Content uploaded by Abdul Mateen
Author content
All content in this area was uploaded by Abdul Mateen on Jan 12, 2022
Content may be subject to copyright.
Bio-Inspired Optimization Techniques for Home
Energy Management in Smart Grid
Abdul Mateen1, Nadeem Javaid1,∗, Muhammad Awais1,
Nasir Khan1, Urva Latif1, Ihtisham Ullah1
1COMSATS Institute of Information Technology, Islamabad 44000, Pakistan
∗Corresponding author: www.njavaid.com, nadeemjavaidqau@gmail.com
Abstract—Smart Grid (SG) plays a noteworthy role in min-
imizing the Electricity Cost (EC) through Demand Side Man-
agement (DSM). Smart homes are the part of SG, pays a lot
in minimizing EC via scheduling the appliances. Home Energy
Management (HEM) have been extensively used for energy
management in smart homes. In this paper, for the effective
utilization of energy in a smart home, we propose a solution
that consists of bio-inspired techniques: Genetic Algorithm (GA),
Flower Pollination Algorithm (FPA) and hybrid of these two,
Genetic Flower Pollination Algorithm (GFPA). All of these
techniques applied to the appliances that are essential in a home.
Our proposed solution leads to find an optimal scheduling pattern
that reduces EC, Peak to Average Ratio (PAR) and maximize
User Comfort (UC). In our work, we have considered one home.
We divide appliances into three categories, non-interruptible,
interruptible and fixed appliances. Simulation results show that
our proposed schemes performed better in terms of EC, UC and
PAR. We have done this work for three different Operational
Time Intervals (OTIs) 15, 30 and 60 minutes for each appliance.
Index terms— Demand response, Genetic algorithm,
Flower pollination, Smart grid, Critical peak pricing, Home
energy management
I. INTRODUCTION
SG uses new and forward-looking technologies comprising
independent controllers, intelligent hardwares, softwares and
other means for two way communication, between utilities and
clients for the organization of data to deliver electricity in an
effectual way. The major aims of the SGs are to increase the
efficiency, trustworthiness and safety of the system. A brief
overview about the smart HEMS is presented by the [8]. They
discuss the functional modules and architectures of HEMS in
their paper.
DSM also called energy demand management, which al-
lows clients to alteration their demand for electricity, so that
electricity practice is minimized during on-peak slots. Demand
Response (DR) plans are in the form of economic inducements
or other time based charges which offer sufficient opportunity
for clients in shifting or dropping appliances load to off-peak
slots. Thus, DR has reflected the most reliable solution to
lessening the peak demand and flat the request curve. DR
offers motivation in the form of price signals including CPP,
Time-Of-Use (TOU), Real-Time Pricing (RTP), Peak Load
Pricing (PLP), Day-Ahead Pricing (DAP) or Day-Ahead RTP
(DA-RTP), Inclined Block Rate (IBR).
EC minimization by appliances scheduling is one of the
challenging task in SG. The EC problem of appliances is
considered as an optimization task and explained using nu-
merous techniques in the literature. An important problem in
SG is UC which is mostly abandoned in EC minimization
problems. Studies expose that consumers wish to lessen their
EC. However, they do not want to negotiation on their comfort.
The simulations are accompanied using three scenarios by
taking different OTIs. GA and FPA also reduces and controls
the PAR of the SG. Moreover, the monthly and yearly EC is
confirmed.
Most Common objectives of electricity management in SG
are PAR reduction, minimization of EC, UC maximization
and minimize the power consumption. A vast amount of
electricity is spent by residential area and its consumption
is growing speedily. Automated Demand Response (ADR)
based optimization is proposed in [6]. Scheduling of different
appliances at domestic level was under consideration. Their
aim was to maximize UC and minimize the EC. According to
[23], through energy production, dispersion and dissemination,
the electricity that is produced, more than 65% is wasted. With
the fruition of smart grid, we have good opportunity to save
maximum energy.
Rest of the paper is organized as follows: section II high-
lights the related work and describes the problem statement,
system model is at section III. GA and FPA is discussed
in Simulation and Discussion‘s section. Section VI covers
Conclusion.
II. RE LATE D WORK
A. Related Work
Many techniques have been proposed which includes mod-
els and implementation in the field of SG using different
algorithms. In these techniques, writers facilitate the customers
and the utility through minimization in cost and load optimiza-
tion. Writers consider different scheduling parameters in these
techniques: such as appliances, price scheme, user demand
and many others. The HEM can provide an efficient way of
optimizing energy usage in smart homes.
A GA base energy management system in SG is proposed in
[1]. In this proposed system, the authors objective is to lower
the EC and other problems that are created by high PAR. The
high PAR creates when the CPP scheme is used because there
is a possibility that, most appliances run during the time when
CPP provide the lowest electricity price. The proposed system
effectively minimized EC and PAR. The authors ignored UC
and cannot explain how much EC minimized. In [2], authors
presents another effective GA-based DSM model in the smart
grid. The objective of the proposed model is to minimize the
power utilization during a day by effectively distributing user
demand during max-peak and min-peak hours. The authors
are adopted evolutionary algorithm GA for minimization the
power utilization.
In [3], proposed model is categorized into three types:
an interruptible, non-interruptible and fixed load. Simulation
performed in MATLAB by giving inputs of many types of
the load. In [4], the authors present Predictive Demand Side
Management (PDSM) system. In [5], a Non-dominated Sorting
Genetic Algorithm (NSGA) technique is used for solving the
DSM multi-objective optimization problem.
The authors proposed DR application in HEM system to
optimize the operational time of appliances [6]. The authors
addressed DR optimization problem in proposed application.
The objectives of proposed application is to minimize the com-
putational cost and operational time of appliances according
to given usage pattern. In [7], the main objective is to reduce
the cost, pollution emission and solve the uncertainty problem
of energy sources. In DSM, the end users alleviate the EC
and minimize PAR by scheduling home appliances from high
peak hours to low peak hours or by integrating Renewable
Energy Sources (RES). Main functionalities of DSM that are
discussed in [8], are management of load and DR.
The authors in [9], proposed a technique for balancing load
in industry, residential and commercial areas. Comparison of
electricity consumption through DSM with GA (GA-DSM)
and without GA is done by the authors. The comprehensive
simulation results shown that the recommended scheme GA-
DSM attained the objectives. During peak hours, reduction in
the electricity consumption is 21.91%. However, the authors
did not discuss about UC and PAR. In [10], smart charging
and appliance scheduling are features while targeting the
residential area. Main objectives are to reduce Cost and PAR,
but ignores UC, the initial and maintenance cost. In [11], the
authors used efficient GA based DSM scheme for SG. What
the things he gains, are cost minimization and PAR, however,
inconsideration of UC and larger appliance delay. Queuing-
based Energy Consumption (QEC) monitoring for different
smart homes in SG is under consideration in [12].
In [13], networks employ low-cost, small devices that in-
tellect biological body structures (e.g. heat, heart beat) and
move forward it to places. The main reason is that the
desirability of these sensor nodes is its energy consumption
and data rate. In [14], the authors presented state of art EMC
model for HEM in order to avoid peak formations while
focusing on utilities electricity cost reduce, by consider UC
level within acceptable limits. The LP and GA algorithms are
used. Authors consider cost minimization, UC maximization.
However, they ignored PAR. In [15], the author deliberated
two different systems to explore the incremental price for
short term arranging of a distribution system with Demand
Grid (DG) units and presented three different scenarios with
various approaches via optimizing benefit functions in a multi-
objective optimization framework. However, the author work
for short term scheduling.
Manageable load modeling is proposed in [16], appropriate
for equally local and direct control from a regulation system
for Distributed Energy Resources (DER) optimization based
on multi objectives. In [17], the author proposed load control
algorithm for DSM. In which author deliberated the problem
of power and load scheduling to lessen the energy expense
of users. In [18], the centralized optimization problem P1
is proposed to minimize PAR and P2 to reduce total energy
consumption to reduce cost.
The author used FPA in [20], for the tuning of Mass Damper
(MD) also known as harmonic absorber, usually used to reduce
the vibration in structures. FPA works when mathematical
equations cannot be applied, as it is meta heuristic just like
others algorithms e.g. GA and Harmony Search Algorithm
(HSA).The author used FPA in [20], for the tuning of Mass
Damper (MD) also known as harmonic absorber, usually used
to reduce the vibration in structures. FPA works when mathe-
matical equations cannot be applied, as it is meta heuristic just
like others algorithms e.g. GA and Harmony Search Algorithm
(HSA). Authors in [25], used GA for scheduling multiple
homes. However, they use to hybrid of GA with Cuckoo
Search Algorithm (CSA). Where, CPP and RTP schemes are
used as an input signal.
B. Problem Statement and Contribution
Most Common objectives of electricity management in SG
are PAR reduction, minimization of EC, UC maximization and
minimize the power consumption. A vast amount of electricity
is spent by residential area and its consumption is growing
speedily. To overcome these issues, we proposed an efficient
HEM. We have used GA and FPA for optimization in our
proposed solution. Results demonstrate that in terms of EC,
GA and FPA techniques performed well with respect to un-
scheduling.The part of a system that connects the consumers of
low voltage from high voltage transmission system is called
distribution system. During primary and secondary distribu-
tion, loss incidence ratio is up to 70%. While remaining 30%
lost in transmission lines [24].
In our work, GA and FPA are used to compare the perfor-
mance with un-scheduling EC. We considered only one home
from residential area, this home contains different appliances.
Each appliances has its own TOU and OTI. Appliances are
categorized into three different categories: interruptible appli-
ances, fixed appliances and non-interruptible appliances, more
details are in TABLE II, III and IV. This category of appliances
has taken from [5] . The main objectives of our work are
•EC minimization
TABLE I: Table for Related Work
Techniques Pricing
Schemes
Objectives Achievements Limitations
GA [1] RTP Minimization in cost and
PAR
Minimization of PAR UC ignore and cost cannot
reduce
GA [2] TOU Reduce the power utiliza-
tion
Minimize cost and PAR UC ignored
NSGA and FDM [3] RTP High energy residential ap-
plication and integration of
RES
Efficient integration of
RES and increase UC
EC increase and the instal-
lation and maintenance cost
of RES ignored
Small HEM [4] TOU Reduction in EC Minimize EC UC ignored and not explain
how much EC is minimized
MOPSO and DR program
[7]
DA-RTP
and TOU
Operational costs and pollu-
tion emissions
Minimize cost and emis-
sions with RES
UC, RES installation and
maintenance cost ignored
Smart charging and appli-
ance scheduling [9]
CPP Residential Area Cost Minimization, PAR
reduction
Lack of Initial Installation
and maintenance Cost of
batteries and UC
Efficient GA Based DSM
[10]
DA-RTP
and TOU
Covered commercial, resi-
dential and industrial areas
PAR reduce and cost min-
imize
Larger appliance delay and
UC ignored
Queuing-based Energy
Consumption (QEC)
monitoring for different
smart homes [11]
- Residential SG Networks Delay reduction and cost
minimization
Do not consider PAR and
RES
Delay and energy consump-
tion analysis [12]
RTP Low duty cycle data rate,
and energy consumption
Minimum energy
consumption
Cannot consider PAR and
delay
Try to meet the power grid
challenges [13]
RTP Stat of the art work EMC
for the residential energy
management system for
avoiding peak formations
PAR reduced and less cost UC ignored
LP, GA, and TLBO [14] - Cost minimization and UC
maximization
Minimize electricity con-
sumption cost
Do not bother PAR
Deliberate two different sys-
tems to explore the incre-
mental price [15]
- Multi-objective
optimization agenda
Multi-objective optimiza-
tion framework
Work for short term
scheduling
Manageable load modelling
[16]
RTP GSO for multiple object and
constrains
New problem of Shift able
load proposed
PAR and UC ignored
Algorithm for load control
[17]
- Less load and power
scheduling
Schedule different type of
appliances
UC and PAR not considered
Game theoretical approach
[18]
DA-RTP Less PAR and EC Lessen PAR and EC High communication over-
head and fail to protect cus-
tomer privacy
•UC maximization
•PAR reduction
•Find out trade-off between UC and EC
III. SYS TE M MOD EL
Heuristic techniques are problem dependent and these can
be used to boost/speed up the process of finding a satisfactory
solution from all possible solutions. This solution may be close
to the best one. However, these techniques never guarantees
you for the best solution.
A. Categorization of Appliances
Appliances can be divided into different categories, in our
case, we divide the appliances into three categories.
•Fixed appliances are those appliances that can not be
shifted from their position, in terms of ON/OFF status.
•Non-interruptible appliances can be shifted from one
position to another. However, once this type of appliance
is ON, it can not be interrupted until it done its work.
•Interruptible appliances can be shifted or interrupted at
any time.
In our case, the fixed appliances category contains (refriger-
ator, Television (TV) and telephone), non-interruptible appli-
ances category contains (Air Conditioner (AC) and lighting),
interruptible appliances contains (printer, hair straightener,
desktop computer, oven, cooker hood, microwave, iron, toaster,
kettle, other fixed, washing machine, dishwasher and hair
dryer).
We tackle the problem of cost by considering one home.
However, it can be applied over several homes. We considered
15, 30 and 60 minute OTI for each of the appliance. OTI varies
from problem to problem. Appliances can be more or less than
that of our case scenario. Length of Operation Time (LOT)
can also vary. The power rating of appliances represented as
follows
Residential Area
HEM
SMART METER
UTILITY
CPP
Electric vehicle Electric vehicle
Electric vehicle
Fig. 1: Proposed System Model
TABLE II: Fixed Appliances
Fixed Appliances Power Rating (kWh)
Refrigerator 1.666
TV 0.083
Telephone 0.005
TABLE III: Non-interruptible Appliances
Non-interruptible appliances Power Rating (kWh)
Air Conditioner 1.14
Lighting 0.1
TABLE IV: Interruptible Appliances
Interruptible appliances Power Rating (kWh)
Printer 0.011
Hair Straightener 0.055
Desktop Computer 0.15
Oven 1.14
Cooker Hood 0.225
Iron 2.4
Microwave 1.2
Toaster 0.8
Kettle 2
Other Fixed 0.05
Washing Machine 1.4
Dishwasher 1.32
Hair Dryer 1.8
IV. PROPOSED METHODOLOGY
GA is nature inspired algorithm which is based on the
theory of Darwin. It was invented by John Holland in the
middle 70s. We can use it for stochastic nature of problems or
when changes are adopted. GA works on the basis of genetic
population like chromosomes [21]. GA is a searched-based
optimization technique. It leads the problem towards its opti-
mal solution. The word optimization means moving towards
best solution. Optimization definition varies from problem to
problem. In our scenario, we use this optimization for cost and
PAR reduction in the HEM. GA applies on a set of population,
it finds the best/fittest solution for this population and all other
population moving towards that solution (Algorithm 1).The
same method was expanded by Agnetis et al., in [22], by using
heuristic approach for household energy consumption, climate
comfort level and timeliness.
A. GA
Selection of the parent (chromosomes) may follow one of
the following procedure
•Roulette wheel
•Tournament
•Truncation selection
•Fittest
•By itself. Pick the best (in our scenario)
Algorithm 1 GA
1: initialization
2: set lower and upper limits
3: for appliances a belong to A do
4: for generate population i=1 to maxpopsize do
5: for generate appliances j=1 to maxappsize do
6: generate initial population
7: evaluate each individual‘s fitness
8: while generation <max generation do
9: select two parents
10: for randomness in population do
11: crossover
12: mutation
13: end for
14: evaluate ending criteria
15: elitism (process of following the
16: best ones)
17: end while
18: end for
19: end for
20: end for
21: return best possible solution
22: end
1) Crossover: Types of crossover
•One-point crossover
•Two-point crossover (in our scenario)
•Uniform crossover
2) Mutation: Types of mutation
•Substitution
•Insertion
•Deletion
•Inversion
B. FPA
FPA is also nature-inspired algorithm and it is developed
by Xin-She Yang in 2012 [19]. It is one of the most recent
developed algorithm. The idea of this technique is taken from
the pollination process of the flowers. The main purpose of a
flower is ultimately reproduction via pollination (Algorithm 2).
The author claims in its paper that FPA performs better than
GA and Partial Swarm Optimization (PSO). Its main objectives
are
•Survival of the fittest
•Optimal reproduction of the plant
There are two types of pollination that are usually done,
a-biotic and biotic. 90% of the plants fall in the category
of biotic while other 10% plants fall in a-biotic. In global
pollination (biotic), pollinators such that animals, bats and
birds are required to transfer the pollens. In local pollination
(a-biotic), wind and diffusion in water caused pollination.
Complete details of FPA are given in Algorithm 1.
1) Basic Steps for FPA:
•Global pollination by using Levy distribution formula
Algorithm 2 FPA
1: initialization
2: for all appliances a belong to A
3: for generate population k=1 to maxpopsize do
4: generate random flowers (population)
5: for generate appliances m=1 to maxappsize do
6: if rand() >probability switch then
7: use levy-flight formula for updation
8: else
9: select random population
10: check limits
11: end if
12: generate random population
13: calculate fitness for each individual
14: end for
15: find local best
16: if new solution is better than previous then
17: update solution
18: end if
19: update the global best solution
20: end for
21: return best solution
22: end
•Local or self pollination
•Reproduction uses flower consistency by considering the
similarity of two flower that are are used in pollination.
•Control switching probability for selecting local and
global pollination.
C. GFPA
GFPA is a hybrid of GA and FPA. We simply add the
crossover and mutation steps from the GA to the FPA (Al-
gorithm 3). More details are discussed in the section of
simulation and discussion.
V. SIMULATION AND DISCUSSION
In this section, we will discuss simulation results in detail.
For showing productiveness of our proposed work. We have
conducted simulations, in order to show optimal scheduling
for single smart home. We have done our scheduling task on
a single home. However, this task (scheduling) can be done
over multiple smart homes. We have successfully implemented
GA and FPA for applying scheduling on a single smart
home. The time schedule is composed of 96, 48 and 24-time
slots respectively in a day, starting from 12 am to 12 am.
Simulations result of our load are shown in Fig. 2
We have applied GA, FPA and GFPA on different OTIs, i.e.
15, 30 and 60. Every OTI generates different results. PAR is
directly proportional to the price (greater the PAR greater the
cost) and vice versa. It can be easily judged that GA, FPA
and GFPA in every OTI has less PAR than unscheduled PAR
(Fig. 3). So, GA, FPA and GFPA performs better for PAR.
Our proposed schemes shuffled appliances and turned them
Algorithm 3 GFPA
1: initialization
2: for all appliances a belong to A
3: for generate population k=1 to maxpopsize do
4: generate random flowers (population)
5: for generate appliances m=1 to maxappsize do
6: if rand() >probability switch then
7: use levy-flight formula for updation
8: else
9: select random population
10: check limits
11: end if
12: generate random population
13: calculate fitness for each individual
14: end for
15: find local best
16: if new solution is better than previous then
17: update solution
18: end if
19: update the global best solution
20: for randomness in population do
21: crossover
22: mutation
23: end for
24: end for
25: return best solution
26: end
on during a beneficial time and also keep avoid to generating
peaks. PAR can be easily understand by the Fig. 3
There is a trade-off (compromise) between UC and cost. If a
person wants more comfort, he has to pay more cost and if he
took less comfort, he has to pay less cost. Unscheduled waiting
time never exists in our scenario, since consumer can turn on
appliances at any time. However, by applying GA, FPA and
GFPA‘s proposed solution, the user has to wait for a proper
time and as a reward he has to pay less bill. There are three
scenarios that we have discussed in our scheduling. If we take
bigger OTI, it increases the waiting time. Similarly, if some
appliance stop before its total running time, then remaining
time is wasted. To avoid the wastage of this time, we work for
different OTIs. Consumers can choose any OTI for scheduling
according to their comfort. Delay against different OTIs shown
in Fig. 4
GA, FPA and GFPA are also performed better while we
discuss the total cost, as these schemes decrease the total cost
as compared with the unscheduled total cost. Our proposed
schemes reduce cost slot by slot (15, 30 and 60) and as a
result of this work, hourly cost is minimized and hourly cost
reduction means minimization in daily cost and so on. The
Fig. 3 shown comparison between cost of GA, FPA, GFPA
and unscheduled cost. It is worth mentioning here that the
cost pattern of our suggested scheduling techniques are quite
optimal as compared to the unscheduled load. By shifting
2 4 6 8 10 12 14 16 18 20 22 24
Time (slots)
0
2
4
6
8
10
12
14
Load (kWh)
Unsheduled
GA
FPA
GFSA
(a) OTI 60 minutes
5 10 15 20 25 30 35 40 45
Time (slots)
0
2
4
6
8
10
12
14
Load (kWh)
Unsheduled
GA
FPA
GFSA
(b) OTI 30 minutes
10 20 30 40 50 60 70 80 90
Time (slots)
0
2
4
6
8
10
12
14
Load (kWh)
Unsheduled
GA
FPA
GFSA
(c) OTI 15 minutes
Fig. 2: Load against different OTIs
the load consumer comfort effects. However, it benefits the
consumer in terms of cost reduction. The more a consumer
shifts the load and tolerates changes in the energy consumption
pattern, the more benefit will be provided by the utility in form
of cost reduction.
A. Performance Trade-off
There is a trade-off or compromise between the cost and
waiting time, if consumer wants less bill, he has to sacrifice
on his comfort. If consumer can not wait for proper time
(suggested by algorithm) and turned on appliances at any time,
he has to pay more bill. As, there is indirect relation between
cost and delaying time. So, consumer can choose only one
of them, either less cost or less comfort. We compared EC
and UC by applying GA, FPA and GFPA with unscheduled
OTI 60 minutes OTI 30 minutes OTI 15 minutes
1
2
3
4
5
6
7
8
9
PAR
Unscheduled
GA
FPA
GFPA
Fig. 3: PAR against different OTIs
OTI 60 minutes OTI 30 minutes OTI 15 minutes
0
0.5
1
1.5
2
2.5
3
3.5
Cost (cents)
×103
Unscheduled
GA
FPA
GFPA
Fig. 4: Cost against different OTIs
OTI 60 minutes OTI 30 minutes OTI 15 minutes
2
4
6
8
10
12
14
16
18
20
Waiting time (hours)
GA
FPA
GFPA
Fig. 5: Delay against different OTIs
load. When these techniques are compared with each other
by considering the simulation results, GFPA perform better as
its cost is minimum in all cases. On the other hand, it takes
average waiting time and PAR, as compared to FPA and GA.
It generates some peak by shifting the load from off-peak to
on-peak hours and price is still minimum.
B. Feasible Region
A region that covers all possible solutions according to our
fitness function. In our scenario we target the EC and PAR,
by means of minimization. The EC is based upon two factors:
electricity price and electricity consumption. The notable thing
is that, we only have authority while we discuss electricity
price. However, we can handle electricity consumption by
shifting the load from on-peak slots to off-peak slots. During
calculation of EC, there are four parameters that we have to
be follow. Feasible regions are shown in Fig. 5
•Minimum price, minimum energy consumption
•Minimum price, maximum energy consumption
•Maximum price, minimum energy consumption
•Maximum price, maximum energy consumption
The Pointers (P1, P2, P3, P4, P5) have shown the possible
feasible region against different OTIs. The area of feasible
region is shaded with cyan color. For more depth details,
follow the Fig. 6.
0 5 10 15
Power consumption (kWh)
0
500
1000
1500
2000
Cost per hour (cents)
P1
P2
P3
P4
P5
P6
(a) OTI 60 minutes
0 2 4 6 8 10 12 14
Power consumption (kWh)
0
500
1000
1500
Cost per hour (cents)
P1
P3
P6
P5
P4
P2
(b) OTI 30 minutes
0 2 4 6 8 10
Power consumption (kWh)
0
200
400
600
800
1000
1200
1400
Cost per hour (cents)
P1
P2
P3
P4
P5
P6
(c) OTI 15 minutes
Fig. 6: Feasible regions against different OTIs
VI. CONCLUSION
In this paper, we proposed DSM scheme using HEM
constructed on GA, FPA and hybrid of these two schemes
to minimize users EC with a reasonable delay. To show the
effectiveness of our proposed scheme we take simulations for
single home by using different OTIs and this home consists of
18 appliances. Each appliance has its own power rating. The
optimization procedure is assessed on the base of EC, PAR
and UC. The UC is dignified in terms of waiting time. The
simulation results show the efficiency of our proposed GFPA
performs even better than GA and FPA in terms of EC, UC
and PAR reduction.
REFERENCES
[1] Sahar Rahim, Nadeem Javaid, Ashfaq Ahmad, Shahid Ahmed Khan,
Zahoor Ali Khan, Nabil Alrajeh, and Umar Qasim. ”Exploiting heuristic
algorithms to efficiently utilize energy management controllers with
renewable energy sources.” Energy and Buildings 129 (2016): 452-470.
[2] Zhuang Zhao, Ziming, Jie Tang, Sangarapillai Lambotharan, Woon
Hau Chin, and Zhong Fan. ”An integer linear programming based
optimization for home demand-side management in smart grid.” In
Innovative Smart Grid Technologies (ISGT), 2015 IEEE PES, pp. 1-
5. IEEE.
[3] Pedram Samadi, 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 7.4 (2016): 1802-
1812.
[4] Abushnaf, Jamal, Alexander Rassau, and Wodzimierz Grnisiewicz. ”Im-
pact on electricity use of introducing timeofuse pricing to a multiuser
home energy management system.” International Transactions on Elec-
trical Energy Systems 26, no. 5 (2016): 993-1005.
[5] Erdinc, Ozan. ”Economic impacts of small-scale own generating and
storage units, and electric vehicles under different demand response
strategies for smart households.” Applied Energy 126 (2014): 142-150.
[6] Sereen Althaher, Pierluigi Mancarella, and Joseph Mutale. ”Automated
demand response from home energy management system under dynamic
pricing and power and comfort constraints.” IEEE Transactions on Smart
Grid 6, no. 4 (2015): 1874-1883.
[7] Bharathi, Rekha, and Vijayakumar. ”Genetic Algorithm Based Demand
Side Management for Smart Grid.” Wireless Personal Communications
93, no. 2 (2017): 481-502.
[8] Bin Zhou, Wentao Li, Ka Wing Chan, Yijia Cao, Yonghong Kuang,
Xi Liu, and Xiong Wang. ”Smart home energy management systems:
Concept, configurations, and scheduling strategies.” Renewable and
Sustainable Energy Reviews 61 (2016): 30-40.
[9] Mohammad Rastegar, Mahmud Fotuhi-Firuzabad, and Hamidreza
Zareipour. ”Home energy management incorporating operational priority
of appliances.” International Journal of Electrical Power & Energy
Systems 74 (2016): 286-292.
[10] Adika, Christopher, and Lingfeng Wang. ”Smart charging and appli-
ance scheduling approaches to demand side management.” International
Journal of Electrical Power & Energy Systems 57 (2014): 232-240.
[11] Muhammad Awais, Nadeem Javaid, Nusrat Shaheen, Zafar Iqbal, Ghaz-
anfar Rehman, Muhammad, and Ahmad. ”An efficient genetic algorithm
based demand side management scheme for smart grid.” In Network-
Based Information Systems (NBiS), 2015 18th International Conference
on, pp. 351-356. IEEE, 2015.
[12] Liu Yi, Chau Yuen, Rong Yu, Yan Zhang, and Shengli Xie. ”Queuing-
based energy consumption management for heterogeneous residential
demands in smart grid.” IEEE Transactions on Smart Grid 7, no. 3
(2016): 1650-1659.
[13] Muhammad Babar Rasheed, Nadeem Javaid, Muhammad Imran, Zahoor
Ali Khan, Umar Qasim, and Athanasios Vasilakos. ”Delay and energy
consumption analysis of priority guaranteed MAC protocol for wireless
body area networks.” Wireless networks 23, no. 4 (2017): 1249-1266.
[14] Mohammad Rastegar and Mahmud Fotuhi-Firuzabad. ”Outage manage-
ment in residential demand response programs.” IEEE Transactions on
Smart Grid 6, no. 3 (2015): 1453-1462.
[15] Babak Yousefi-khangah, Saeid Ghassemzadeh, Seyed Hossein Hosseini,
and Behnam Mohammadi-Ivatloo. ”Short-term scheduling problem in
smart grid considering reliability improvement in bad weather condi-
tions.” IET Generation, Transmission & Distribution 11, no. 10 (2017):
2521-2533.
[16] Ogwumike, Chris, Michael Short, and Mouloud Denai. ”Near-optimal
scheduling of residential smart home appliances using heuristic ap-
proach.” In Industrial Technology (ICIT), 2015 IEEE International
Conference on, pp. 3128-3133. IEEE, 2015.
[17] Ye Feng, Yi Qian, and Rose Qingyang Hu. ”A real-time information
based demand-side management system in smart grid.” IEEE Transac-
tions on Parallel and Distributed Systems 27, no. 2 (2016): 329-339.
[18] Nguyen, Hung Khanh, Ju Bin Song, and Zhu Han. ”Distributed demand
side management with energy storage in smart grid.” IEEE Transactions
on Parallel and Distributed Systems 26, no. 12 (2015): 3346-3357.
[19] Xin-She Yang. ”Flower pollination algorithm for global optimization.”
In UCNC, pp. 240-249. 2012.
[20] Nigdei, Bekda, and Xin-She Yang. ”Optimum tuning of mass dampers
for seismic structures using flower pollination algorithm.” International
Journal Theor. Appl. Mech 1 (2017): 264-268.
[21] Whitley, Darrell. ”A genetic algorithm tutorial.” Statistics and computing
4, no. 2 (1994): 65-85.
[22] Agnetis, Alessandro, Gianluca de Pascale, Paolo Detti, and Antonio
Vicino. ”Load scheduling for household energy consumption optimiza-
tion.” IEEE Transactions on Smart Grid 4, no. 4 (2015): 2364-2373.
[23] Evangelisti, Lettieri, Clift, and Borello. ”Distributed generation by
energy from waste technology: a life cycle perspective.” Process Safety
and Environmental Protection 93 (2015): 161-172.
[24] Prasad Reddy, Dinakara Prasad, Veera Reddy and Gowri Manohar.
”Application of flower pollination algorithm for optimal placement and
sizing of distributed generation in distribution systems.” Journal of
Electrical Systems and Information Technology 3, no. 1 (2016): 14-22.
[25] Sheraz Aslam, Zafar Iqbal, Nadeem Javaid, Zahoor Ali Khan, Khursheed
Aurangzeb, and Syed Itaza Haider. ”Towards Efficient Energy Manage-
ment of Smart Buildings Exploiting Heuristic Optimization with Real
Time and Critical Peak Pricing Schemes.” Energies 10, no. 12 (2017):
2065..
