Cost and Load Reduction Using Heuristic Algorithms in Smart Grid

Abstract

Due to smart grid applications the consumers and producers are able to meet the demand of each others and thus take part in demand side management and demand response program. Hence smart grid leads to optimization of energy consumption and reduce high cost in today extensive demand of energy. In this research work we are reducing electricity consumption cost and load consumption using scheduling the appliances. The twenty appliances are used to schedule their energy consumption and load using heuristics techniques i.e. binary particle optimization, genetic algorithm and wind driven optimization, using the same data set for each technique and their results are compared with each other in order to find which technique do better optimization. Simulations are performed in matlab to show the cost and load reduction by the above three techniques and validate the experiment. The simulation results show that binary particle swarm optimization perform better than the other two techniques and wind driven optimization is better than genetic algorithm but not able to perform as binary particle swarm optimization, similarly genetic algorithm is least efficient as compared to both methods. Our research work is beneficial to meet the demand side management and help in reducing electricity cost and load for consumers.

Full text

Cost and Load Reduction using
Heuristic Algorithms in Smart Grid
Zafar Iqbal1, Nadeem Javaid2,∗, Mobushir Riaz Khan1,
Imran Ahmed3, Zahoor Ali Khan4,5, Umar Qasim6
1Pir Mehr Ali Shah Arid Agriculture University, Rawalpind 46000, Pakistan
2COMSATS Institute of Information Technology, Islamabad 44000, Pakistan
3Institute of Management Sciences (IMS), Peshawar 25000, Pakistan
4Faculty of Engineering, Dalhousie University, Halifax, NS, B3J 4R2 Canada
5Higher Colleges of Technology, Fujairah 4114, United Arab Emirates
6University of Alberta Edmonton, Alberta T6G 2J8, Canada
∗Corresponding author: nadeemjavaidqau@gmail.com, www.njavaid.com
Abstract—Due to smart grid applications the consumers and
producers are able to meet the demand of each others and thus
take part in demand side management and demand response
program. Hence smart grid leads to optimization of energy
consumption and reduce high cost in today extensive demand
of energy. In this research work we are reducing electricity
consumption cost and load consumption using scheduling the
appliances. The twenty appliances are used to schedule their
energy consumption and load using heuristics techniques i.e.
binary particle optimization, genetic algorithm and wind driven
optimization, using the same data set for each technique and
their results are compared with each other in order to find which
technique do better optimization. Simulations are performed in
matlab to show the cost and load reduction by the above three
techniques and validate the experiment. The simulation results
show that binary particle swarm optimization perform better
than the other two techniques and wind driven optimization is
better than genetic algorithm but not able to perform as binary
particle swarm optimization, similarly genetic algorithm is least
efficient as compared to both methods. Our research work is
beneficial to meet the demand side management and help in
reducing electricity cost and load for consumers.
Keywords:-Appliance Scheduling, Demand Side Manage-
ment, Demand Response Program, Cost Reduction, Load
Reduction, Binary Particle Swarm Optimization, Wind Driven
Optimization, Genetic Algorithm, Smart Grid.
I. INTRODUCTION
The world is now moving toward a high technological
era due to advancement in information and communication
technologies. Among these technologies, Smart Grid (SG) is
the advanced form of traditional grid due to incorporation of
two way communication network to facilitate both end users
and utility. SG is helpful in providing intelligent monitoring,
control system, communication and self-healing services to
consumers [1].
Smart grid have some general features which include; (1)
Reliability, SG has the ability to cope with faults and provides
self-healing mechanisms which shows its reliability and less
vulnerability to some natural disaster or intruder attacks,
(2) Flexibility in network topology, its network topology is
very flexible and it allow bidirectional information flow, (3)
Efficiency, it provides flexibility in implementation of different
demand side management and demand response programs
to manage end user energy demand, (4) Peak curtailment,
as the prices of electricity changes throughout the day due
to energy demand variations and there are chances of high
peaks during low pricing hours which can disturb the stability
of electric grid. Through the implementation of different
energy management techniques, load can be scheduled in
order to avoid high peaks, (5) Stainability, it has the ability
to sustain even if greater amount of other distributed power
sources penetrate like solar power, wind power etc, (6) Market
enabling, through mart metering infrastructure, smart grid
provides greater flexibility to suppliers and consumers to
sell extra energy back to grid, (7) Demand response, smart
grid provides support to consumers in the form of demand
response programs, where supplier or generators can estimate
the end user energy demands at particular time slots to manage
electricity generation. (8) Platform for advance services i.e. it
also provides some advance services like fire monitoring and
alarm which can shut off power system and emergency call
service [2].
Now a days energy crises are everywhere and the expensive
energy production mechanisms are harder to use in a man-
ageable way. Along with energy demand management, SG
minimizes the emission of greenhouse gases and make the
chances of global warming less to occur. In this work we con-
sider twenty appliances, their power rating, price signal, and
length of operation time to evaluate the energy consumption
and electricity bill. We have done cost and load optimization
in this work,and have used three evolutionary and heuristic
techniques to calculate cost and load for these appliances.
These techniques are binary particle swarm optimization,
genetic algorithms and wind driven optimization.
We have evaluated the performance of each technique and
compare their values with each other in order to find which

techniques perform better in term of cost and load optimiza-
tion. For this purpose to calculate the exact values and find
accurate comparison among these techniques we have taken
the same power rating, same price signal, same time slots,
length of operation time, same appliances bits pattern,same
number of appliances and same values for all the parameters
for all these twenty appliances and feed them to these three
techniques one by one and generated the results and then these
results are compared as shown in a table V and VI, in the form
of scheduled and unscheduled cost and load, their total cost,
total load reduction, their total percentage cost and percentage
load reduction. If we see the table V and VI it is shown that
Binary particle swarm optimization perform better than other
two techniques.
The rest of the paper is organize as follows.sectiob II
describes the existing work,motivation of idea is in section
III and section IV compare the optimization techniques used
in this paper.Cost and load optimization is described in section
V,and section Vi discusses the Simulation results,and finally
section VII provides conclusion.
II. RE LATE D WOR K
Some of the existing work related to our work are listed
below. In [3] the authors are using binary particle swarm
optimization for solving the optimization problem and re-
ducing electricity bill. The authors proposing home energy
management system with RES, without RES, with smart
scheduler, without smart scheduler,with smart scheduler plus
RES and conventional system. The consumers are classified
into three classes i.e. conventional users, smart users and
prosumers, scheduling is performed via BPSO technique to
reduce cost and load. In [4] the authors divide the appliances
into categories and proposing an optimal scheduling algorithm
and using binary particle swarm optimization technique to
reduce electricity cost. They consider supplier and consumer
scenario also with renewable energy resources. They consider
the personal habits and characteristics of appliances and draw
a table of time of use electricity prices for appliances operation
with the objective to reduce cost. When the power demand is
high or low the supplier will get a signal and will ask the
consumers to on/off some appliances or switch off some high
power consuming appliances so they will be able to achieve
load shifting and thus avoid peak hours cost.In [5] the authors
use particle swarm optimization for scheduling the operation
of distributed energy resources. They use decision support tool
to optimize energy services of residential end users. In [6]
the authors used PSO and DER for scheduling the energy
services, they use different cases for scheduling and optimizes
cost. They also use case study and decision support tool to
optimize energy services to smart homes.In [7] the authors
present a scheduling strategy for cost optimization and de-
mand response management using particle swarm optimization
technique to simulate their different scheduling scenarios. In
[8] the authors use non cooperative game theoretic model for
cost optimization using scheduling during power outages. In
[9] the authors use binary particle swarm optimization for
home energy management and appliance scheduling. They
have taken six appliances for their test system and simulated
them via matlab. They generated load and cost plots for sched-
uled and unscheduled cases. In [10] the authors use particle
swarm optimization and binary particle swarm optimization for
scheduling of different scenario as shown in Table.1,in their
work to optimize electricity cost and reduce load consumption.
They schedule DER with no battery storage, with battery
storage, with net feed-in tariff and no battery storage, with net
feed-in tariff and battery storage. They also used case study
for energy service provisioning in a smart home, and thus via
scheduling they achieve cost and load optimization. In [11]
the authors use APSO a variant of PSO to schedule the Power
consumption or charging of two PHEV-1 and PHEV-2.They
have taken some scenarios for scheduling from Monday to
Friday and shift the Peak load demand to off peak hours
and thus reducing total energy consumption cost. They are
using grid power, solar power for charging and ToU pricing
scheme. They have calculated hourly unit power price, hourly
solar power generation, hourly regular power demand, hourly
power demand by PHEV-1, hourly power demand by PHEV-2,
hourly combined power demand, hourly grid power demand,
hourly grid power price through simulation and thus show
optimization in cost via scheduling. In [12] the authors use
genetic algorithm to optimize cost and scheduled appliances
in a smart home. They schedule their appliances in the home
for the purpose of reducing electricity expenses and to reduce
Peak to Average Ratio (PAR).They have introduced general
architecture of energy management system in a smart home
and then propose an efficient scheduling method for home
power usage. They have used RTP and IBR pricing model, by
using this combine model their proposed scheduling strategy
has efficiently reduced cost and peak to average ratio. In [13]
the authors use genetic algorithm to find the optimum schedule
arrangement for all the energy consumption by appliances in a
smart home to reduce the electricity cost. They used intelligent
task scheduling module to minimize the entire energy expense
in a smart home, if the module could schedule the appliances
start time. They used the genetic algorithm approach for
minimizing the residential total electricity cost in demand
response services. Their approach consider task or appliance
on /off time constraints and the circuit maximum load con-
straints. Further they compared their approach with other two
techniques i.e. SA and greedy method and showed via sim-
ulation that their technique has obtained optimal scheduling
solution for residential customers as well as satisfying the
equipments operation time constraints and the entire power
load constraints. Hence their simulations prove that genetic
algorithm can efficiently reduce electricity cost for residential
customers.In [14] the authors in this work proposes NSGA-2
a genetic based algorithm to solve the optimization problem.
The optimization problem is to reduce the total energy cost and
peak to average ratio of the total energy demand. They have
taken strict energy consumption scheduling and soft energy
consumption scheduling for appliances and generate plots for
scheduled energy consumption and their corresponding cost

without optimization and scheduled energy consumption and
corresponding cost with optimization and thus reduces total
consumption cost and peak to average ratio. In [15] the authors
used genetic algorithms for the retailers to maximize the
profit. They have scheduled the appliances and the retailer use
the appliance scheduling information to maximize its profit
by solving the profit maximization problem. They have also
used other techniques besides i.e. Stackelberg Game Approach
and linear programming approach besides genetic algorithm
maximize profit and reduce cost and peak to average ratio.
They have divided 24 hours time slot in to three groups i.e.
on-peak hours (5 PM-12 PM), mid-peak hours (8 AM-5 PM)
and off-peak hours (12 PM-8 AM).
III. MOTI VATI ON
Energy crises and problems are everywhere in the world
and increasing day by day, Hence experts, scientists and
researcher are constantly trying to develop and explore new
ways of energy production and safe utilization. Researchers
and scientists are trying to develop such ways to minimize
Energy wastage and also provide less expensive and smart
energy resources with a minimum customer cost. Hence costly
energy production and costly distribution to consumers as well
as wastage of this costly energy is a big problem these days for
every country unless they did not formulate a special strategy
or design a special solution to mitigate these issues. Hence
smart grid is one such solution of the many existing solutions
for providing less expensive energy to consumers, preventing
wastage of energy and smart utilization and smart distribution
of energy is provided. In short SG means computerizing
the electric grid or conventional power grid into smart one,
so it can perform smart functions like minimize wastage of
energy during distribution to consumers, reduce electricity
cost, maximize user comfort, minimize user frustration level.
The customers are always novice users and are not aware of
demand side management or DR program so they are unable
to use energy according to their use and thus wastage of
energy occurs and they always pay more cost what they have
used. Hence by the use of smart grid the above issue can be
solved or minimize to a maximum extent. In this work we are
taking the similar motivation of electricity cost, electricity bill
reduction and load reduction. We are using three biologically
inspired evolutionary techniques in our work for electricity
cost reduction and load optimization and reduction. i.e. binary
particle swarm optimization (BPSO), genetic algorithms (GA),
and wind driven optimization (WDO).These three optimization
techniques are used to optimize or reduce electricity bill and
electricity load i.e. to generate scheduled and unscheduled
cost, scheduled and unscheduled Load. For this purpose simi-
lar data set i.e. total twenty appliances which consume wattage
above five hundreds watts, having same price signal, same
power rating, and having the same time slots and length of
operation time (LOT) are taken for each of the above three
techniques mentioned in Table.2, to generate scheduled and
unscheduled cost and load results, ToU pricing, and these
results are then compared as shown in Table.2. And explained
and analyzed in section simulation results and discussion.
Time (hours)
0 5 10 15 20
Cost (Rs)
5
10
15
20
25
30
ToU Pricing
Fig. 1. TOU Pricing Signal
IV. COS T AN D LOAD OPTIMIZATION
Cost Reduction:-The purpose of our work is to achieve cost
and load optimization using evolutionary techniques. These
techniques have reduced the cost and load successfully.
Objectives:-The objective of our work is to reduce electricity
bill and load consumption and as the energy these days are
very costly and if we switch on our appliances during high
peak hours then it will add extra cost to our bill. We have
used three evolutionary techniques to reduce electricity cost
using the same data set for all appliances. The efficiency of
the three techniques are described below.
1. Most Efficient Technique.Binary particle swarm opti-
mization is and evolutionary and heuristics technique which is
used in many fields for optimization. Kennedy and Ebarhart
in 1995 developed the particle swam algorithm by studying
the social and cognitive behavior of ants. The individuals or
objects called particles are flown through a multidimensional
search space [16]. The scheduled cost and unscheduled cost
generated by BPSO for twenty appliances are Rs 1454.8 and
Rs 2341.6 respectively, and the total reduction in cost Rs 886.8
and total percentage reduction in cost is 37.88% as shown in
Table I. Hence BPSO cost reduction is more than the other
two techniques so it show best performance as compared to
other two techniques.
2. Efficient Technique. The wind driven optimization was
initially developed by Dr. Zikri Bayraktar during his graduate
studies at the Pennsylvania State University. The WDO is a
novel nature-inspired global optimization algorithm based on
atmospheric motion. It is a population based iterative heuris-
tic global optimization algorithm for multi-dimensional and
multi-model problems having ability to implement constraints
on the search domain. A population of very small air particles
navigate over an N-dimensional search space using second
law of motion, which is also used to explain the motion
of air parcels within the earth atmosphere.WDO add extra

TABLE I
COS T REDUCTION
Cost Category BPSO WDO GA
Scheduled Cost Rs 1454.8 Rs 1827.8 Rs 2077.8
Unscheduled Cost Rs 2341.6 Rs 2288.1 Rs 2341.6
Reduction in Cost Rs 886.8 Rs 460.3 Rs 263.8
%age Reduction in Cost 37.88% 20.2% 11.26%
term in velocity update equation i.e. gravitation and Coriolis
forces which provides robustness and extra degree of freedom
to fine tune the optimization. They claim that WDO can
perform better than PSO and that it is well-suited for problems
with both discrete and continuous-value parameters [17]. The
scheduled and unscheduled cost by WDO in our work is Rs
1827.8 and Rs 2288.1 and the total reduction in cost is Rs
460.3 and the percentage reduction in cost is 20.2%, which
show enough reduction in cost by WDO and as they claimed
but our simulation via Matlab show that WDO cost reduction
is less than BPSO as BPSO is 37.88% and WDO is 20.2%.
Hence in our scenario WDO is least efficient than BPSO as
shown in Table I.
3. Least Efficient Technique.Genetic algorithms were ini-
tially developed by John Holland in early 1970.Genetic al-
gorithm were developed to show some behavioral processes
observed in natural evolution. The idea with GA is to use
this power of evolution to solve optimization problems. we
have used genetic algorithm to optimize cost and load using
scheduling. We have taken twenty appliances and calculate
their scheduled and unscheduled cost. The scheduled and
unscheduled cost are Rs 2077.8 and Rs 2341.6 respectively,
and their difference is Rs 263.8 which is actually reduction in
cast after we run the scheduled load. The percentage reduction
in cast by GA is 11.26% which in our scenario is less from
both BPSO and WDO as shown in Table V below. Hence GA
in our scenario is least efficient in cost reduction as compared
to two other techniques.
TABLE II
NONLINEAR MOD EL RES ULTS
Case Method#1 Method#2 Method#3
1 50 837 970
2 47 877 230
3 31 25 415
4 35 144 2356
5 45 300 556
Load Reduction:-The second goal was to reduce or balance
load consumption by consumers and end users so that the
peak to average ratio could be maintained and there should
be balance load on the utility company and the consumers
will not have any problem from utility side, and demand side
management purpose will be achieved.
Objectives:-In this work our objectives was to optimize cost
and load consumption using below three evolutionary algo-
rithms and then compare the resulted values with each other to
find out which techniques perform best. If the load distribution
TABLE III
LOAD REDUCTION
Load Category BPSO WDO GA
Scheduled Load 118.0 kwh 148.2 kwh 153.2 kwh
Unscheduled Load 153.2 kwh 148.2 kwh 153.2 kwh
Reduction in Load 35.2 kwh 00.0 kwh 00.0 kwh
%age Reduction in Load 22.98% 00.0% 00.0 kwh
is uniform and consumption is balanced by consumers there
will be no high peaks and hence the peak to average ratio will
be small.It is the requirements of demand side management
and demand response program that peaks should be balanced
and the PAR should be less.
1. Most Efficient Techniques.As binary particle swarm op-
timization is explained above. We have used BPSO for load
reduction. The scheduled and unscheduled load by BPSO are
118.0 Kwh and 153.2 kWh respectively and their difference is
35.2 kWh which is actually reduction in load. The percentage
reduction in load by BPSO is 22.98% which show that BPSO
is better than the other two techniques, as shown in Table II.
2. Efficient Techniques. (WDO) wind driven optimization as
explained above are used for load reduction. The scheduled
and unscheduled load by wind driven optimization are 148.2
kWh and 148.2 kWh respectively and their difference is 00.0
Kwh which is actually load reduction. The percentage load
reduction by wind driven optimization is 00.0%. Hence wind
driven optimization is better than GA but not than BPSO in
our case as shown in Table II.
3. Least Efficient Techniques. Genetic algorithm as ex-
plained above is used for load optimization in our work. The
scheduled and unscheduled load by genetic algorithm is 153.2
kWh and 153.2kwh respectively and their difference is 00.0%
which is reduction in load. The percentage reduction in load
is also 00.0% which is similar to wind driven optimization.
Hence in this case the genetic algorithm shows least perfor-
mance as shown in Table II.
V. SIMULATION RESU LTS AND DISCUSSION
In this paper the three techniques discussed above are used
for which plots are explained and analyzed below.
The ToU pricing scheme plot is shown in fig.3 which show
cost in rupees and time slots in hours. The time slots curve
ranging from 10 to 20 for cost and 0 to 25 on x-axis for time
slots. The time of use pricing scheme is dynamic and its price
depends on time slots.It usually includes pricing depending on
very low peak hours, low peak hours, very high peak hours,
high peak hours, medium peak hours, and shoulder peak hours.
fig.4 show bar graph for scheduled and unscheduled load
and cost. The red bar show unscheduled load and unscheduled
cost which are 153.2 kWh and Rs 2341.6 and the blue bar
show scheduled load and scheduled cost which are 118.0
Kwh and Rs 1454.8 respectively as shown in Table V and
VI.Hence the difference between scheduled and unscheduled
load and scheduled and unscheduled cost is Rs 886.8 and 35.3
kwh respectively as shown in fig.4 and Table V and VI,which
show a great reduction in load and cost as evident from the

Time(hours)
0 5 10 15 20 25
Cost (Rs)
0
50
100
150
200
250
300
350
ScheduledCost
UnscheduledCost
(a) Scheduled and Unscheduled Cost of BPSO
Time(hours)
0 5 10 15 20 25
Energy consumption (KWh)
0
5
10
15
20
25
30
35
Scheduledload
Unscheduledload
(b) Scheduled and Unscheduled Energy Consumption of BPSO
Fig. 2. Electricity cost and energy consumption comparison of BPSO algorithm
bar graph. This reduction in load and cost after scheduling
is much greater than the other two techniques as shown in
Table V and VI and figure 4. Percentage reduction in cost is
37.88% and percentage reduction in load is 22.98% as can
be seen in bar graph and Table V and VI.37.88% reduction
in cost and 22.98% reduction in load for BPSO show that
this optimization techniques perform much better than the
other two techniques as can be seen in Table V and VI. And
fig.1,fig.2,fig.3,fig.4. The second optimization techniques used
in our work is genetic algorithm.fig.5 show the scheduled
and unscheduled cost curve. The red line is scheduled cost
and blue line is unscheduled cost.The time slots are taken in
hours and electricity cost are taken in rupees. The blue line or
unscheduled cost have only one sharp peak which show the
peak or maximum load at that time slot. The total scheduled
and unscheduled cost are Rs 2077.8 and Rs 2341.6 which
shows that unscheduled cost is much more than scheduled
cost and the difference between scheduled and unscheduled
cost is Rs 263.8 which is actually total reduction in cost. The
percentage reduction in cost is 11.26% by GA which is less
than 1/3 of the BPSO as shown in Table .V and VI.It show
that for our same data set and parameters GA perform very
poor as compared with BPSO as the percentage cost reduction
can be seen in Table V.
In fig.6 scheduled and unscheduled load curve are
shown.The red line show scheduled load and blue line show
unscheduled load. energy consumption is shown in Kwh and
time slots in hours. The blue line i.e. unscheduled load show
only one sharp peak and the rest of the curve are smooth.
The total scheduled and unscheduled load calculated by GA
are 153.2 kWh and 153.2 kWh respectively which are similar
and hence their difference will be 00.0 kWh and so the total
reduction in cost will also be Rs 00.0 and the percentage
reduction in load is also zero as shown in table VI.Hence GA
total load in our scenario is zero as compared to BPSO. fig.7
show the ToU pricing scheme for GA, which show similar
curve as for BPSO.The cost are shown in rupees and time
slots are shown in hours and the curve ranging from 15 to 20
on y-axis and from 0 to 25 on the x-axis which show total 24
hours time slots. fig.8 shows the bar graph of scheduled and
unscheduled load and cost for genetic algorithm. The cost are
shown as Rs/hour, the red bar show scheduled load and cost
and blue bar show unscheduled load and cost. The scheduled
load and cost of GA are 153.2 kWh and Rs 2077.8 respectively
while unscheduled load and cost are 153.2 kwh and Rs 2341.6
respectively, which show that scheduled and unscheduled load
of GA are same while cost have Rs 263.8 difference which
is actually total reduction in cost by GA which is less than
1/3 of the cost reduction by BPSO.Its percentage reduction in
cost is 11.26% and load is zero percent, which is significantly
less as compared to 37.88% of BPSO as shown in Table .V
and VI and fig.8.
fig.9 show scheduled and unscheduled cost curve. Time slots
is shown in hours and cost in rupees. scheduled cost is shown
by blue line and unscheduled cost by red line. The peaks are
shown as in fig.9.The scheduled and unscheduled cost are Rs
1827.8 and Rs 2288.1 respectively and their difference is Rs
460.3 which is actually total reduction in cost by WDO.The
percentage reduction in cost by WDO is 20.2%,which is less
than BPSO but better than GA.Hence WDO perform better
than GA for cost reduction.
fig.10 show load curve for scheduled and unscheduled load.
energy consumption is shown in watts and time slots are shown
in hours. Both the curve have very low peaks. The red line
show unscheduled load and black line show scheduled load.
The scheduled and unscheduled load for WDO are 148.2 kWh
and 148.2 kWh respectively which are similar hence their
difference is zero and same is reduction in load.WDO reduce
cost 20% but load reduction is zero percent which is similar
to GA. fig.11 show the plot of ToU pricing scheme for WDO
which show time slots in hours and cost in cents/kwh.The Total
time slots are 24 i.e. from 0 to 25.The wind driven optimization
ToU signal is similar to BPSO and GA and varies between Rs

Time(hours)
0 5 10 15 20 25
Electricity Cost (Rs)
0
50
100
150
200
250
300
350
400
450
ScheduledCost
UnscheduledCost
(a) Scheduled and Unscheduled Cost of GA
Time(hours)
0 5 10 15 20 25
Energy consumption (KWh)
0
5
10
15
20
25
30
ScheduledLoad
UnscheduledLoad
(b) Scheduled and Unscheduled Energy Consumption of GA
Fig. 3. Electricity cost and energy consumption comparison of GA
Time (Hours)
0 5 10 15 20 25
Electricity Cost (Cents)
0
50
100
150
200
250
300
UnscheduledCost
ScheduledCost
(a) Scheduled and Unscheduled Cost of WDO
Time (Hours)
2 4 6 8 10 12 14 16 18 20 22 24
Energy Consumption (Watt)
0
5
10
15
20
25
30
UnscheduledLoad
ScheduledLoad
(b) Scheduled and Unscheduled Energy Consumption of WDO
Fig. 4. Electricity cost and energy consumption comparison of WDO
10 and Rs 20. fig.12 show the bar graph of scheduled and
unscheduled load and cost. The cost are shown in rupees. The
red bar show unscheduled load and blue bar show scheduled
load. scheduled and unscheduled cost by WDO are Rs 1827.8
and Rs 2288.1 respectively which show the difference of Rs
460.3 which is reduction in cost and the percentage reduction
is 20.2%,as shown in bar graph and Table.V and VI.The
scheduled and unscheduled load by WDO are 148.2 kwh and
148.2 kwh which is same.
All simulation results and discussion show that BPSO is
best than GA and WDO in cost and load reduction while GA
perform much poor than the other two techniques as seen in
all the figures and Table V and VI.
VI. CONCLUSION
In this work we have used evolutionary and heuristics
techniques i.e. binary particle swarm optimization, genetic
algorithm and wind driven optimization for cost and load
optimization, using twenty appliances and using the same
data set i.e. power rating, electricity price signal, time slots
and length of operation time for the three techniques and
obtained different results by each, as the energy optimization
is a very critical issue these days, our purpose of this paper
is to minimize and reduce expensive energy cost and balances
the load consumption by residential consumers. We have also
compared the results of scheduled and unscheduled cost, load,
total reduction in cost, load and their percentage reduction
in cost and load by these techniques and found that which
techniques perform best for cost and load optimization. Our
simulation results and table V and VI show the total reduction
in cost and load and also show that BPSO perform the best as
compared to other two techniques. The %age cost reduction
by BPSO is 37.88%, WDO 20.2% and GA is 11.26%.our
simulation results show that we have achieved cost reduction,
and BPSO is best as its reduction is 37.88%. In future we
intend to use these techniques for more optimized values to

obtain and also reduce PAR value, user comfort maximization
and energy minimization.
REFERENCES
[1] http://www.iec.ch/smartgrid/background/explained.htm..
[2] https://en.wikipedia.org/wiki/Smart grid..
[3] Michael Angelo A. Pedrasa1,2, Ted D. Spooner1 and
Iain F. MacGill1,1School of Electrical Engineering and
Telecommunications, and Centre for Energy and Environmen-
tal Markets,University of New South Wales, Sydney, Aus-
tralia,2Electrical and Electronics Engineering Institute, University
of the Philippines, Quezon City, Philippines, “The Value of Accu-
rate Forecasts and a Probabilistic Method for Robust Scheduling
of Residential Distributed Energy Resources”, 978-1-4244-5721-
2/10/.
[4] Michael Angelo Pedrasa,Ted Spooner,Dr Iain MacGill,April
2010,“An Energy Service Decision-Support Tool for Optimal En-
ergy Services Acquisition”,Centre for Energy and Environmental
Markets,UNSW,the university of new south wales.
[5] Anil Kumar Pathak, Dr. S. Chatterji, Mahesh S. Narkhede, “Ar-
tificial Intelligence Based Optimization Algorithm for Demand
Response Management of Residential Load in Smart Grid”,
International Journal of Engineering and Innovative Technology
(IJEIT) Volume 2, Issue 4, October 2012,ISSN: 2277-3754,ISO
9001:2008 Certified.
[6] Amit S. Closepet,Spectrum Consultants, Bangalore, India,Email:
amitsclosepet@gmail.com, “Simple Real Time Non-Co-
Operative Game Theoritic Model for Energy Cost Optimization
in Developing Countries”, Journal of Power and Energy Engi-
neering, 2014, 2, 220-226,Published Online April 2014 in SciRes.
http://www.scirp.org/journal/jpee,http://dx.doi.org/10.4236/jpee.2014.24031.
[7] Zain Ul Abedin1, Uruj Shahid1, Anzar Mahmood1,Umar
Qasim2, Zahoor Ali Khan3, Nadeem Javaid1,1COMSATS
Institute of Information Technology, Islamabad,
Pakistan,2University of Alberta, Alberta, Canada,3CIS,
Higher Colleges of Technology, Fujairah Campus,
UAE,*www.njavaid.com, nadeemjavaidqau@gmail.com
“Application of PSO for HEMS and ED in Smart Grid ”,
http://www.researchgate.net/publication/275464014.
[8] M. A. Pedrasa, Student Member, IEEE, E. D. Spooner, and
I. F. MacGill,Centre for Energy and Environmental Markets
and School of Electrical Engineering and Telecommunications,
University of New South Wales, Sydney, Australia (email:
m.pedrasa@student.unsw.edu.au)., “Improved Energy Services
Provision through the Intelligent Control of Distributed Energy
Resources”.
[9] MOHD. MONIRUZZAMAN, KAMARUZZAMAN SOPIAN,
SALEEM H. ZAIDI,Solar Energy Research Institute
Universiti Kebangsaan Malaysia 43600 Bangi, Selangor
Malaysia,Mohd.Moniruzzaman@gmail.com, “Smart Appliance
Scheduling Scheme for Smart Meters using Adaptive Particle
Swarm Optimization”,Latest Trends in Renewable Energy and
Environmental Informatics,ISBN: 978-1-61804-175-3.
[10] Zhuang Zhao, Won Cheol Lee, Member, IEEE, Yoan Shin,
Senior Member, IEEE, and Kyung-Bin Song, “An Optimal Power
Scheduling Method for Demand Response in Home Energy Man-
agement System”, IEEE TRANSACTIONS ON SMART GRID,
VOL. 4, NO. 3, SEPTEMBER 2013.
[11] Hui Miao1, Xiaodi Huang2, Guo Chen3, “A Genetic Evolution-
ary Task Scheduling Method for Energy Efficiency in Smart
Homes”,International Review of Electrical Engineering (I.R.E.E.),
Vol. 7, N. 5,ISSN 1827- 6660 September - October 2012.
[12] Mahdi Mehrshad*1, Abdolreza Dehghani Tafti 2, Reza Effatnejad
3 Science and Research branch, Islamic Azad University, Al-
borz, Iran, *mahdimehrshad@gmail.com; 2 dehghani@kiau.ac.ir;
3 reza.effatnejad@kiau.ac.ir, “Demand-side Management in the
Smart Grid Based on Energy Consumption Scheduling by NSGA-
II”, International Journal of Engineering Practical Research
(IJEPR) Volume 2 Issue 4, November 2013 www.seipub.org/ijepr.
[13] Fan-Lin Meng,School of Computer Science,University of
Manchester,Manchester,UK,Email: mengf@cs.man.ac.uk,Xiao-
Jun Zeng,School of Computer Science University of
Manchester,Manchester,UK,Email: x.zeng@manchester.ac.uk, “A
Stackelberg Game Approach to Maximise Electricity Retailer’s
Profit and Minimise Customers’ Bills for Future Smart Grid”, .
[14] Eunji Lee1 and Hyokyung Bahn2,1 Department of EECS, Uni-
versity of Michigan, Ann Arbor, MI 48105, USA,2Department
of Computer Engineering, Global Top 5 Research Institute,
Ewha University, Seoul 120-750, Republic of Korea, “Electric-
ity Usage Scheduling in Smart Building Environments Using
Smart Devices”, Hindawi Publishing Corporation The Scien-
tificWorld Journal Volume 2013, Article ID 468097, 11 pages
http://dx.doi.org/10.1155/2013/468097.
[15] Jyh-Yih Hsu 1, Chien-Hua Chu 2, Chih-Ching Chen 3 1, 2
Department of Management Information Systems 3 Department
of Computer Science and Engineering, National Chung-Hsing
University, No. 250, Kuo-Kuang Road,Taichung City, 402, Tai-
wan, ROC, “An Integrated Fuzzy Multi-Objective Genetic Algo-
rithm for Optimization of Residential Appliances Scheduling”,
2011 International Conference on Signal, Image Processing and
Applications With workshop of ICEEA 2011 IPCSIT vol.21
(2011) (2011) IACSIT Press, Singapore.
[16] Zikri Bayraktar, Member, IEEE, Muge Komurcu, Jeremy A.
Bossard, Member, IEEE, and Douglas H. Werner, Fellow, IEEE,
“The Wind Driven Optimization Technique and its Application
in Electromagnetics”, IEEE TRANSACTIONS ON ANTENNAS
AND PROPAGATION, VOL. 61, NO. 5, MAY 2013.
[17] http://www.doc.ic.ac.uk/∼nd/surprise 96/journal/vol1/hmw/
article1.html..