An Optimized Priority Enabled Energy Management System for Smart Homes

Abstract

With the advent of smart grid (SG) and the emergence of information and communication technology, smart meters , bidirectional communication, smart homes and storage systems the energy consumption patterns at the consumer premises have been revolutionized. Moreover, with the rise of renewable energy sources (RESs), storage systems and electric vehicles (EVs) a profound amelioration in the energy management systems has been observed. Home energy management systems (HEMSs) help to control, manage and optimize the energy in smart homes. In this paper, we present a HEMS using multi-agent system (MAS) for smart homes. The HEMS uses priority techniques with the integration of electrical supply system (ESS). Furthermore, a bio-inspired technique, binary particle swarm optimization (BPSO), is used for the optimal scheduling of appliances in a smart home. Simulation results illustrate the effectiveness of the HEMS in terms of electricity cost, demand, user comfort and peak to average ratio (PAR).

Full text

An optimized priority enabled energy
management system for smart homes
Samia Shah, Rabiya Khalid, Ayesha Zafar,
Sardar Mehboob Hussain, Hassan Rahim, Nadeem Javaid∗
COMSATS Institute of Information Technology, Islamabad 44000, Pakistan
∗Corresponding author: www.njavaid.com, nadeemjavaidqau@gmail.com
Abstract—With the advent of smart grid (SG) and the emer-
gence of information and communication technology, smart me-
ters, bidirectional communication, smart homes and storage sys-
tems the energy consumption patterns at the consumer premises
have been revolutionized. Moreover, with the rise of renewable
energy sources (RESs), storage systems and electric vehicles (EVs)
a profound amelioration in the energy management systems has
been observed. Home energy management systems (HEMSs) help
to control, manage and optimize the energy in smart homes. In
this paper, we present a HEMS using multi-agent system (MAS)
for smart homes. The HEMS uses priority techniques with the
integration of electrical supply system (ESS). Furthermore, a bio-
inspired technique, binary particle swarm optimization (BPSO),
is used for the optimal scheduling of appliances in a smart
home. Simulation results illustrate the effectiveness of the HEMS
in terms of electricity cost, demand, user comfort and peak to
average ratio (PAR).
Index Terms—- Smart homes; HEMS; MAS; demand side
management; renewable energy sources
I. INTRODUCTION
Electricity is provided to more than 80% of people in
this world. Due to increasing modern society more people
have access to electricity in the last decade. Nowadays it is
becoming a more challenging task for the modern society to
optimally schedule the energy resources, which in turn helps
to solve the energy dilemma.
Electricity provision system which exploits the operations
of electricity generation, transmission, distribution and control
is termed as grid. The SG is an enhancement of traditional
power grid. It consist of two way communication technologies,
smart meters, smart appliances and RES to help achieve a
system which is reliable, self-healing, cost-effective, secure
and dynamically controllable. On the other hand, a smart home
is a very necessary unit of smart grid. Since early nineties the
concept of smart homes has been evolved [14]. The usage of
electricity is rapidly increasing in every smart home. With the
increase of electricity usage, the need to improve the overall
efficiency of electrical grids is also increasing. Hence, energy
efficiency is becoming a more challenging task for both SGs
and smart homes [16].
The EMS helps to optimize, control and manage the op-
erations of electricity generation, transmission, distribution
and control. Two important functions of EMS are demand
side management (DSM) and supply side management (SSM).
DSM is a strategy proposed by utility which helps electricity
consumers to change their electricity consumption patterns
through education. Usually, the aim of DSM is to educate the
energy consumer to use less electricity during on peak hours
and fulfill their energy requirements on off peak hours. On the
other hand, SSM helps to manage various energy generation
resources and improve the reliability of supply.
MAS technology plays a very potential role in SG and
smart homes. In literature it has been used for the purpose of
simulations, demand response, operation, SG control, control
of microgrid, storage systems, smart home control and for
service restoration [11] [12] [15]. In MAS, multiple agents
communicate with each other to achieve a specific goal in
a very efficacious way. This allows agents to communicate,
interact and parley with each other for the efficient use of
energy resources.
In this research work HEMS is implemented using MAS for
smart homes. An algorithm is designed to maintain a balance
between demand and supply. Smart appliances are scheduled
using a swarm intelligence optimization technique, BPSO, to
help find the best schedule for each appliance. This system
integrates ESS which consist of EV, wind turbine (WT),
storage system and main grid (MG). This system enables two
priority techniques [12] with the integration of ESS. Time of
use (TOU) pricing tariff is used which has low, medium and
high peak hours.
The rest of this paper is organized as follows. Section II
presents a brief state of the art. Section III illustrates the
proposed methodology and section IV presents simulations
and results. Finally, section V concludes the overall work.
II. STATE OF THE ART
A lot of research has been conducted for the optimal
scheduling of appliance to achieve a number of objectives.
On this subject, few papers are discussed below.
Authors in [3], comparatively evaluate the performance of
home energy management controller (EMC) which is designed
by using three algorithms named: ant colony optimization
(ACO), BPSO and genetic algorithm (GA). In this research
work, objective function is formulated by using multiple
knapsack problem. The model of EMC has been designed
for the energy management of residential system to avoid the

formation of peaks. Moreover, according to the consumption
pattern and TOU, appliances have been classified into three
categories. Later, it is mentioned that this model is designed
for three types of users in residential area. This categorization
is done on the basis of their energy consumption pattern.
The main objectives of this paper include the minimization
of PAR, electricity bill, execution time and maximization of
user comfort level. In this work, combined model of inclined
block rate (IBR) and TOU are used for energy pricing. It has
been concluded from the results that the proposed EMC model
works more efficiently with GA as compared to the other
two heuristic algorithms for the reduction of electricity bill
and minimizing PAR while maximizing user comfort level.
The execution time of this model using GA is less than the
execution time of EMC models using other two algorithms.
However, no priority technique is discussed and authors did
not consider privacy and security issues between utility and
end user.
For the optimization of execution time of scheduler and
energy consumption in smart grid a novel swarm intelligence
multi-objective algorithm named multi-objective brainstorm
algorithm (MOBSA) is presented in [4]. Brainstorming is a
process which consists of a group of people from different
backgrounds that propose several ideas to solve a particular
problem. This algorithm optimizes response time and energy
consumption by grid scheduling. Another novel and swarm
algorithm based on the firefly behavior, called multi-objective
firefly algorithm (MO-FA) is presented in this work. To
prove their efficiency and consistency, the two algorithms are
compared with non-dominated sorting GA II (NSGA-II). All
the above mentioned algorithms have been implemented in
GridSim simulator. MOBSA shows good performance regard-
ing energy efficiency and execution time over the other two
algorithms.
Authors in [5], identified the problem of easy and rapid
convergence in computation and poor performance in high
dimensional search in bee colony optimization (BCO) tech-
nique. The energy management of microgrids is a complex
and high dimensional problem. Therefore, to overcome the
problem of fast convergence an enhanced BCO (EBCO) algo-
rithm is proposed. The proposed technique helps to improve
the efficiency and accuracy of search by introducing a self-
adaption and repulsion factor. EBCO, in this research work,
is intended to optimize the use of distributed generators to
improve the efficiency of microgrid energy production. The
efficiency of the proposed algorithm is tested in two modes:
grid-connected and islanded mode. To prove its effectiveness
the proposed algorithm has been compared with evolutionary
programming (EP), GA, PSO and BCO algorithms. From the
results, the algorithm demonstrates that its average execution
time is less than GA but still greater than EP, BCO and
PSO which needs to be optimized. However, the authors
did not consider consumers demand in this work. Although
for an efficient energy management of power system both
demand and supply need to be considered. In our work, energy
management is done on the basis of consumers’ demand and
producers’ supply.
It has been mentioned in [6] that the existing DSM strate-
gies use specific algorithms and techniques. Besides, a very
small range of controllable loads are controlled through these
strategies. To overcome this problem, the proposed algorithm
handles a large number of loads. A DSM strategy is proposed
in this research work which can be used for the optimization
purpose of future SGs. The objective of the proposed work
is increased saving by reducing the peak demand on SG.
Bacteria foraging optimization algorithm (BFOA) is used as
a load shifting technique as DSM strategy for SG. This
technique is very effective in handling different types of loads
in three customer areas. The considered areas are residential,
commercial and industrial areas handling different types of
devices. Using BFOA the operational cost and peak demand
has been reduced in the three mentioned areas by 7.4011%,
5.9378% and 10.0911% respectively. The proposed algorithm
has been compared with evolutionary algorithm; however, the
results have not been shown in this research work. Further-
more, the integration of RES and priority techniques need
to be considered in this work. A heuristic methodology that
is proposed in [7] is called signaled PSO (SPSO) can be
applied to a large number of electricity consumers. It is
used to address the energy resources management problem
in SG. The comparison of the SPSO method with other
methodologies named mixed-integer non-linear programming
(MINLP), GA, original PSO, evolutionary PSO, and new PSO
revealed the superiority of the proposed scheme in terms
of faster convergence, better robustness, cost, time execution
and absolute error. For the optimization of the inter-operation
of SG-building EMS (SG-BEMS) framework an agent-based
approach has been proposed in [8].
A multi-agent control system is proposed in [9]. This system
optimally optimizes the energy and it helps to oversee the en-
ergy consumption of grid. To control the energy consumption
in grid, four intelligent agents are used. These agents are: load,
switch, central control and local coordinator agent. A heuristic
optimization technique, PSO, is used to control the mentioned
agents. In [10], a novel energy management architecture has
been presented. It envisions energy consumer to use on-site
RES. Real time price signal (RTP) scheme is used. Based
on user preferences and their demand, electricity consumers
are categorized into three types. Furthermore, priority enable
early deadline first (PEEDF) and modified first come first
serve (MFCFS) algorithms tackle different loads. However, the
authors did not consider any heuristic optimization technique
to schedule the load.
The focus of authors in [11] is on developing a HEMS
for smart homes. A MAS is proposed in this work. It helps
to handle the usage of surplus power when supply is greater
than demand and vice versa. However, in this work the authors
did not consider any scheduling technique for their model.
Moreover, the problem of peak formation and user comfort are
also neglected. Whereas, in [12], the implementation of HEMS
is presented using multi-agent system for smart homes. For this
purpose, the final consumer demand is calculated on the base

of two priority techniques i.e., priority of user comfort (PUC)
and priority of lowest consumption (PLC). These techniques
help to minimize the electricity cost, energy consumption
and power demand of electricity consumer. However, the
integration of RES and PAR reduction is neglected in this
work.
To overcome the aforementioned problems, we have pre-
sented a multi-agent HEMS. The HEMS helps to reduce
electricity cost and consumer demand. It also helps to reduce
electricity supply from main grid which consequently trims
down the PAR. The overall working of the proposed model is
presented in the next section.
III. METHODOLOGY
Smart grid literature shows a potential way to implement
smart grid with the help of multi-agent systems. Smart homes
are one of the requisite units of smart grid.
This section provides an overview of the working of HEMS.
We have presented an optimization based HEMS model. In
this work fourteen home appliance agents, four ESS agents
and three management agents are considered. The ESS agent
further consists of WT, EV, storage system and MG agents.
Furthermore, a heuristic optimization technique named BPSO
is used to get the best schedule for each appliance used in
the smart home. HEM, DSM and SSM are the three main
controller agents. SSM agent manages power flow from the
electrical supply systems. DSM agent manages the power flow
to the home appliances and HEM agent manages the two main
controller agents,the SSM agent and DSM agent.
This model considers RES and priority of household ap-
pliances. The appliances are categorized according to three
different types named fixed, interruptible and non-interruptible
appliances. In this work the priority of all the appliances
is assumed to be different depending on the user com-
fort/preference and their energy consumption. To calculate
electricity bill for energy consumer TOU electricity price tariff
is used. The electricity cost is calculated on the basis of energy
bought from MG to fulfil the consumer demand.
T D =T Dh−
14
X
a=1
ALa(1)
In this case the demand is fulfilled through WT, EV, battery
(B) and MG:
T D =
24
X
h=1
(MGh+W Th+EVh+Bh)(2)
However,
TEC =
24
X
h=1
(MGh·T OUh)(3)
In equation 1 the total demand is calculated after applying
the priority techniques [12]. Here, ALarepresents the load
of appliances, T D represents the total demand based on the
priority of appliance as given in table 1. In equation 2, TEC
shows the total electricity cost, MGhrepresents the main
grid power used to fulfil the consumer demand and T OUh
represents the electricity price signal for each hour.
Algorithm 1 Algorithm of HEM
1: Initialize parameters (τa, αa, βa, %a)
2: For all appliances an∈Ado
3: For all time slots t∈Tdo
4: For all electrical supply systems en∈Edo
5: Generate population randomly
6: Initialize velocity of particle to 0
7: Initialize particle position to Pbest
8: Initialize appliance on/off state [0 or 1]
9: For i = 1:T
10: Evaluate fitness function
11: Update particle velocity as in [2]
12: Update particle direction as in [2]
13: if v < 4then
14: v←4
15: if v > 4then
16: v←4
17: Apply sigmoid function as in [2]
18: For each bit
19: if rand < 1
1+e(−v)then
20: bit = 1
21: else
22: bit = 0
23: Save Gbest
24: end For
25: Return best schedule
26: For i = 1:T
27: Get demand and supply from DSM and SSM
28: if Demand > Supply then
29: Calculate insufficient power
30: Check energy stored in EV and storage device
31: if Insuf f icientpower!=0then
32: Buy electricity from main grid (Max 1000 W/h)
33: if Insuf f icientpower!=0then
34: Turn on the appliances based on the priority using eq. 1
35: else
36: if Demand < Supply then
37: Calculate Surplus power
38: Charge EV and storage device
39: if Surplus!=0 then
40: Sell surplus power to main grid
41: else
42: Turn on all the appliances
43: end For
Algorithm 1 represents the overall working of the proposed
scheme. It is designed considering various electrical supply
systems (ESSs) and priority settings. The status of the power
plug of each appliance is represented by ’1’ or ’0’. The
state ’1’ represents that the appliance is ON and the state 0
represents that the appliance is OFF. The DSM agent calculates
the total power demand by adding the power demand of ON
appliances in that hour. Then the total power demand of each
hour is sent to HEMS agent. In the same way, SSM checks
the power supply from the ESS and sends the final supply to
the HEMS agent. HEMS agent checks total demand of the
consumer and energy supplied from the SSM in each hour. If
the demand of the consumer is greater than the energy supplied
from the WT then HEMS agent shall calculate the insufficient
power and check if the power supplied from electrical supply
systems is sufficient to meet the total demand. If the supply
is still less than demand, consumer will buy electricity from
the main grid. Another thing that we have considered in our
work is that, the consumer cannot afford to buy more than
1000 W/h from main grid. After buying electricity from main
grid, if the power supply is still insufficient to meet consumer
demand than the priority techniques will be applied to turn

on the appliances of highest priority and turn off the low
priority appliances. In case of surplus power, extra power
shall be utilized to charge EV and battery. If the surplus
energy is still available, it will be sold back to the main grid
providing beneficial incentives to the consumer. Table I shows
the power rating and the arrangement of devices based on the
priority of importance to the house consumer comfort level
and lowest energy consumption of appliances. The priority of
each appliance is listed from the number 1 to 14 which shows
the importance of the appliance.
TABLE I: Priority and power rating of appliances.
1 Appliance power rating
(W/h)
PUC PLC
1 Light 225 3 4
2 Iron 2200 11 14
3 Fridge 300 1 6
4 Vacuum cleaner 1600 10 13
5 Tower fan 40 12 1
6 Television 230 6 5
7 Desktop computer 620 2 8
8 Standing fan 55 13 2
9 Air conditioner 1400 14 11
10 Rice cooker 820 8 10
11 Bulb 60 7 3
12 Electric dispensing pot 800 9 9
13 Bathroom heater 1500 4 12
14 Washing machine 400 5 7
In table I numbers from 1 to 14 are assigned to each
appliance indicating their priority. For example, the PLC of
appliance ’Tower fan’ is 1 which indicates that this appliance
is more important. Which means the smaller the number, the
more important an appliance is.
A. Different categories of appliances
For the proposed HEMS scheme, the household appliances
are divided into three groups, i.e., fixed, interruptible and non-
interruptible appliances. These appliances are categorized ac-
cording to their power consumption patterns. These categories
are explained bellow.
1) Fixed appliances: As the name suggests these appliances
are called fixed appliances because their length of operation
time and energy consumption patterns cannot be modified.
For example fridge, tower fan, light etc. We represent a fixed
appliance by fafor faF . The power rating of these appliances
is represented by fa%. The respected energy consumption of
these appliances is represented by αTand ςfashows either
the appliance is on or off in the respected time. The equation
4 shows the total energy consumption of fixed appliances for
a day [3].
αT=
24
X
h=1
(X
faF
fa%·ςfa)(4)
2) Interruptible appliances: Interruptible appliances are
also called elastic appliances. The energy consumption pattern
and usage time of these appliances can be changed The An In-
terruptible appliance is represented by Iafor IaI. The power
rating of these appliances is represented by Ia%. respected
energy consumption of these appliances is represented by βT
and ςIashows either the appliance is on or off in the respected
time. The equation 5 depicts the total energy consumption of
fixed appliances for a day [3].
βT=
24
X
h=1
(X
IaI
Ia%·ςIa)(5)
3) Non-interruptible appliances: These appliances are also
called burst load appliances because their operation time
should not be interrupted. The starting and ending time of
their operation is predefined by the electricity consumer. We
symbolize a non-interruptible appliance by NIafor NIaN I.
The power rating of these appliances is represented by NIa%.
The respected energy consumption of these appliances is
represented by γTand ςNIashows either the appliance is on
or off in the respected time. The equation 6 depicts the total
energy consumption of fixed appliances for a day [3].
γT=
24
X
h=1
(X
NIaN I
NIa%·ςN Ia)(6)
B. Electricity price tariff
TOU is a commonly used residential electricity pricing
scheme. In TOU, the price of electricity remains same for fixed
hours in a day. This pricing scheme has low peak, medium
peak and high peak prices. Mathematically, this pricing tariff
can be represented as follows [1]:
δ(h) = 


δ1hH1
δ2hH2
δ3hH3
(7)
C. PAR
PAR is the parameter of utility. It helps to maintain balance
between consumer demand and supply. It is defined as the
ratio of maximum load to the average load in a day. The Par
is mathematically formulated as [13]:
Let LMG indicate the total load demand fulfilled through MG
at each time slot t{1,·· ·, T }. We denote the peak and average
load as PLand AL, respectively.
PL= max
tT LMG (8)
AL=1
T
24
X
t=1
LMG (9)
Hence, the PAR can be acquired as
P AR =PL
AL
(10)
D. BPSO
BPSO is an improved version of PSO. It is a swarm
intelligence optimization technique which depicts the social
behaviour of flocking birds.It helps to optimize various op-
timization problems in the field of computer science, engi-
neering etc.. In this technique, the particles search the entire
space in the form of a swarm and update their velocity and

direction in each iteration which consequently helps to get
a near optimal solution. Moreover, this scheme is easy to
implement and it has only a few parameters to deal with. In
BPSO, each particle of the population takes a decision which
is either 1/0 or yes/no. Furthermore, the probability of each
particle to take a decision can be computed as given below
[2]:
P(Xid = 1) = f(Xid(t−1)), Vid (t−1), pid, pgd (11)
Applying (12) this threshold can be modeled. Using this
sigmoidal function, the state dth state of the ith individual
can be modeled at any time 0t0:
Xid =1ifρid < s(Vid)
0otherwise (12)
In Xid, the best state of individual will be stored. Moreover,
the importance of social and individual factors vary from one
decision to another, so random weights are applied to each
part, as given in 13:
Vid =Vid(t−1) + ϕ1·rand1·(pid −Xid(t−1)) + ···
ϕ2·rand2·(pgd −Xid (t−1))
(13)
IV. SIMULATIONS AND RESULTS
In this section, simulations and results are discussed in
detail. We conduct simulations in java agent development
(JADE) framework, to evaluate the performance of different
parameters. Through simulations we have achieved our ob-
jectives: consumer demand, electricity cost, PAR, main grid
supply reduction on the basis of priority techniques and RES
with scheduled and unscheduled load. The scheduling of
home appliances is done using BPSO technique. In Fig. 1 ,
consumer demand is calculated using unscheduled load, load
calculated after applying the PLC on unscheduled load and
integrating ESS . It is also calculated by scheduling the home
appliances using BPSO, applying PLC, and integrating RES.
The results show that, the consumer demand, calculated by
using the PLC and integrating RES on scheduled load is less
as compared to the demand calculated through unscheduled
load and by applying the priority techniques and integrating
ESS.
In Fig. 2 electricity cost is computed on the base of demand
illustrated in Fig. 1. It shows that electricity cost using PLC
with BPSO and integrating ESS is less as compared to the
PLC + ESS and electricity cost of unscheduled load with
no technique. It is worth mentioning here that, in this work
electricity demand is fulfilled by using the ESS. However, the
electricity cost is computed as given in equation 3.
Similarly, Fig. 3 and Fig. 4 illustrate the electricity demand
and electricity cost of consumer for each hour. However, in
the mentioned figures PUC technique is applied. Results show
that, this priority technique gives best results for both demand
and electricity cost when applied on scheduled load and ESS is
integrated. This is because the demand using this technique is
least, hence, electricity to be bought from main grid will also
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
0
500
1000
1500
2000
2500
3000
3500
4000
Time (hours)
Consumer Demand (Watts)
Unscheduled
PLC+ESS
PLC+ESS with BPSO
Fig. 1: Consumer demand using PLC
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
0
5
10
15
20
25
30
35
Time (hours)
Electricity cost (cents)
Unscheduled
PLC+ESS
PLC+ESS with BPSO
Fig. 2: Electricity cost using PLC
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
0
500
1000
1500
2000
2500
3000
3500
4000
Time (hours)
Consumer Demand (Watts)
Unscheduled
PUC+RES
PUC+ESS with BPSO
Fig. 3: Consumer demand using PUC

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
0
5
10
15
20
25
30
35
Time (hours)
Electricity Cost(cents)
Unscheduled
PUC+ESS
PUC+ESS with BPSO
Fig. 4: Electricity cost using PUC
1 2 3 4 5
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5 x 104
Total consumer demand (Watts)
Unscheduled
PUC + ESS
PUC + ESS with BPSO
PLC + ESS
PLC + ESS with BPSO
Fig. 5: Total consumer demand
be less. From Fig. 2 and 4, it is clear that the performance
of PUC is better in terms of electricity cost. This is because,
the demand calculated through this technique is less, hence
less electricity will be required from main grid. This scheme
consequently benefits the consumer in terms of electricity cost
and increases comfort as well.
Fig. 5 epitomizes the total demand of the electricity con-
sumer computed when no technique is applied on the unsched-
uled load, PUC with integration of ESS, PUC with integration
of ESS and BPSO, PLC with integration of ESS and PLC
with BPSO and ESS integrated. From this figure, it is clear
that the demand of electricity consumer is least while using
PUC with BPSO and integrating RES. However, electricity
demand of rest of the techniques is less as compared to the
demand computed when no technique is applied.
Whereas, Fig. 6 shows the total cost to be paid by electricity
consumer considering the demand shown in Fig. 5. It is
clear from this figure that the electricity bill of consumer
12345
0
50
100
150
200
250
300
350
Total Electricity Cost (Cents)
Electricity Cost
Unscheduled
PUC + ESS
PUC + ESS with BPSO
PLC + ESS
PLC + ESS with BPSO
Fig. 6: Total electricity cost
1 2 3 4
0
1000
2000
3000
4000
5000
6000
7000
Total main grid supply(Watts)
PUC + ESS
PUC + ESS with BPSO
PLC + ESS
PLC + ESS with BPSO
Fig. 7: Electricity bought from main grid
1 2 3 4 5
0
0.5
1
1.5
2
2.5
PAR
Unscheduled
PUC + ESS
PUC + ESS with BPSO
PLC + ESS
PLC + ESS with BPSO
Fig. 8: PAR

will be least while using the technique PUC with BPSO and
integrating RES. However, the overall electricity cost of rest
of the techniques is less when compared with electricity cost
of unscheduled load.
Fig. 7 illustrates the total amount of electricity bought from
main grid. Whereas, Fig. 8, depicts the PAR generated by
applying the techniques mentioned in the labels. This figure
depicts there is a trade-off between PAR and electricity cost.
However, the overall PAR of all the techniques is less when
compared with PAR of unscheduled load. From the simulation
results we conclude that, in case of unscheduled load, the
performance of the priority technique PUC is better than PLC
when integrated with ESS. In case, when load is scheduled
using BPSO, PLC again performs better than PLC when
integrated with ESS. However, scheduled PLC gives better
results than unscheduled PUC.
V. CONCLUSION
The HEMS plays a very important role in the energy con-
sumption patterns of smart homes. In the future, exploitation
of RES and energy consumption patterns shall be changed
through the immense use of HEMS. In this paper we have
presented a HEMS using MAS for smart homes. For this
purpose, we have used a bio-inspired technique BPSO which
helps to find best schedule for each appliance. TOU pricing
tariff is used. This system enables two types of priority
techniques and the integration of ESS. The simulation results
show that the technique PUC with BPSO gives best results
in terms of electricity cost, PAR and consumer demand as
compared to the technique PLC with BPSO.
REFERENCES
[1] Rastegar, Mohammad, Mahmud Fotuhi-Firuzabad, and Hamidreza
Zareipour. ”Home energy management incorporating operational priority
of appliances.” International Journal of Electrical Power and Energy
Systems 74 (2016): 286-292.
[2] Del Valle, Yamille, Ganesh Kumar Venayagamoorthy, Salman Mo-
hagheghi, Jean-Carlos Hernandez, and Ronald G. Harley. ”Particle swarm
optimization: basic concepts, variants and applications in power systems.”
IEEE Transactions on evolutionary computation 12, no. 2 (2008): 171-
195.
[3] Rahim, Sahar, 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.
[4] Arsuaga-Ros, Mara, and Miguel A. Vega-Rodrguez. ”Multi-objective
energy optimization in grid systems from a brain storming strategy.” Soft
Computing19, no. 11
[5] Lin, Whei-Min, Chia-Sheng Tu, and Ming-Tang Tsai. ”Energy manage-
ment strategy for microgrids by using enhanced bee colony optimiza-
tion.”Energies 9, no. 1 (2015): 5.
[6] Esther, B. Priya, et al. ”Demand Side Management Using Bacterial
Foraging Optimization Algorithm.” Information Systems Design and
Intelligent Applications. Springer India, 2016. 657-666. (2015): 3159-
3172.
[7] Soares, Joo, Marco Silva, Tiago Sousa, Zita Vale, and Hugo Morais.
”Distributed energy resource short-term scheduling using Signaled Parti-
cle Swarm Optimization.” Energy 42, no. 1 (2012): 466-476.
[8] Hurtado, L. A., P. H. Nguyen, and W. L. Kling. ”Smart grid and smart
building interoperation using agent-based particle swarm optimization.”
Sustainable Energy, Grids and Networks 2 (2015): 32-40.
[9] Wang, Lingfeng, Zhu Wang, and Rui Yang. ”Intelligent multiagent control
system for energy and comfort management in smart and sustainable
buildings.” IEEE transactions on smart grid 3, no. 2 (2012): 605-617.
[10] Rasheed, Muhammad Babar, Nadeem Javaid, Ashfaq Ahmad, Muham-
mad Awais, Zahoor Ali Khan, Umar Qasim, and Nabil Alrajeh. ”Priority
and delay constrained demand side management in realtime price envi-
ronment with renewable energy source.” International Journal of Energy
Research 40, no. 14 (2016): 2002-2021.
[11] V Li, W., T. Logenthiran, and W. L. Woo. ”Intelligent multi-agent system
for smart home energy management.” In Smart Grid Technologies-Asia
(ISGT ASIA), 2015 IEEE Innovative, pp. 1-6. IEEE, 2015.
[12] Li, Weixian, et al. ”Implementation of demand side bmanagement of a
smart home using multi-agent system.” Evolutionary Computation (CEC),
2016 IEEE Congress on. IEEE, 2016.
[13] Liu, Yi, Chau Yuen, Shisheng Huang, Naveed Ul Hassan, Xiumin
Wang, and Shengli Xie. ”Peak-to-Average Ratio Constrained Demand-
Side Management With Consumer’s Preference in Residential Smart
Grid.” IEEE Journal of Selected Topics in Signal Processing 8, no. 6
(2014): 1084-1097.
[14] Lobaccaro, Gabriele, Salvatore Carlucci, and Erica Lfstrm. ”A Review of
Systems and Technologies for Smart Homes and Smart Grids.” Energies
9, no. 5 (2016): 348.
[15] Narkhede, Mahesh S., S. Chatterji, and Smarajit Ghosh. ”Article: Multi-
agent systems (mas) controlled smart grid-a review.” In IJCA Proceedings
on International Conference on Recent Trends in Engineering and Tech-
nology, pp. 12-17. 2013.
[16] Zhou, Bin, 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 Sus-
tainable Energy Reviews 61 (2016): 30-40.