approach. The first approach gives
the prominent results, however, this approach is computationally expensive and
converges slowly. Whereas, the second approach is computationally inexpensive
and provides near to optimal results [9].
3Thesis by: Sajjad Khan
CHAPTER 1. 1.1. INTRODUCTION
Our work focuses on the energy management of a Smart Home (SH). The mo-
tivation for PAR minimization in SH is due to two reasons: i) industrial and
commercial users are reluctant to alter their energy utilization pattern: ii) energy
utilization in residential sectors is more than 40% of the total energy. Further-
more, about 65% of electricity consumption can be minimized by SHs and smart
buildings [10]. A SH comprises of several components such as SM, SS, smart elec-
tric appliances and smart storage devices, etc. These appliances are able to react,
in response to the signals received in order to shift or reduce their electricity load.
The SM provides two-way communication between SH and utility using AMI. In
this way, consumers are informed about the varying prices and power failures that
may occurred due to natural disasters or equipment failures.
1.1.1 Thesis contributions
The main contribution of this thesis are following:
•A novel scheme (i.e., UC Maximization (UCM)) to maximize UC is developed
[11].
•An Improved Algorithm for PAR Reduction (IAPR) is developed to minimize
the PAR [12].
•A comparison analysis of UCM, with the well known meta-heuristic tech-
niques is performed for minimizing appliances waiting time within a SH.
•A comparison analysis of renowned meta-heuristic techniques with the pro-
posed IAPR is performed using CPP and RTP pricing schemes.
1.1.2 Thesis organization
In this thesis, we developed a SHEMS to alter the operational pattern of the
home appliances using real time energy management. Price base DR strategies
with meta-heuristic techniques are used to assess the efficacy of our schemes. The
rest of this thesis is organized as follows: Chapter 2is the related work and
problem statement. Chapter 3describes the proposed system model. Existing
and proposed optimization techniques are presented in chapter 4. Simulation
results and discussions are described in chapter 5. Finally, the thesis is concluded
in chapter 6.
4Thesis by: Sajjad Khan
Chapter 2
Literature review and problem statement
5
CHAPTER 2. 2.1. LITERATURE REVIEW
2.1 Literature review
Many DR schemes have been proposed by different authors in the last few years.
These schemes mainly focuses on minimizing PAR and energy consumption cost
while maximizing UC. A brief summary of the well known schemes is given in
Table 2.1.
In [7], a SHEMS is developed to monitor as well as schedule the household elec-
trical appliances in a conventional house. The main objective of this work is to
minimize electricity consumption cost. In this scheme, if the total load on the
power grid exceeds a predefined limit, the new appliances are either shifted to
other time periods or operated using battery power. Batteries are either charged
from the utility in the low price intervals or using Photovoltaic (PV) cells in day
time. Experimental results demonstrate that the scheme reasonably minimizes
electricity consumption cost. However, this work did not consider the various
pricing schemes.
Yang et al. [13] present a scheme using Improved Particle Swarm Optimization
(IPSO) for scheduling home appliances. The scheme mainly focuses on minimiz-
ing cost. It is evident that electricity tariff and objective curve have an inverse
relationship with each other. Experimental results show that the desired objective
curve is achieved. Moreover, this work aims to achieve power system stability. For
this purpose, the high electricity consumption is reduced in high price intervals.
As a result, UC is compromised.
The authors in [14] propose a Mixed Integer Linear Programming (MILP) based
cost minimization and load equalizing scheme for domestic users. In this scheme,
electricity consumer is considered as a prosumer. Prosumer is the producer as well
as the consumer of electricity. The work consider that prosumers have installed
PV cells in their homes for electricity generation. Furthermore, it is assumed that
the SH is directly connected to the utility to fulfill the energy requirement. In this
work, prosumers store energy in batteries when electricity generation exceeds the
user demand. The stored energy is utilized in high price intervals. However, this
work ignored exporting surplus energy to maximize their income.
The work in [15] design a scheme for minimizing cost and PAR. The proposed
scheme use two renowned meta-heuristic techniques namely Strawberry Algorithm
(SA) and Cuckoo Search Optimization Algorithm (CSOA). The work considered
that each home is equipped with an energy production system, i.e., Micro Grid
6Thesis by: Sajjad Khan
CHAPTER 2. 2.1. LITERATURE REVIEW
(MG). MG in the proposed work includes a wind turbine along with PV system.
MG for electricity generation is an intermittent source. This intermittent nature
of MG affects the reliability and stability of the power system. In an effort to make
the power system stable, Energy Storage System (ESS) is also considered in this
work. Experimental results demonstrate that this scheme reasonably minimized
cost and PAR using CSOA and SA. Moreover, CSOA is effective in minimizing
consumption cost and maximizing profit as compare to SA.
The authors in [16] develop a fuzzy based energy management system. In this
work, consumer is involved in changing the status of appliances from ON-state to
OFF-state. The work mainly focuses on minimizing cost along with PAR. The
UC in this scheme is compromised due to increase in the appliances waiting time.
Israr et al. [17] present an optimization function to maximize the UC. The pro-
posed scheme aims to retain an intended environment inside a SH. Genetic Algo-
rithm (GA) and PSO techniques are used to maximize the UC. For this purpose,
the recorded data for a month in an indoor laboratory environment is analyzed.
Experimental results demonstrate that the proposed scheme achieves desired re-
sults. However, this work did not consider electricity consumption cost and PAR.
A hybrid technique for SHEMS is developed to schedule home appliances according
to user preferences [18]. The scheme used Lightening Search Algorithm (LSA) with
Artificial Neural Network (ANN) to compute learning rate along with the best case
values used in the hidden layers of ANN in the SHEMS. It is deliberated in the
simulation results that this scheme minimizes the electricity bill. The efficiency
of the proposed system is further enhanced by incorporating Renewable Energy
Sources (RES). However, PAR alleviation and power system stability are ignored
in this work.
The work in [19] propose a scheme to minimize the load demand using GA. In this
scheme the electricity consumption pattern (of industrial, commercial enterprises
and domestic users) from multiple data-sets is evaluated. Performance of the
scheme is evaluated by examining the energy consumption pattern through GA,
i.e., GA-DSM and DSM without GA. Experimental results demonstrate that the
scheme achieves the desired results, i.e., minimization of electricity consumption.
Moreover, GA-DSM performs better as compared to DSM without GA. However,
in this work, UC and PAR are ignored.
In [20], a hybrid technique using Harmony Search Algorithm (HSA) and Enhanced
7Thesis by: Sajjad Khan
CHAPTER 2. 2.1. LITERATURE REVIEW
Differential Evaluation (EDE) algorithm is developed. The proposed scheme for-
mulate the demand and supply problem as Multiple Knapsack Problem (MKP).
The work incorporates ESS to maximize the UC and alleviate PAR. Experimental
results demonstrate that the hybrid scheme reasonably minimizes the consumer
bill as compare to HSA and EDE. However, PAR reduction using EDE is better
as compared to their proposed technique.
A chance constrained DR scheme is proposed in [21] to schedule electrical ap-
pliances in SHEMS. The scheme uses PSO and two-Point Estimation Method
(2-PEM). Experimental results demonstrate that PSO along-with 2-PEM outper-
forms gradient based PSO with Latin hyper cube sampling in terms of minimizing
computational load. However, this work, ignores PAR and electricity consumption
cost. PAR burdens the utility that has a direct impact on consumer.
Wen Zheng et al. [22] propose SHEMS using reinforcement learning technique to
meet the DR of residential buildings. The scheme adds intelligence to the SHEMS
via a three step Observe, Learn and Adopt (OLA) algorithm. In this technique,
an algorithm learns the electricity consumption pattern of the appliances based
on consumer preferences. Experimental results demonstrate that the proposed
scheme reasonably minimizes the user consumption cost and PAR. However, UC
is compromised due to shifting or delaying the operational time of the electricity
consuming appliances.
The work in [23] present a scheme to tackle the problem of load demand in critical
hours using appliance scheduling and smart charging. A community based schedul-
ing algorithm is proposed to make sure that the electricity consumers are always
connected to electricity source. In this scheme, the consumers electricity load and
associated batteries are merged together. In this work, the authors merged the
RTP and ToUP scheme to shift the electricity load from high price intervals. The
operational time of home appliances is generated using the day-ahead electricity
pricing. Experimental results demonstrate that this scheme achieves minimum
cost and PAR by merging RTP with ToUP. Furthermore, this work also achieves
an acceptable trade-off in UC and cost minimization.
In [24], a scheme is developed for load scheduling and power trading by incor-
porating RES. The scheme uses dynamic programming to shift the appliances to
different time intervals. In this work, the consumers not only produces the electric-
ity from RES but also sells surplus electricity to the utility or neighboring homes.
Experimental results demonstrate that the proposed work reduces electricity cost
8Thesis by: Sajjad Khan
CHAPTER 2. 2.1. LITERATURE REVIEW
of the consumers. It is due to the reason that competing users exist in the neigh-
borhood who are willing to sell their surplus electricity at low price. The work
adopts a game theory-based approach for power trading between consumers and
prosumers.
The authors in [25] propose a scheme to schedule the electric appliances by in-
corporating RES and storage resources. In this scheme, the appliances scheduling
problem is first formulated as Multiple Knapsack Problem (MKP). Then using
meta-heuristic techniques namely GA, Hybrid GA-PSO (HGPSO), Wind Driven
Optimization (WDO), Bacterial Foraging Optimization (BFO) and Binary PSO
(BPSO) a balance schedule of home appliances is generated. Experimental results
demonstrate that HGPSO outperforms GA, WDO, BFO and BPSO in reducing
PAR and energy utilization cost.
The work in [26] minimizes PAR in SG via DR strategies. This work considers
the optimization problem as two stage problem. The consumers try to boost their
pay-offs and the utility tries to maximize its revenue. At the consumer end, elec-
tricity consumption cost is calculated using an iterative algorithm. Whereas, at
the retailer end, the work used Simulated-Annealing based Price Control Algo-
rithm (SAPCA). SAPCA is used to maximize retailer’s profit using RTP scheme.
Experimental results demonstrate that the proposed scheme is beneficial for the
consumers as well as utility. Furthermore, the proposed work reduces the PAR.
Li Dan et al. in [27] develop a hierarchical framework for DSM. The proposed
framework comprises of three layers namely utility layer, DR-aggregator layer and
consumer layer. In this work, the main objective is to reduce the operational
cost of the utility. When the operational cost of the utility reduces, it gives part
of its revenue to DR-aggregator which in turns incentivize its consumers to alter
their energy utilization pattern. To ensure fairness for all the participants, this
work formulates the energy optimization problem as a multi-objective problem.
This problem is then solved using Artificial Immune Algorithm (AIA). It leads
to Pareto optimal set which did not favor any participant. Experimental results
demonstrate that this scheme is beneficial for all the participants, i.e., consumers
save money in the form of minimum electricity bill, utility reduces its operational
cost, and DR-aggregator makes revenue by providing DSM services.
The work in [28] proposed SHEMS using meta heuristic techniques to plan the
operational time of electric appliances. This work uses GA, CSOA and Crow
Search Algorithm (CSA) to minimize PAR, appliances waiting time as well as
energy utilization cost. This work incorporates ESS for stable grid operation. It
9Thesis by: Sajjad Khan
CHAPTER 2. 2.1. LITERATURE REVIEW
is shown in the experimental results that the performance of CSOA in terms of
minimizing PAR, appliances waiting time and electricity cost is better as compare
to GA and CSA . However, this work neglected the operational and maintenance
cost of ESS.
The authors in [29] present a scheme for scheduling home appliance on the basis
of assigning priorities to the home appliances. These priorities vary according to
user preferences. On the basis of the time and device-based priority, an absolute
comfort level of consumers is computed. The SHEMS then schedule the electric
appliance using the absolute comfort level, consumer’s budget and the total power
available to the consumer. GA is used to generate the schedule of the home
appliances based on maximum UC. Experimental results demonstrate that the
proposed scheme reasonably maximizes the UC within a small budget.
Huang et al. in [30] propose DR scheme for peak load reduction. In an effort to
reduce the peak load and fairly allocate solar power for efficient management of
load, this work adopted a multi-agent minority game theory strategy. Supervised
learning techniques are employed to predict the demand of energy. However, this
scheme considered a similar profile and energy consumption pattern for all con-
sumers. Experimental results demonstrate that the scheme reasonably minimizes
the peak load. However, this work is affected by the changing weather conditions.
Farrokhifar et al. in [31] develop a scheme for scheduling home appliances to
reduce consumption cost within a building. The scheme considered optimization
problem as the ILP problem. Simulations results demonstrate that the proposed
scheme reduces consumption cost. However, this work ignores the PAR.
Asif et al. in [32] present a priority induced DSM scheme to efficaciously sched-
ule the home appliances in SHEMS. The main objective of this work is to mini-
mize cost and mitigate rebound peaks. In this work, the authors used renowned
meta-heuristic techniques namely EDE, BPSO, and GA. Furthermore, knapsack
capacity limit is used to enhance the grid reliability. Simulation results shows that
this scheme minimized the consumer bill by 60%. However, this work considered
a limited number of home appliances.
The authors in [33] propose a scheme to maximize the UC in a residential building.
This scheme uses the well known meta heuristic technique namely Bat Algorithm
(BA). Three environmental parameters (i.e., air quality, illumination and temper-
ature) are optimized via BA. Then the difference between the optimized values
and non-optimized values is computed. Afterwards, the difference is then given
10 Thesis by: Sajjad Khan
CHAPTER 2. 2.1. LITERATURE REVIEW
Table 2.1: Summary of related work
Scheme(s) Achievement(s) Limitation(s)
LP [7] Consumer electricity cost is
minimized
PAR and UC are ignored
IPSO [13] Consumer electricity cost is
minimized
UC is ignored
MILP [14] Consumer electricity cost is
reduced
PAR and UC are ignored
CSOA and SA
[15]
Consumer electricity cost
and PAR are alleviated
UC is ignored
Fuzzy Logic [16] Consumer electricity cost
and PAR are minimized
UC is neglected
GA and PSO
[17]
UC is maximized Electricity cost and PAR
are not considered
Hybrid LSA-
ANN [18]
Consumer electricity cost is
minimized
Neglected PAR and power
system stability
GA [19] Consumer electricity cost is
minimized
PAR and UC are ignored
Hybrid HSA-
EDE [20]
Consumer electricity cost is
minimized
Performance in PAR reduc-
tion is not optimal
PSO and 2-PEM
[21]
Minimized the computa-
tional cost
Electricity cost and PAR
are ignored
Reinforcement
Learning OLA
[22]
Consumer electricity cost
and PAR are minimized
UC is ignored
LP[23] Consumer electricity cost
and PAR are minimized
UC is ignored
Game theoretic
approach [24]
Consumers electricity cost
is minimized
PAR and UC are ignored
GA,WDO,BFO,
BPSO and
HGPSO [25]
Consumer electricity cost
and PAR are minimized
UC is ignored
SAPCA [26] Consumer electricity cost
and PAR are minimized
UC is ignored
AIA [27] Consumers electricity cost
is reduced
PAR and UC are ignored
GA,CSOA and
CSA [28]
Consumers electricity cost
is alleviated
Ignored the operational and
maintenance cost of the ESS
EACA and GA
[29]
Reasonable UC is achieved
within a small budget
PAR is ignored
Game theory ap-
proach [30]
PAR is minimized Electricity cost and UC are
ignored
ILP [31] Consumers electricity cost
is minimized
PAR is ignored
GA, BPSO,
EDE [32]
PAR and cost are mini-
mized
UC is ignored
BA [33] UC is maximized PAR is not considered
11 Thesis by: Sajjad Khan
CHAPTER 2. 2.2. PROBLEM STATEMENT
as an input to the fuzzy controller. Fuzzy controller provides the required power
to operate the electrical appliances. It is shown in the experimental results that
this scheme reasonably minimized the error difference in the actual and optimized
environment parameters.
2.2 Problem statement
DR optimization techniques play a key role to overcome the increasing demand
of electricity. Many techniques have been proposed to overcome the varying elec-
tricity demand. However, there exists a trade off between conflicting objectives.
Normally, PAR and electricity cost is minimized at the expense of UC. The work
proposed in [28] minimized cost and PAR with a reasonable trade off in user waiting
time by incorporating ESS. However, this work did not consider the operational
and maintenance cost of the ESS. The work in [22],[23],[25] and [26] minimizes
electricity consumption cost. However, UC is completely ignored in this work.
The work in [29] developed an absolute UC level by using time and device based
priorities. However, in this work, PAR reduction is not taken into account. The
existing schemes only focuses on either maximizing UC or minimizing electricity
cost along with PAR alleviation. The work in [33] maximized UC. In this work UC
is referred as thermal comfort. None of the existing schemes considered minimizing
appliances waiting time. In [15] and [20], an ESS is incorporated in SHEMS for
PAR and cost minimization. The work in [16],[22],[23] and [25] minimized PAR
and energy consumption cost using the meta-heuristic techniques. However, these
scheme compromised UC. Moreover, none of these schemes considered rebound
peaks. The work in [32] mitigate rebound peaks using meta heuristic techniques.
However, this work considered a limited number of appliances.
12 Thesis by: Sajjad Khan
Chapter 3
Proposed system model
13
CHAPTER 3. 3.1. PROPOSED SYSTEM MODEL:
3.1 Proposed system model:
This thesis investigates energy consumption in a SH. The SH is comprised of
numerous smart appliances. Each appliance has a distinct power rating and Length
of Operational Time (LOT). Figure 3.1 represents the proposed system model.
The appliances considered in this thesis are taken from [36]. The SH is directly
connected with an electric grid. The electric grid has the feature of a SG (i.e.,
bi-directional communication between the utility and consumers). The use of SM
allows the consumer to extract the pricing signals and maintenance schedule from
the utility. Generally appliances scheduling is done for a single day. A day is
divided into multiple time slots. Each time slot comprises of one hour. Equation
3.1 represents the total time slots for a single day.
T=
24
X
t=1
ti(3.1)
3.1.1 Appliances categories
According to figure 3.1, the proposed SHEMS consist of various appliances. Based
on their distinguishing role in UC maximization and PAR minimization these
appliances are divided into three sub-classes. Class 1 contains the Deferrable
Interruptible (ADI ), class 2 is the Deferrable Non-interruptible (AD N ), whereas,
Utility
Smart meter
Figure 3.1: System model
14 Thesis by: Sajjad Khan
CHAPTER 3. 3.1. PROPOSED SYSTEM MODEL:
class 3 represents Non-Deferrable Appliances (AN D). AD I can be flexibly adjusted
in any time interval. AD N can also be adjusted in any time interval, however, these
appliances cannot be interrupted during their operations. ADN are the appliances
that should be continuously ON without any interruption in their respective time
intervals. Table 3.1 represents the set of appliances used in this work. It is
assumed that each appliance has the capability to communicate with the local
Energy Management Controller(EMC) using a home area network. The EMC
controls these appliances by switching them from high to low price intervals.
Table 3.1: Appliances categorization and power ratings
Class Name Power (kWh) LOT(h)
Class 1
Vacuum cleaner 0.7 6
Laptop 0.1 8
Desktop 0.3 6
Dish washer 1.8 8
Electric car 2.5 14
Class 2 Washing ma-
chine
1 3
Cloth dryer 3 5
Class 3
Cooker top 2.15 2
Refrigerator 0.3 24
Microwave 1.7 2
Interior lights 0.84 6
Cooker oven 5 2
3.1.2 Electricity price tariff
According to previous studies, the well known pricing schemes are CPP, IBRP,
RTP and ToUP. In this thesis, the pricing schemes used are RTP and CPP. RTP
is also known as dynamic pricing. It is commonly updated for every hour. In
CPP, the normal electricity price is replaced with a pre-defined high value to
minimize the energy demand as well as stress on the power grid. Figure 3.2 and
3.3 represents the per hour electricity cost announced by the utility.
15 Thesis by: Sajjad Khan
CHAPTER 3. 3.2. WAITING TIME OBJECTIVE
1 4 8 12 16 20 24
Time (hour)
Price (cent/kWh)
CPP
0
20
40
60
80
100
120
140
Figure 3.2: CPP signal
Time (hours)
0
5
10
15
20
25
30
RTP
1 4 8 12 16 20 24
Price (cent/kWh)
Figure 3.3: RTP signal
3.2 Waiting time objective
Figure 3.4 represents the execution pattern of an appliance. Let ωbe the time
when the consumer intended to switch an appliance ON, σrepresents the time
when the SHEMS initiates the appliance and Ω represents the finishing time of
appliance. The Waiting Time (WT) of an appliance is computed as the difference
between ωand σ. The objective of this thesis is to generate an optimal operational
pattern of home appliances such that the overall WT of appliances and electricity
cost is minimized. Hence, the objective function can be stated as in 3.2. In this
thesis, maximizing UC and reducing the appliances WT are used alternatively.
16 Thesis by: Sajjad Khan
CHAPTER 3. 3.3. PAR OBJECTIVE
ω
Ω
σ
WT LOT
Figure 3.4: Execution pattern of an appliance
minimize W T
s.t. CostScheduled < CostU nscheduled
(3.2)
The cost (in terms of money), paid to the utility by the consumers is computed as
the power consumed by an appliance in one hour weighted with the electricity price
announced by the utility in that hour. The per hour electricity cost announced
by the utility is shown in figure 3.2. From this figure, it can be seen that the per
hour cost during peak hours is very high. It means that high energy consumption
cost during these hours will result in high cost. Therefore, our goal is to minimize
electricity consumption during these hours. The energy utilization cost of home
appliances is computed using equation 3.3 and 3.4
CostH=ρ(t)∗η∗α(3.3)
CostT=
24
X
t=1
CostH(t) (3.4)
Here CostHshows the cost consumed by home appliances in one hour, CostTis the
total cost consumed by SH in one day, ηis the electricity pricing tariff announced
by the utility and αis the ON and OFF status of an appliance. ρ(t) represents
the power consumed by an appliance in one time slot.
3.3 PAR objective
In this thesis, the aim is to modify the execution time of home appliances such
that PAR and energy consumption cost are minimized. In order to minimize the
PAR, the appliances scheduling problem is formalized as an optimization problem.
17 Thesis by: Sajjad Khan
CHAPTER 3. 3.3. PAR OBJECTIVE
The objective function for PAR minimization can be written as:
minimize P AR
s.t. CostScheduled < CostU nscheduled
(3.5)
In order to compute the PAR, equation 3.6 is used.
P AR =LP eak
LAverage
(3.6)
Here LP eak is the maximum load consumed by the home appliances in SH for a
single time slot and LAverage is the average load consumed in one day. To compute
the per hour and total load consumed by the SH we use equation 3.7 and 3.8.
LoadH=ρ(t)∗α(3.7)
LoadT=
24
X
t=1
LoadH(t) (3.8)
Here LoadHrepresents the load consumed in one hour, LoadTis the total load
consumed in one day.
18 Thesis by: Sajjad Khan
Chapter 4
Existing and proposed schemes
19
CHAPTER 4. 4.1. OPTIMIZATION SCHEMES
4.1 Optimization schemes
The traditional optimization mechanisms have long been applied to solve the appli-
ances scheduling problem in a SHEMS. The trend to solve optimization problems
using meta-heuristic techniques has increased for the last few years. Some of the
well known meta-heuristic techniques are Flower Pollination Algorithm (FPA),
GA, PSO, CSO ant colony optimization and GWO. Due to their simplicity, con-
vergence to global solution and derivation free mechanism, scientists propose new
techniques to solve optimization problems from various fields.
4.1.1 Existing schemes
This section discusses the existing meta heuristic schemes.
4.1.1.1 FPA
In [34], Yang et al. develop a nature inspired algorithm on the basis of flower
pollination process. Typically, pollination occurs due to transfer of pollens. This
transfer is mainly linked with pollinators such as honeybees and insects etc. Pol-
lination in flowers can either be abiotic or biotic. Abiotic pollination in flowers
occurs due to water diffusion or wind blowing. From the view point of biological
studies, flower pollination is termed as survival of the fittest. Therefore, it is con-
sidered as an optimization process in plants reproduction where the objective is to
minimize the reproduction of weak flowers. FPA is idealized using the following
rules.
1. Cross and biotic pollination is termed as global pollination. It occurs due to
the Levy flights of pollinators.
2. Local pollination is termed as self and abiotic pollination.
3. Flower constancy is termed as the probability of reproduction in similar
flowers. It is directly proportional to the similarity of the flowers involved in
reproduction.
4. Switch probability decides the status of local or global pollination. It is
directly affected by physical proximity and environmental factors such as
wind blowing.
20 Thesis by: Sajjad Khan
CHAPTER 4. 4.1. OPTIMIZATION SCHEMES
4.1.1.2 Jaya Optimization Algorithm (JOA)
The population based meta-heuristic techniques are commonly based on proba-
bilistic approaches such as mutation probability in GA and switch probability in
FPA. Furthermore, these algorithms rely on algorithm specific parameters such as
mutation and crossover probability in GA. Improper tuning of these parameters
degrades the performance of these algorithms. Keeping in view, the parameter
specific nature of the meta-heuristic techniques R. Venkata Rao proposed JOA
[35]. JOA has no algorithm specific parameters. It is evident that the solution to
any constrained or un-constrained optimization problem can either be success or
failure. JOA is based on the idea that the algorithm proposed to solve a particular
problem must be victorious in acquiring the best solution and has to move away
from the worst solution. JOA is idealized using the following rules.
1. Given a population, identify the best and worst solutions.
2. Modify the candidate solutions based on the best and worst solutions.
4.1.1.3 SA
Plants are connected to the earth by means of their roots. It is evident that
these roots cannot move to different places like other animals and birds which
migrate to different places for their survival. However, there exist some plants
(e.g., Strawberry plant) which propagates by means of runners (stolon). Keeping
in view the vegetative propagation nature of Strawberry plants F. Merrikh-Bayat
in [37] proposed SA. Strawberry plant propagates by means of runners as well as
roots hair. If the strawberry plant is in a location which is not favorable for its
survival, the plant performs exploration. Exploration in Strawberry plants is the
process of sending long runners in different directions for its survival. A runner is a
creeping stalk which emerges from leaf axils of a parent or mother plant. If a runner
succeeds in locating a favorable condition for its survival, it generates additional
roots hairs and runners which directly effects the growth rate of Strawberry plant.
SA for optimization problem can be modeled using the following facts.
•All the Strawberry plants propagates by means of runners. These runners
arise randomly. Runners explore the global search space.
•All the Strawberry plants randomly generate roots and root hair. It allows
the plant to exploit the local search space.
21 Thesis by: Sajjad Khan
CHAPTER 4. 4.1. OPTIMIZATION SCHEMES
•If a strawberry plant have easy access to nourishing resources, the plant will
grow faster. Furthermore, that plant will generate more runners and roots.
Whereas, if a plant do not have access to nourishing resources, they have a
high death rate.
4.1.1.4 Salp Swarms Algorithm (SSA)
Inspired by the swarming behavior of Salp swarms, Mirjalilia et al. in [38] proposed
SSA. Salps belongs to the family of Salpidae. They possess a transparent barrel-
shaped body. Their tissue structure resembles the tissue structure of jelly fishes.
Moreover, their movement behavior also resembles jelly fish i.e., to move forward,
they pumped water through their body as repulsion. The biological research about
Salp swarms is at its early stages. Because Salps are mostly found in deep oceans.
Furthermore, it is very difficult to create a favorable living environment for Salps
in a laboratory. In deep oceans, Salps form a Salp chain. The main reason for
forming a Salp chain is not yet clear. However, scientists believe that foraging is
one of the many reasons. The Salp chain is divided into two groups i.e., leader and
followers. The leader of the Salp chain is at the front, whereas, the followers follow
their leader directly or indirectly. For modeling the Salp swarming behavior to
optimization problems, it is assumed that there exist a food source in the search
space. The leading Salp targets the food source by changing its position. Whereas,
the followers gradually follow the leading Salp.
4.1.2 Proposed schemes
In this section, the proposed schemes for UC maximization and PAR minimization
are discussed.
4.1.2.1 UCM
UCM is a population based meta-heuristic scheme. It starts by initializing ran-
dom candidate solutions. Every candidate solution is evaluated according to user
preferences. The best and worst candidate solutions are recorded so that the
performance of our scheme is enhanced. In this scheme, probabilistic approach is
22 Thesis by: Sajjad Khan
CHAPTER 4. 4.1. OPTIMIZATION SCHEMES
Algorithm 1 Algorithm for UCM
1: Objective min(W T ), Cost
2: Input:Random solutions, termination criteria,
Maximum Iteration
3: Determine the best and worst candidate solutions in the initial population
4: Output: Optimal schedule of home appliances
5: ittr = 0, transf ormation probability
6: while ittr < M aximum Iteration do
7: for i= 1 to P opulation do
8: flag= rand
9: if flag < p then
10: transf orm local candidate solution ;
11: else
12: transf orm global candidate solution ;
13: end if
14: end for
15: Identify the best and worst candidate solutions
16: if Solution moves towards the best then
17: U pdate solution ;
18: else
19: Modif y solution via mutation ;
20: end if
21: if New solution moves towards worst solutions then
22: U pdate solution ;
23: else
24: Discard new solution and keep old solution ;
25: end if
26: end while
27: return
employed to iteratively transform the candidate solutions. To transform the can-
didate solutions, the proposed scheme uses local or global transformation. Levy
flights are used for global transformation. The status of local or global transfor-
mation is determined randomly. After every transformation, the proposed scheme
evaluate the candidate solution. If the new solution meets the consumer require-
ments, the population is updated. However, if the solution does not accord with
the consumers requirements, then the candidate solution is modified on the basis
of the best and worst candidate solution. A mutation policy is adopted for this
23 Thesis by: Sajjad Khan
CHAPTER 4. 4.1. OPTIMIZATION SCHEMES
purpose. This policy to update candidate solutions ensures that the proposed so-
lution moves towards achieving the best solution. Moreover, this policy also helps
to reduce the number of iterations required to compute an optimal solution. The
pseudo code of UCM scheme is given in algorithm 1.
4.1.2.2 IAPR
It is evident that plant survives by their vegetative propagation nature. For this
purpose, plants send runners and root hair randomly. If the runners or the roots of
a plant succeed in locating a nourishing resource, the plant survives, otherwise the
plant dies. Motivated by the vegetative propagative nature of plants, we proposed
a survival based meta-heuristic scheme. In this scheme, first a pre-defined number
of explorers are initialized. With a rich nourishing resource available in the search
space, these explorers move towards that source independently. It is possible that
some of the explorers may not be able to reach the food source. This might
endanger the plant survival. Therefore, to tackle this problem, we evaluate the
fitness of the explorers and categorize the explorers as best and worst. A best
explorer is the one who succeed in locating the nourishing resource, whereas, the
worst explorer is the one who failed to do so. In this scheme, to prevent the worst
explorers from endangering the plant survival, the worst explorers are mutated
with best explorers. If the new explorers produced by mutation move towards the
food source, the explorers are updated by replacing the old explorers with the new
ones else if the new explorers do not move in the direction of nourishing resources,
they are discarded. In modeling the proposed IAPR to an optimization problem,
in this thesis the food source is termed as the global optimum solution. However,
the global optimal solution to any optimization problem is not known in advance,
therefore, in this scheme, it is assumed that a nourishing source located first will
remain a global solution unless a better nourishing resource is located. The pseudo
code of our proposed scheme is given in 2.
24 Thesis by: Sajjad Khan
CHAPTER 4. 4.1. OPTIMIZATION SCHEMES
Algorithm 2 Algorithm for IAPR
1: Randomly generate a nourishing resource
2: while stopping criteria not met do
3: Compute c using the current and total iterations
4: for i= 1 to size(explorers)do
5: Movement Probability= rand()
6: if P < M ovement P robability then
7: U pdate explorer position by adding c ;
8: else
9: Update explorer position by subtracting c ;
10: end if
11: end for
12: for all explorers do
13: if An explorer lies outside the legal region then
14: P lace it on the boundry ;
15: end if
16: end for
17: Evaluate the f itness of the explorers
18: Classify the explorers as best and worst ;
19: if New explorer moves towards the f ood source then
20: Replace new explorer with old worst explorer ;
21: else
22: Discard new explorer and keep old explorer ;
23: end if
24: end while
25: return Optimal operational pattern of home appliances
25 Thesis by: Sajjad Khan
Chapter 5
Simulation results and discussions
26
CHAPTER 5. 5.1. SIMULATION RESULTS AND DISCUSSIONS
5.1 Simulation results and discussions
In this chapter, we illustrate the effectiveness of our proposed schemes. Simulations
are performed for a single home with multiple appliances. The purpose of this
thesis is to develop a SHEMS that minimize the appliances waiting time and PAR.
The power ratings, operational hours and class of the appliances used in this thesis
are shown in table 3.1. The proposed and existing schemes were simulated using
MATLAB 2017a installed on Intel(R) Core(TM) i5-6430 CPU. In meta-heuristic
techniques, one of the most important features is randomization. To deal with it,
we have taken average results of ten experiments.
The performance of UCM is evaluated in terms of minimizing appliance waiting
time with minimum cost. To illustrate that UCM scheme is able to minimize the
appliances waiting time, the performance is compared with two well-known meta-
heuristic techniques namely FPA and JOA. According to figure 5.1, UCM scheme
effectively minimized the waiting time as compare to FPA and JOA. Table 5.1
presents the empirical results obtained via simulations for UCM scheme.
JOA FPA UCM
0
1
2
3
4
5
6
Waiting time (hour)
Figure 5.1: Appliances waiting time of UCM
Table 5.1: Simulation results of UCM scheme
Techniques WT Cost (cents) Load (kWh) PAR
Unscheduled 0 69449 140.0800 6.8715
JOA 5.9501 39489 140.0800 2.0260
FPA 2.9150 22706 140.0800 3.8496
UCM 2.7408 29149 140.0800 4.5112
27 Thesis by: Sajjad Khan
CHAPTER 5. 5.1. SIMULATION RESULTS AND DISCUSSIONS
In figure 3.2, it is shown that time slots 11-17 are high price intervals. Furthermore,
it is clear that increasing the load in these hours will increase the overall cost.
In this thesis, we reduced the hourly electricity consumption cost in the high
price intervals. This ultimately minimized the total electricity cost. Figure 5.2
represents the hourly electricity consumption cost of the home appliances. In this
Time (hour)
0
500
1000
1500
2000
Per hour cost (cent)
Unscheduled
JOA
FPA
UCM
1 4 8 12 16 20 24
Figure 5.2: Per hour cost consumption of UCM scheme
figure, it is clearly seen that the proposed scheme effectively minimized the electric-
ity consumption cost. Time slots 14-17 are recorded as the highest price intervals
time slots in our proposed scheme. However, it is due to the use of AD N in these
time slots. These appliances do not play any role in minimizing electricity con-
sumption cost. The overall electricity consumption cost for a single day is shown
in figure 5.3.
Unscheduled JOA FPA UCM
0
2000
4000
6000
8000
Total cost (cent)
Figure 5.3: Total cost of UCM scheme
28 Thesis by: Sajjad Khan
CHAPTER 5. 5.1. SIMULATION RESULTS AND DISCUSSIONS
This figure shows that UCM scheme minimizes electricity cost by 58.02%, whereas,
JOA and FPA minimized the cost by 43% and 67 % as compared to unscheduled
cost. However, FPA has a higher WT.
Time (hour)
0
5
10
15
Per hour load (kWh)
Unscheduled
JOA
FPA
UCM
1 4 8 12 16 20 24
Figure 5.4: Per hour load consumption of UCM scheme
Figure 5.4 shows the per hour load consumed by home appliances. In this thesis,
we focus on minimizing the load consumption during high price intervals. The load
consumption in the unscheduled scenario creates PAR in the high price intervals.
It is due to the irregular consumption pattern of unscheduled appliances. This is
the main reason of high electricity cost for unscheduled case. From this figure,
it is clear that the proposed scheme successfully shift the unscheduled load from
high to low price intervals. FPA also shows a similar pattern. Whereas, the
performance of JOA in cost consumption is not good. Shifting load from high to
low price intervals creates peak during low price intervals. As it can be seen from
Unscheduled JOA FPA UCM
0
2
4
6
8
PAR
Figure 5.5: PAR of UCM scheme
29 Thesis by: Sajjad Khan
CHAPTER 5. 5.1. SIMULATION RESULTS AND DISCUSSIONS
Unscheduled JOA FPA UCM
0
50
100
150
Total load (kWh)
Figure 5.6: Total load consumed by all schemes for UCM scheme
figure 5.4, the load consumption in time slots 21-24 is high. This is the trade-off
in our scheme. Figure 5.5 represents the PAR of all the schemes. This figure
shows that JOA outperforms the proposed scheme and FPA. PAR reduction in
JOA, FPA and proposed scheme is 70%, 56% and 34% respectively. Moreover, the
trade-off in all schemes is shown in this figure. JOA minimized PAR, however, its
performance in WT and cost minimization is not good. Similarly, FPA minimized
the total cost, however, the performance in WT and PAR is not good. The
proposed scheme efficiently minimized the appliances WT an acceptable trade off
in the consumption cost. However, the PAR is increased. Here, it is noteworthy
that the total load consumed in all schemes is the same as shown in figure 5.6.
Figure 5.7 and 5.8 represents that IAPR outperforms the well known meta-heuristic
techniques SSA and SA. The proposed IAPR minimized PAR using CPP by 72%.
Whereas, SA and SSA minimized PAR by 51% and 9% respectively. In RTP,
PAR alleviation is 17%, 48% and 68% in SSA, SA and IAPR respectively. Figure
5.13 and 5.14 shows the per hour load consumption pattern of SH. These figures
gives a complete picture of the performance of IAPR using CPP and RTP. The
IAPR balanced the per hour load consumption, as a result PAR in minimized.
The maximum load recorded in a single time slot using CPP and RTP schemes is
10.54, 7.80, 9.60 and 7.80 kWh for unscheduled, SA, SSA and IAPR respectively.
30 Thesis by: Sajjad Khan
CHAPTER 5. 5.1. SIMULATION RESULTS AND DISCUSSIONS
Unscheduled SA SSA IAPR
0
2
4
6
PAR
Figure 5.7: PAR reduction using CPP scheme in IAPR
Unscheduled SA SSA IAPR
0
2
4
6
PAR
Figure 5.8: PAR reduction using RTP scheme in IAPR
Here it is worth mentioning that PAR is minimized by balancing load consumption
throughout the day and total load consumed by all algorithms with both pricing
Table 5.2: Simulation results of IAPR
Algorithm Pricing
scheme
PAR Cost
(cents)
Waiting
time
(hours)
Total load
(kWh)
Unscheduled CPP 5.9002 4956 0 104.1400
SA CPP 2.8343 3269.9 5.4061 104.1400
SSA CPP 5.3111 1187.2 4.5620 104.1400
IAPR CPP 1.6270 3362.2 6.1070 104.1400
Unscheduled RTP 5.9002 1447.8 0 104.1400
SA RTP 3.0147 1221.6 2.7135 104.1400
SSA RTP 4.8947 868.5276 2.7135 104.1400
IAPR RTP 1.8759 1197 6.7942 104.1400
31 Thesis by: Sajjad Khan
CHAPTER 5. 5.1. SIMULATION RESULTS AND DISCUSSIONS
Unscheduled SA SSA IAPR
0
1000
2000
3000
4000
5000
Total cost (cents)
Figure 5.9: Total cost using CPP scheme in IAPR
Unscheduled SA SSA IAPR
0
500
1000
1500
Total cost (cents)
Figure 5.10: Total cost using RTP scheme in IAPR
scheme was same as shown in table 5.2. Figure 5.9 and 5.10 shows the performance
in overall cost reduction. It can be seen from these figures that all the algorithms
minimized cost. In CPP tariff, cost reduction as compare to unscheduled scenario
is 34%, 76% and 32% using SA, SSA and IAPR respectively. Whereas, in RTP
scheme cost is minimized by 15%, 40% and 17% using SA, SSA, and IAPR. The
proposed IAPR balanced the hourly load consumption which has a direct impact
on the total cost. Figure 5.11 and 5.12 shows the per hour cost consumption in
SH. These figures reveal that IAPR successfully minimized unscheduled cost by
load shifting. Furthermore, cost reduction in IAPR with CPP scheme is higher as
compare RTP scheme.
The WT of all the algorithms using CPP and RTP scheme is shown in 5.15 and
5.16. The proposed IAPR has a higher WT as compare to the other optimization
algorithms. It can be seen from figure 5.7-5.16 that all the algorithms with both
the pricing strategies are confronted with trade-offs. SSA minimized the electricity
32 Thesis by: Sajjad Khan
CHAPTER 5. 5.1. SIMULATION RESULTS AND DISCUSSIONS
Time (hours)
Per hour cost (cents)
Unscheduled
SA
SSA
IAPR
1 4 8 12 16 20 24
0
200
400
600
800
1000
1200
Figure 5.11: Per hour cost using CPP scheme in IAPR
Time (hours)
Per hour cost (cents)
Unscheduled
SA
SSA
IAPR
1 4 8 12 16 20 24
200
150
100
50
0
Figure 5.12: Per hour cost using RTP scheme in IAPR
consumption cost. However, the performance in PAR minimization is not good
and multiple peaks are created in Off-peak hours. SA minimized the PAR as
compare to SSA. However its performance in waiting time reduction is higher
than SSA. Similarity IAPR minimized PAR, however, it increased the appliances
waiting time as compare to SA and SSA. Figure 5.17 represents the total load
consumed by all schemes. It is note worthy that the total load consumed by all
schemes using CPP and RTP is equal.
33 Thesis by: Sajjad Khan
CHAPTER 5. 5.1. SIMULATION RESULTS AND DISCUSSIONS
Time(hours)
Per hour load (kWh)
Unscheduled SA SSA IAPR
1 4 8 12 16 20 24
8
6
4
2
0
10
12
Figure 5.13: Per hour load using CPP scheme in IAPR
Time (hours)
Per hour load (kWh)
Unscheduled SA SSA IAPR
1 4 8 12 16 20 24
8
6
4
2
0
10
12
Figure 5.14: Per hour load using RTP scheme in IAPR
Unscheduled SA SSA IAPR
0
2
4
6
8
Waiting time (hours)
Figure 5.15: Waiting time using CPP scheme in IAPR
34 Thesis by: Sajjad Khan
CHAPTER 5. 5.1. SIMULATION RESULTS AND DISCUSSIONS
Unscheduled SA SSA IAPR
0
1
2
3
4
5
6
7
Waiting time (hours)
Figure 5.16: Waiting time using RTP scheme in IAPR
Unscheduled SA SSA IAPR
0
20
40
60
80
100
120
Total load (kWh)
Figure 5.17: Total load consumed using CPP and RTP in IAPR
35 Thesis by: Sajjad Khan
Chapter 6
Conclusion and future work
36
CHAPTER 6. 6.1. CONCLUSION
6.1 Conclusion
SG possess the potential to solve the increasing demand of electricity. In this
thesis, a SHEMS is developed to minimize the appliances waiting time and PAR.
For appliances waiting time reduction a novel meta heuristic technique namely
UCM is developed. UCM successfully shifted the operational time of home appli-
ances from high to low price intervals, taking into account the appliances waiting
time. Performance of UCM scheme is compared with the well known popula-
tion based meta-heuristic techniques namely JOA and FPA. Experimental results
demonstrate that in terms of appliances waiting time reduction, UCM is 5.97%
and 53.9% efficient as compare to FPA and JOA respectively. Moreover, UCM
minimized the total electricity cost by 58% and PAR by 56% as compared to un-
scheduled scenario using the CPP scheme. Furthermore, in this thesis, an IAPR
is developed to encourage the residential consumers to eagerly participate in en-
hancing the reliability of power system. To validate the effectiveness of IAPR,
the performance in PAR reduction is compared with the renowned meta-heuristic
techniques namely SA and SSA using the RTP and CPP scheme. Simulation re-
sults show that IAPR in PAR minimization with both the pricing schemes exceeds
SA and SSA. PAR minimization using the CPP scheme in IAPR, SA and SSA
is 72%, 51% and 9% as compare to the unscheduled scenario. Whereas, with the
RTP scheme, PAR is minimized by 17%, 48% and 68% using SSA, SA and IAPR.
6.2 Future work
In future, we plan to develop a SHEMS by integrating DES for single and multi
homes scenario. Moreover, this work can also be extended by incorporating RES
at consumers’ end to curtail the energy utilization cost, maximize UC as well as
trade surplus energy with the neighboring homes within a community.
37 Thesis by: Sajjad Khan
Chapter 7
References
38
References
[1] Awais, Muhammad, Nadeem Javaid, Khursheed Aurangzeb, Syed Haider, Za-
hoor Khan, and Danish Mahmood. ”Towards Effective and Efficient Energy
Management of Single Home and a Smart Community Exploiting Heuristic
Optimization Algorithms with Critical Peak and Real-Time Pricing Tariffs in
Smart Grids.” Energies 11, no. 11 (2018): 3125.
[2] Fang, Xi, Satyajayant Misra, Guoliang Xue, and Dejun Yang. ”Smart
grid—The new and improved power grid: A survey.” IEEE communications
surveys & tutorials 14, no. 4 (2012): 944-980.
[3] Samuel, Omaji, Sakeena Javaid, Nadeem Javaid, Syed Ahmed, Muhammad
Afzal, and Farruh Ishmanov. ”An efficient power scheduling in smart homes
using Jaya based optimization with Time-of-Use and critical peak pricing
schemes.” Energies 11, no. 11 (2018): 3155.
[4] Farhangi, Hassan. “The path of the smart grid.” IEEE power and energy mag-
azine 8, no. 1 (2010).
[5] Economics, Frontier, and Sustainability First. “Demand side response in the
domestic sector-a literature review of major trials.” Final Report, London,
August (2012).
[6] Hussain, Bilal, Nadeem Javaid, Qadeer Hasan, Sakeena Javaid, Asif Khan, and
Shahzad Malik. ”An Inventive Method for Eco-Efficient Operation of Home
Energy Management Systems.” Energies 11, no. 11 (2018): 3091.
[7] Shakeri, Mohammad, Mohsen Shayestegan, SM Salim Reza, Iskandar Yahya,
Badariah Bais, Md Akhtaruzzaman, Kamaruzzaman Sopian, and Nowshad
Amin. “Implementation of a novel home energy management system (HEMS)
architecture with solar photovoltaic system as supplementary source.” Renew-
able Energy 125 (2018): 108-120.
[8] Pilloni, Virginia, Alessandro Floris, Alessio Meloni, and Luigi Atzori. “Smart
home energy management including renewable sources: A QoE-driven ap-
proach.” IEEE Transactions on Smart Grid 9, no. 3 (2018): 2006-2018.
[9] Shareef, Hussain, Maytham S. Ahmed, Azah Mohamed, and Eslam Al Has-
san. “Review on Home Energy Management System Considering Demand
Responses, Smart Technologies, and Intelligent Controllers.” IEEE Access 6
(2018): 24498-24509.
39
REFERENCES REFERENCES
[10] Lior, Noam.“Sustainable energy development: the present (2009) situation
and possible paths to the future.” Energy 35, no. 10 (2010): 3976-3994.
[11] Sajjad Khan, Zahoor Ali Khan, Nadeem Javaid, Waleed Ahmad, Raza Abid
Abbasi and Hafiz Muhammad Faisal, “On Maximizing User Comfort using a
Novel Meta-Heuristic Technique in Smart Home”, in 33rd International Con-
ference on Advanced Information Networking and Applications (AINA).
[12] Sajjad Khan, Nadeem Javaid, Zahoor Ali khan, Sahibzada Muhammad Shuja,
Muhammad Abdullah and Annas Chand, “Energy Efficient Scheduling of
Smart Homes”, in 33rd International Conference on Advanced Information
Networking and Applications (AINA), 2019
[13] Yang, Hong-Tzer, Chiao-Tung Yang, Chia-Chun Tsai, Guan-Jhih Chen, and
Szu-Yao Chen. “Improved PSO based home energy management systems inte-
grated with demand response in a smart grid.” In Evolutionary Computation
(CEC), 2015 IEEE Congress on, pp. 275-282. IEEE, 2015.
[14] Van Der Stelt, Sander, Tarek AlSkaif, and Wilfried van Sark. “Techno-
economic analysis of household and community energy storage for residential
prosumers with smart appliances.” Applied Energy 209 (2018): 266-276.
[15] Aslam, Sheraz, Nadeem Javaid, Farman Ali Khan, Atif Alamri, Ahmad Almo-
gren, and Wadood Abdul. “Towards Efficient Energy Management and Power
Trading in a Residential Area via Integrating a Grid-Connected Microgrid.”
Sustainability 10, no. 4 (2018): 1245.
[16] Collotta, Mario, and Giovanni Pau. “A novel energy management approach
for smart homes using bluetooth low energy.” IEEE Journal on Selected Areas
in Communications 33, no. 12 (2015): 2988-2996.
[17] Ullah, Israr, and DoHyeun Kim. “An improved optimization function for
maximizing user comfort with minimum energy consumption in smart homes.”
Energies 10, no. 11 (2017): 1818.
[18] Ahmed, Maytham S., Azah Mohamed, Raad Z. Homod, and Hussain Shareef.
“Hybrid LSA-ANN based home energy management scheduling controller for
residential demand response strategy.” Energies 9, no. 9 (2016): 716.
[19] Bharathi, C., D. Rekha, and V. Vijayakumar. “Genetic algorithm based de-
mand side management for smart grid.” Wireless Personal Communications
93, no. 2 (2017): 481-502.
[20] Kazmi, Saqib, Nadeem Javaid, Muhammad Junaid Mughal, Mariam Akbar,
Syed Hassan Ahmed, and Nabil Alrajeh. “Towards optimization of metaheuris-
tic algorithms for IoT enabled smart homes targeting balanced demand and
supply of energy.” IEEE Access (2017)
[21] Huang, Yantai, Lei Wang, Weian Guo, Qi Kang, and Qidi Wu. “Chance con-
strained optimization in a home energy management system.” IEEE Transac-
tions on Smart Grid 9, no. 1 (2018): 252-260.
40 Thesis by: Sajjad Khan
REFERENCES REFERENCES
[22] Wen, Zheng, Daniel O’Neill, and Hamid Maei. “Optimal demand response
using device-based reinforcement learning.” IEEE Transactions on Smart Grid
6, no. 5 (2015): 2312-2324.
[23] Adika, Christopher O., and Lingfeng Wang. “Smart charging and appliance
scheduling approaches to demand side management.” International Journal of
Electrical Power & Energy Systems 57 (2014): 232-240.
[24] Samadi, Pedram, Vincent WS Wong, and Robert Schober. “Load schedul-
ing and power trading in systems with high penetration of renewable energy
resources.” IEEE Transactions on Smart Grid 7, no. 4 (2016): 1802-1812.
[25] Ahmad Adnan, Asif Khan, Nadeem Javaid, Hafiz Majid Hussain, Wadood
Abdul, Ahmad Almogren, Atif Alamri, and Iftikhar Azim Niaz. “An optimized
home energy management system with integrated renewable energy and storage
resources.” Energies 10, no. 4 (2017): 549.
[26] Qian Li Ping, Ying Jun Angela Zhang, Jianwei Huang, and Yuan Wu. “De-
mand response management via real-time electricity price control in smart
grids.” IEEE Journal on Selected areas in Communications 31, no. 7 (2013):
1268-1280.
[27] Li Dan, Wei-Yu Chiu, Hongjian Sun, and H. Vincent Poor. “Multiobjective
Optimization for Demand Side Management Program in Smart Grid.” IEEE
Transactions on Industrial Informatics 14, no. 4 (2018): 1482-1490.
[28] Aslam, Sheraz, Zafar Iqbal, Nadeem Javaid, Zahoor Ali Khan, Khursheed
Aurangzeb, and Syed Itaza Haider. “Towards efficient energy management of
smart buildings exploiting heuristic optimization with real time and critical
peak pricing schemes.” Energies 10, no. 12 (2017): 2065.
[29] Khan, Asif, Nadeem Javaid, and Majid Iqbal Khan. “Time and device based
priority induced comfort management in smart home within the consumer
budget limitation.” Sustainable Cities and Society (2018).
[30] Huang, Hantao, Yuehua Cai, Hang Xu, and Hao Yu. “A multiagent minority-
game-based demand-response management of smart buildings toward peak
load reduction.” IEEE Transactions on Computer-Aided Design of Integrated
Circuits and Systems 36, no. 4 (2017): 573-585.
[31] Farrokhifar, Meisam, Farid Momayyezi, Nasser Sadoogi, and Amin Safari.
“Real-time based approach for intelligent building energy management using
dynamic price policies.” Sustainable cities and society 37 (2018): 85-92.
[32] Khan, Asif, Nadeem Javaid, Adnan Ahmad, Mariam Akbar, Zahoor Ali Khan,
and Manzoor Ilahi. “A priority-induced demand side management system to
mitigate rebound peaks using multiple knapsack.” Journal of Ambient Intelli-
gence and Humanized Computing (2018): 1-24.
41 Thesis by: Sajjad Khan
REFERENCES REFERENCES
[33] Fayaz, Muhammad, and Do Hyeun Kim. “Energy Consumption Optimization
and User Comfort Management in Residential Buildings Using a Bat Algorithm
and Fuzzy Logic.” Energies 11, no. 1 (2018): 161.
[34] Yang, Xin-She. “Flower pollination algorithm for global optimization.” In In-
ternational conference on unconventional computing and natural computation,
pp. 240-249. Springer, Berlin, Heidelberg, 2012.
[35] Rao, R. “Jaya: A simple and new optimization algorithm for solving con-
strained and unconstrained optimization problems.” International Journal of
Industrial Engineering Computations 7, no. 1 (2016): 19-34.
[36] Pooranian, Zahra, Jemal H. Abawajy, and Mauro Conti. “Scheduling Dis-
tributed Energy Resource Operation and Daily Power Consumption for a
Smart Building to Optimize Economic and Environmental Parameters.” En-
ergies 11, no. 6 (2018): 1348.
[37] Merrikh-Bayat, Farshad. “A numerical optimization algorithm inspired by
the strawberry plant.” arXiv preprint arXiv:1407.7399 (2014).
[38] Mirjalili, Seyedali, Amir H. Gandomi, Seyedeh Zahra Mirjalili, Shahrzad
Saremi, Hossam Faris, and Seyed Mohammad Mirjalili. “Salp Swarm Algo-
rithm: A bio-inspired optimizer for engineering design problems.” Advances in
Engineering Software 114 (2017): 163-191.
42 Thesis by: Sajjad Khan
Appendices
43
.A. DETAIL OF APPENDICES
.A Detail of appendices
This section presents the detail of appendices for the thesis entitled “Multi-Objective
Home Energy Management System with Multi-Class Appliances using Meta-Heuristic
Techniques ”. This code is developed by Sajjad Khan under the supervision of
Dr. Nadeem Javaid. To execute this code, copy the code given below and paste
in MATLAB files with the respective name as given at the start of each function
with .m extension. If you have any query regarding execution of this code, please
feel free to contact. The best way to reach me is via email.
Email address: nadeemjavaidqau@gmail.com, ciit.sajjad@gmail.com
The detail of appendices are as follows.
1 Appendix B contains the MATLAB code for waiting time objective.
2 Appendix C contains the MATLAB code for PAR reduction objective.
44 Thesis by: Sajjad Khan
.B. IMPLEMENTATION OF WAITING TIME OBJECTIVE
.B Implementation of waiting time objective
Code for waiting time objective1
2
3
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%4
%% Ma t l ab co d e f or : Wa i t in g _ T im e _ Ob j e ct i v e . m5
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%6
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%7
%% Ele c tr i cit y p ri c ing s ign al8
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%9
EP =[ 11. 4 1 1. 4 11. 4 11 . 4 11 .4 11 .4 11 .4 1 1 .4 1 1.4 1 1. 4 1 23 . 4 12 3 .410
12 3 .4 1 2 3. 4 1 23 . 4 123 .4 11. 4 11. 4 11 . 4 11 .4 11 .4 1 1 .4 1 1.4 1 1.4 ]11
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%12
%% Ini tia l O pe r ati o na l P att ern of Ap pli a nc e s13
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%14
Vac u um _ c le a ner =[ 1 0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 1 1 1 0 1 ];15
La p top = [1 0 0 0 0 0 0 0 1 0 0 1 1 0 1 0 1 0 0 1 0 0 0 1];16
De s kto p = [1 1 1 1 0 0 0 0 0 0 0 1 1 0 0 0 1 0 0 1 0 0 1 1] ;17
Dis h _w a she r = [0 0 0 0 0 0 1 1 1 0 0 1 1 1 0 0 0 0 0 0 0 1 1 0] ;18
Ele c tr i c_C a r = [1 1 1 1 1 1 0 1 0 0 1 0 1 0 1 1 0 1 1 1 0 0 0 0] ;19
Was h in g _ ma c h in e = [ 0 0 0 0 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0];20
Clo t h_ d rye r = [0 0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0] ;21
Coo k er _ Top = [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0] ;22
Ref r ig e rat o r = [1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1] ;23
Mic row a ve = [ 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0];24
Int e ri o r _L i g ht 1 = [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 ];25
Coo k er _ Ove n = [1 1 0 0 0 0 0 0 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0] ;26
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%27
%% For min g a mat r ix fo r c o mp u tat i on28
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%29
I nt r r up t i bl e _ a pp l i an c e s =[ Va c u um _ cl e an e r ’, L a pt op ’ , D es kt o p ’ , D i sh _ wa s he r ’ ,30
Electric_Car ’];31
N on _ in t rr u p ti b le _ ap p li a n ce s = [ W as hi n g_ m ac h in e ’ , C l ot h_ d ry e r ’ ];32
N on _ sc h e du l a bl e _ Ap p li a n ce s = [ C oo k er _T o p ’ , R ef r ig e ra t or ’ , M ic ro w av e ’,33
In t er i or _ Li g ht 1 ’ , Co o ke r_ O ve n ’] ;34
Appliances=[Intrruptible_appliances Non_intrruptible_appliances35
Non_schedulable_Appliances];36
time_slots_Intrruptible=[6 8 6 8 14];37
time_slots_NonIntrruptible=[3 5];38
time_slots_Nonschedulable=[2 24 2 6 2];39
time_slots=[time_slots_Intrruptible time_slots_NonIntrruptible40
time_slots_Nonschedulable ];41
42
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%43
%% Po w er r at i ng s of Ap p lia n ce s44
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%45
pIn t rr u pti b le =[0 .7 0. 1 0. 3 1.8 2 .5] ;46
pNonIntrruptible=[1 3];47
pNo n sc h e du l a bl e =[ 2 .1 5 0.3 1 .7 0 .84 5 ];48
p =[ pI n t r ru p t i bl e p No n I n tr r u p ti b l e pN o n s ch e d u la b l e ] ;49
50
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%51
%% Com put i ng the o p er a tio n al ti me sl o ts o f app l ia n ces52
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%53
54
45 Thesis by: Sajjad Khan
.B. IMPLEMENTATION OF WAITING TIME OBJECTIVE
tVC1=strfind (Vacuum_cleaner ,(1));55
tW H 1 = st r fi n d ( La p to p , ( 1) ) ;56
tW P 1 = st r fi n d ( De s kt op , ( 1) ) ;57
tD W 1 = st r fi n d ( D is h_ w as h er , ( 1) ) ;58
tI R 1 = st r fi n d ( E le c tr ic _ Ca r , (1 ) ) ;59
tW M 1 = st r fi n d ( W as h in g _m a ch i ne , ( 1) ) ;60
tC D 1 = st r fi n d ( C lo th _ dr y er , ( 1) ) ;61
tO V 1 = st r fi n d ( Co o ke r _T o p ,( 1) ) ;62
tB L 1 = st r fi n d ( R ef r ig er a to r , (1 ) ) ;63
tL O 1 = st r fi n d ( Mi c ro w av e ,( 1 ) );64
tL T 1 = st r fi n d ( I nt e ri o r_ L ig h t1 , ( 1) ) ;65
tL T H1 = s t rf in d ( C oo k er _O v en , (1 ) ) ;66
67
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%68
%% % Co m pu t in g Un s c he d ul e d L oa d , c os t a n d P AR69
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%70
71
fo r h ou r =1 :2 472
E le c tr i c it y _ co s t =[ E P ( ho u r )* p ( 1) EP ( h ou r ) * p (2 ) EP ( h o ur ) * p (3 ) E P ( ho ur ) * p ( 4)73
EP ( h ou r ) * p (5 ) E P( h o ur ) * p (6 ) E P ( ho ur ) * p ( 7) E P ( ho u r )* p ( 8) EP ( h ou r ) * p( 9 ) EP (74
ho u r )* p ( 10 ) E P ( ho u r )* p ( 11 ) E P ( ho u r )* p ( 12 ) ] ;75
u ns c he d u le d lo a d ( h ou r ) = Ap p l ia n ce s ( h ou r , :) * p ’;76
u ns c he d u le c os t ( h ou r ) = A pp l ia n c es ( h ou r , : )* El e ct r ic i t y_ c os t ’;77
en d78
s um _ u ns c h ed u l ed l o ad = su m ( u n sc h ed u l ed l o ad );79
s um _ u ns c h ed u l ed _ c os t = s u m ( un s c he d u le c o st );80
P AR _ u ns c h ed u l ed = m a x ( u ns c h ed u l ed l oa d ) ^ 2 /( s um _ u ns c h ed u l ed l o ad / 24 ) ^ 2;81
D_ E P = ro u nd ( m e an ( E P) ) ;82
lo a d_ (1 : le n gt h( u n sc he du l ed lo a d) ) =( u ns c he du l ed lo ad (1 : le n gt h(83
u ns c h e du l e d lo a d ) ) - m in ( un s c h ed u l e dl o a d ) ) /( ma x ( u n s ch e d u le d l o ad ) - m i n (84
u ns ch e du l ed l oa d ) );85
t_ l o ad = s um ( lo a d_ ) ;86
n=80;87
N_ it e r =10 0;88
d=12;89
90
91
FPA Code92
93
94
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%