Content uploaded by Nadeem Javaid
Author content
All content in this area was uploaded by Nadeem Javaid on Mar 22, 2019
Content may be subject to copyright.
Minimizing Daily Cost and Maximizing
User Comfort Using a New Metaheuristic
Technique
Raza Abid Abbasi1, Nadeem Javaid1(B
), Sajjad Khan1,ShujaturRehman
2,
Amanullah2, Rana Muhammad Asif3, and Waleed Ahmad1
1COMSATS Institute of Information Technology, Islamabad 44000, Pakistan
nadeemjavaidqau@gmail.com
2Quaid-i-Azam University, Islamabad 44000, Pakistan
3NCBA&E, Multan, Pakistan
http://www.njavaid.com
Abstract. A home energy management system intended to improve the
energy consumption pattern in a smart home is proposed in this research.
The objective of this work is to handle the load need in an adequate man-
ner such that, electrical energy cost and waiting time is minimized where
Peak to Average Ratio (PAR) is maintained through coordination among
appliances. The proposed scheme performance is assessed for PAR, user
comfort and cost. This work assess the behavior of advised plan for real-
time pricing and critical peak pricing schemes.
Keywords: Smart Grid ·DSM ·HEMS ·BCS
1 Introduction
In recent decades an exponential growth in the energy consumption is observed,
therefore, the gap between demand and generation is increasing. Because of the
rise in the energy need, the conventional grid is facing a number of problems
like reliability, maintenance and sustainability. Residential consumer are grow-
ing rapidly at a pace of 0.3% annually, and has large share of 27% among others
in electricity consumption [1]. In past most of the power generation was achieved
using fossil fuels to fulfill the gap between energy demand and generation. How-
ever in recent years scientists have worked on exploring new means of energy
generation, i.e., renewable and sustainable energy resources [2]. The addition of
Renewable Energy Sources (RESs) resulted in an increase in the power system
complexity, where existing grids were not able to manage that. Smart Grid (SG)
is introduced to handle the above mentioned challenges. A SG is equipped with
Information and Communication Technology (ICT), that enabled it to incorpo-
rate RESs and manage the stability and reliability of power systems. The most
important feature of SG is the control of energy production, transmission and
distribution through advanced ICTs.
c
Springer Nature Switzerland AG 2019
L. Barolli et al. (Eds.): WAINA 2019, AISC 927, pp. 80–92, 2019.
https://doi.org/10.1007/978-3-030-15035-8_8
HMS Using Hybrid Meta-heuristic Optimization Technique 81
The state of the art techniques for DSM and SG are discussed in Sect. 2.
Different problems related to the DR and DSM are reflected along with their
consequences in Sect. 3. Section 4explains the suggested scheme for discussed
problemettes. Section 5highlights formerly employed techniques, Strawberry
Algorithm (SBA), Earthworm optimization Algorithm (EWA), and schemed
algorithm. Simulations outcomes are depicted and shown in Sect.6. At last,
Sect. 7draws conclusion of the work done.
2 Related Work
SG enabled the utility to manage the DR of customers to minimize the overall
load at power generating sources.
In [6] Huang et al. proposed to enhance the control of appliances in a HEMS,
considering the uncertain electricity loads and cost. A new scheme, involving,
Gradient-based Particle Swarm Optimization (GPSO) for solving the global
optimum solution and Two Point Estimation Method (2PEM) for handling the
indeterminate behavior of must load runs was introduced. Thus, the presented
GPSO-2PEM can minimize processing cost and can be easily applied in an
embedded device having limited resources. Hansen et al. proposed HEMS scheme
for reducing the household electricity bill in [7] using RTP pricing scheme. They
introduced non-myopic Partially Observable Markov Decision Process (POMDP)
technique for reducing the peak demand which ultimately reduces the overall
electricity cost.
In [8] authors proposed a new HEMS which aims at minimizing the next day
energy cost and disturbance to the user with RTP and household photo-voltaic
penetrations. In proposed system, first user defined electricity usage restrictions
are set. Then, an optimal scheduling model is proposed based on the predicted
solar output and energy prices. It will support the decision making for the RES
operations. They used advanced adaptive thermal comfort model for sensing the
users indoor thermal comfort degree which supports in scheduling the heating,
ventilating and air conditioning systems. Where, user disturbance value met-
ric is suggested to understand the psychological disturbance of an appliance
schedule on the users preference. Luo et al., worked on optimize scheduling of
Distributed Residential Energy Resources (DRERs) in [9]. They worked for a SH
having varying electricity pricing and high RES penetrations. They used Monte
Carlo sampling technique for handling the uncertainties of solar power output.
Users preferences are also taken into account with different levels of priority,
i.e., normal and high. Considering all this, an optimal DRER scheduling model
is proposed to minimize the electricity bill. Authors used natural aggregation
algorithm to solve the proposed model.
Elghitani et al., proposed a methodology for residential demand aggrega-
tion, based on a multi-class queuing system in [10]. Authors used this model for
reducing the cost of the energy power consumption below day-ahead predicted
pricing. Proposed scheme achieves a cost that is close to the best solution. In [11]
Mosaddek Hossain et al., proposed a Real-time Decentralized DSM (RDCDSM)
82 R. A. Abbasi et al.
to adjust the real-time residential load to follow a preplanned day-ahead energy
generation by the microgrid. based on predicted customers aggregate load. A
deviation from the predicted demand at the time of consumption is assumed
to result in additional cost or penalty. Wu et al. in [12] considered the energy
storage uncertainties due to intermittent renewable energy supplies and energy
storage opportunity. Authors proposed a new stochastic dynamic programming
model for the optimal energy management of a smart home with Plug-in Electric
Vehicle (PEV) energy storage. They worked on minimizing the energy cost while
satisfying the user energy demand and PEV storage requirements.
In [13]Moonet al., used 2-stage forecasting scheme for prediction of opti-
mal operation of power system in educational buildings using Short Term Load
Forecasting (STLF) model. Authors collected load data of a university campus
for last five year. Where, using the moving average method found the electric
load pattern of day of a week. Random forest method is being used with time
series cross-validation for forecasting the daily electric load. Keles et al. in [14]
proposed a new scheme that uses Artificial Neural Network (ANN) for predicting
the electricity prices. In a ANN forecasting scheme the output performance is
dependent on the parameter sets, therefore, selection and preparation of funda-
mental data that has prominent effect on the electricity prices is in more focus.
3 Problem Statement
From the time of its evolution, SG has a number of complications as well, i.e.,
energy transferring, equality, surveillance and secrecy at DSM level [5]. The
uneven use of electricity on user side rises the burden on the power distribution
and power generating system. Due to irregular usage, energy required in some
hours (peak hours) increases exponentially. Such high demand results in using
extra generation units to provide the required energy, which subsequently results
increase in cost. Therefore, DSM needs a proficient scheme that can distribute
the load between peak hours and off peak hours evenly.
Although, authors have already done a lot of work on DSM for reducing cost,
waiting time and PAR. Authors in [6,10,17,18] worked on DR optimization con-
sidering uncertainties of must run load, aimed at reducing the upper bound of
energy consumption cost, however PAR was not considered. In [7,8,11] authors
worked on reducing the peak demand through scheduling techniques for appli-
ances incorporating the RES, however user comfort was not considered.
4 Proposed System Model
For efficient usage of electricity, DSM is the base for HEMS. Smart appliances in
a SH are connected to one another and Energy Management Controller (EMC)
through Home Area Network (HAN). Energy usage data from smart appliances
is sent to EMC and then EMC process that data for making important deci-
sions. Utility get the information sent from SM and executes required processing
accordingly. Utility propagates the pricing signal and the consumer required
HMS Using Hybrid Meta-heuristic Optimization Technique 83
energy to SM. SM sends the information coming through the utility with the
EMC. Now, EMC possesses the data collected from appliances and the cost infor-
mation received from utility. Information collected from utility and appliances
in used by EMC for optimum appliances scheduling. EMC uses the Strawberry
Algorithm (SBA), Earthworm Optimization Algorithm (EWA) and proposed
scheme for scheduling. The goal of the scheduling is to reduce PAR, maximize
user satisfaction through minimizing waiting time and reducing cost. The dis-
cussed system model is displayed in Fig. 1.
In our proposed model, we have a house, using 11 appliances and those appli-
ances are distributed among different classes. We classify appliances in three
main classes, i.e., Fixed, shift-able and interruptible appliances. Fixed appli-
ances require to be ON according to the specified schedule. Those appliances
that are optional to scheduling but the can not be interrupted while they are
working are classified as Shiftable appliances. Furthermore, there is a condition
for washing dryer that it can start its operation when washing machine has com-
pleted its operation. Appliances those are schedulable as well as interruptible are
placed in Interruptible appliances category. Table1, displays the classified smart
appliances as well as their LoT and power rating in respective columns. The pro-
posed schemes objective is to: minize PAR, maximize user comfort by reducing
waiting time and reducing total cost.
Table 1. Control parameters
Classes Appliances LoT (h) PR (kWh)
Fixed appliances Light 12 0.1
Coffee maker 40.5
Oven 9 3
Blender 41.2
Shiftable appliances Washing machine 50.5
Cloth dryer 4 4
Dish washer 41.5
Interruptible appliances Water heater 12 1.1
Space heater 12 1.5
Iron 61.1
Vacuum cleaner 50.5
5 Metaheuristic Optimization Algorithms
Meta-heuristic algorithms are structured to provide solution to almost any type
of problem. Therefore, we will use metaheuristic techniques for finding the solu-
tion to scheduling home appliances problem. In this study we are dedicated to
84 R. A. Abbasi et al.
Fig. 1. System model
total energy consumption cost minimization and user comfort maximization.
Now we will discuss the selected metaheuristic algorithms.
5.1 SBA
SBA is a nature inspired numerical optimization algorithm. This algorithm is
derived from the strawberry plant behavior for searching the resources and is
being used for solving complex computing problems. Strawberry plant uses both
runners and root for searching resources like water and minerals. Where, these
runners being used for global search and roots being used for local search. Runner
and roots are randomly initialized at the beginning. While algorithm running,
when these runners and roots finds new resources then the reproduction starts
which results in more runners and roots. The locations of roots and runners at
a specific iteration is calculated using Eq. 4. Where, Xroot(i)andXrunner (i)are
matrices containing locations of roots and runners respectively. Equations5 and
6 are used for optimization and is the mathematical representation of strawberry
algorithm.
5.2 EWA
Earthworm executes numerous optimization runs during the reproduction pro-
cess. Reproduction process of the earthworm follows following steps.
•Every earthworm is capable of participating in regeneration process, it can
reproduce at most two earthworms.
•The singular generated earthworm contains the complete genetic behavior of
its parent that is of same length as parent.
•The single fit earthworm supports a straight next generation and can not be
changed by operators. It guarantees that earthworm can not end with the
increment in generations.
HMS Using Hybrid Meta-heuristic Optimization Technique 85
5.3 Proposed
Our proposed scheme is based on the behavior of plant roots and earthworm.
Plant root search resources, i.e., minerals and water and grows in the direction of
best possible route to the resource. Earthworm performs numerous runs during
the reproduction process. Every earthworm has the ability to reproduce and has
same capability of reproduction like parent earthworm. In our proposed scheme
we consider plants as solution to the problem, plant roots for finding the global
optimum solution and earthworm capability for local optimum solution. The
resource found quality is mapped as fitness function in our proposed scheme.
The steps followed in our proposed scheme are listed below.
Algorithm 1. Proposed Scheme Algorithm for Appliance Scheduling.
Require: Input: [Initialized the population]
while iteration < M axN umbererI teration do
for i←1tosize(population)do
Generate offspring through Reproduction 1
Generate offspring through Reproduction 2
end for
for j←1tosize(Primes)do
Perform Crossover
end for
Evaluation of fitness
for i←1tosize(population)do
Select random location by chance as mother of the next iteration
end for
Apply mutate
Local best solution is extracted for iteration j
end while
Global best solution is achieved
Appliances are scheduled
6 Simulation Results and Discussion
We evaluated the performance of our proposed technique for reducing the cost.
The results obtained from extensive simulation are discussed here in this section.
We compared results obtained from simulations with other techniques, i.e.,
Earthworm optimization Algorithm (EWA) and Strawberry Algorithm (SBA).
Critical Peak Pricing (CPP) and Real Time Pricing (RTP) pricing signals are
used to examine the usage pattern of the electricity by the users. We consid-
ered single home with eleven appliances, i.e., D = 11 for simulation purpose.
We classified these appliances in three different classes, i.e., Fixed Appliances
(FA), Shiftable Appliances (SA) and Interruptible Appliances (IA). These appli-
ances are selected due to their frequent use in winter. Light, coffee maker, oven
and blender are included in FA. FA can neither be shifted nor interrupted.
86 R. A. Abbasi et al.
Washing machine, cloth dryer and dish washer are included in SA. These appli-
ances can not be interrupted once they have started their working, however these
appliances can be shifted. Water heater, space heater, iron and vacuum cleaner
are included in IA. This group of appliances can be shifted as well as interrupted.
CPP and RTP are used for load, cost and PAR calculation.
6.1 Pricing Tariff
As mentioned earlier, we will evaluate our proposed scheme using CPP and RTP
pricing signals. Here we will discuss CPP and RTP pricing tariff.
6.1.1 CPP Tariff
During hot summer weekdays utilities face emergency conditions. Due to critical
condition during a specific time period utility impose high price rate. The energy
cost through this interval are commonly above average. CPP has two different
types. In the first variant specific instance of time and amount of peak prices is
already known. In the second variant, electricity prices changes with the use of
electricity. If electricity use is increased electricity price also increases to reduce
the load on the utility. Maximum fifteen time these critical hours are allowed in
a season.
6.1.2 RTP Tariff
The RTP pricing scheme is also called dynamic price rate due to its varying
nature. It depends on amount of electricity used per hour. It feeds details about
actual cost of power at a specific time to the user. It empowers the user to
acclimate their working hours from on peak hours to off peak hours, which
results in saving. It is aided using SM which allows the mutual exchange of
information between utility and consumer. Utilities calculate the bill in RTP as
sum of two different elements.
(a) Base cost is computed based on Customer Baseline Load (CBL). It is a
standard defined tariff.
(b) Per hour usage cost is put in based on the time of use of electricity. This is
the basic difference between CBL and the actual usage.
Figures 4and 5are showing the CPP and RTP power consumption pattern for
unscheduled and scheduled schemes, i.e., SBA, EWA and proposed respectively.
The proposed scheme is scheduling the energy in an efficient way in comparison
to alternatives. However, proposed scheme load consumption in last hour is high
which is due to reducing load consumption during peak hours (Fig. 2).
Figures 6and 7are showing the unscheduled, SBA, EWA and proposed scheme
hourly cost respectively for CPP and RTP pricing schemes respectively. It is
evident from figures that proposed scheme is performing better then other schedul-
ing schemes while transferring the high energy demand to off peak hours. Further-
more, we can see that even SBA and EWA are performing better than unscheduled
(Fig. 3).
HMS Using Hybrid Meta-heuristic Optimization Technique 87
Fig. 2. CPP per hour load
Fig. 3. RTP per hour load
The PAR for unscheduled, SBA, EWA and proposed scheme using CPP
and RTP pricing schemes is displayed in Figs. 8and 9respectively. In case of
CPP pricing scheme PAR is reduced by 58.77%, 62.29% and 58.15% using SBA,
EWA and proposed scheme. Here, EWA outperformed other schemes. In case of
RTP pricing scheme PAR is reduced by 54.99%, 54.86% and 51.42% using SBA,
EWA and proposed scheme. Here, SBA outperformed other schemes. It is worth
mentioning that all schemes performed better than unscheduled.
Figures 10 and 11 are displaying the waiting time for unscheduled, SBA, EWA
and proposed scheme using CPP and RTP pricing schemes respectively. In case
of CPP pricing scheme waiting time is reduced to 23.37%, 25.41% and 18.77% for
SBA, EWA and proposed scheme. Here, proposed scheme outperformed other
schemes. In case of RTP pricing scheme waiting time is reduced to 24.16%,
24.28% and 18.13% using SBA, EWA and proposed scheme. Here, proposed
scheme outperformed other schemes. It is also noticed that all schemes performed
better than unscheduled.
88 R. A. Abbasi et al.
Fig. 4. CPP hourly cost
Fig. 5. RTP hourly cost
Fig. 6. CPP peak to average ratio
HMS Using Hybrid Meta-heuristic Optimization Technique 89
Fig. 7. RTP peak to average ratio
Fig. 8. CPP waiting time
Figures 12 and 13 are displaying the total cost for unscheduled, SBA, EWA
and proposed scheme using CPP and RTP pricing schemes respectively. In case
of CPP pricing scheme cost is reduced by 18.75%, 30.65% and 26.99% using
SBA, EWA and proposed scheme. Here, EWA outperformed other schemes. In
case of RTP pricing scheme cost is reduced by 13.99%, 15.25% and 27.05% using
SBA, EWA and proposed scheme. Here, our proposed scheme outperformed other
schemes.
The proposed scheme minimizes the waiting time, however, it compromises
for PAR and cost in case of CPP pricing signal. Where, proposed scheme mini-
mizes the waiting time and cost, however, it compromises PAR in case of RTP
pricing signal.
90 R. A. Abbasi et al.
Fig. 9. RTP waiting time
Fig. 10. CPP total cost
Fig. 11. RTP total cost
HMS Using Hybrid Meta-heuristic Optimization Technique 91
7 Conclusion
DSM is focused to manage the appliances on the user end for efficient use of
energy. Load shifting technique of DSM is used in proposed HEMS to attain
the desired results. A single house consisting of various appliances is considered
for scheduling. Each appliance of house is scheduled using three heuristic tech-
niques; SBA, EWA, and proposed scheme. The proposed scheme helps to find the
most optimal schedule of each house appliance considering the constraints. The
performance of applied techniques is assessed with respect to basic parameters:
PAR, user comfort and energy consumption cost. The cost of energy consump-
tion is calculated using RTP and CPP pricing schemes. Control parameters and
categorization of appliances is kept same, however, power rating of electrical
devices and their types is kept same for a fair comparison. Results are depicting
that waiting time is reduced to 18.77% in case of CPP and 18.13% for RTP
pricing scheme which is best with respect to SBA and EWA. Cost is reduced by
27.05% in case of RTP which is best in comparison to the other two techniques.
However, there is a trade-off, as we cannot achieve the best values for all three
performance parameters simultaneously.
References
1. U.S. Energy Information Administration, Annual Energy Outlook 2018, p. 121,
122, Residential and commercial energy consumption grows gradually from 2017
to 2050. https://www.eia.gov/outlooks/aeo/pdf/AEO2018.pdf. Accessed October
2018
2. Luo, F., Ranzi, G., Wan, C., Xu, Z., Dong, Z.Y.: A multi-stage home energy
management system with residential photovoltaic penetration. IEEE Trans. Ind.
Inf., 11 (2018)
3. Strbac, G.: Demand side management: benefits and challenges. Energy Policy 36,
4419–4426 (2008). Energies 11, 1426 (2018)
4. Gellings, C.W.: The concept of demand-side management for electric utilities. Proc.
IEEE 73(10), 1468–1470 (1985)
5. Baharlouei, Z., Hashemi, M.: Demand side management challenges in smart grid:
a review. In: Proceedings Smart Grid Conference (SGC), pp. 96–101, December
2013
6. Huang, Y., Wang, L., Guo, W., Kang, Q., Wu, Q.: Chance constrained optimization
in a home energy management system. IEEE Trans. Smart Grid 9(1), 252–260
(2018)
7. Hansen, T.M., Chong, E.K.P., Suryanarayanan, S., Maciejewski, A.A., Siegel, H.J.:
A partially observable markov decision process approach to residential home energy
management. IEEE Trans. Smart Grid 9(2), 1271–1281 (2018)
8. Luo, F., Ranzi, G., Kong, W., Dong, Z.Y., Wang, F.: Coordinated residential energy
resource scheduling with vehicle-to-home and high photovoltaic penetrations. IET
Renew. Power Gener. 12(6), 625–632 (2018)
9. Luo, F., Dong, Z.Y., Xu, Z., Kong, W., Wang, F.: Distributed residential energy
resource scheduling with renewable uncertainties. IET Gener. Transm. Distrib.
12(11), 2770–2777 (2018)
92 R. A. Abbasi et al.
10. Elghitani, F., Zhuang, W.: Aggregating a large number of residential appliances for
demand response applications. IEEE Trans. Smart Grid 9(5), 5092–5100 (2018)
11. Tushar, M.H.K., Zeineddine, A.W., Assi, C.: Demand-side management by regu-
lating charging and discharging of the EV, ESS, and utilizing renewable energy.
IEEE Trans. Ind. Inf. 14(1), 117–126 (2018)
12. Wu, X., Hu, X., Yin, X., Moura, S.J.: Stochastic optimal energy management of
smart home with PEV energy storage. IEEE Trans. Smart Grid 9(3), 2065–2075
(2018)
13. Moon, J., Kim, K.-H., Kim, Y., Hwang, E.: A short-term electric load forecasting
scheme using 2-stage predictive analytics. In: 2018 IEEE International Conference
on Big Data and Smart Computing (2018)
14. Keles, D., Scelle, J., Paraschiv, F.: Wolf Fichtner Extended forecast methods for
day-ahead electricity spot prices applying artificial neural networks. Appl. Energy
162, 218–230 (2016)
15. Ghofrani, M., Hassanzadeh, M., Etezadi-Amoli, M., Fadali, M.: Smart meter based
short-term load forecasting for residential customers. In: North American Power
Symposium (NAPS), pp. 1–5, August 2011
16. Huang, H., Cai, Y., Xu, H., Yu, H.: A multi-agent minority-game- based demand-
response management of smart buildings towards peak load reduction. IEEE Trans.
Comput.-Aided Des. Integr. Circuits Syst. 36(4), 573–585 (2017)
17. Ruelens, F., Claessens, B.J., Vandael, S., Schutter, B.D., Babuˇska, R., Belmans,
R.: Residential demand response of thermostatically controlled loads using batch
reinforcement learning. IEEE Trans. Smart Grid 8(5), 2149–2159 (2017)
18. Pourmousavi, S.A., Patrick, S.N., Nehrir, M.H.: Real-time demand response
through aggregate electric water heaters for load shifting and balancing wind gen-
eration. IEEE Trans. Smart Grid 5(2), 769–778 (2014)
19. Jindal, A., Dua, A., Kaur, K., Singh, M., Kumar, N., Mishra, S.: Decision tree
and SVM-based data analytics for theft detection in smart grid. IEEE Trans. Ind.
Informat. 12(3), 1005–1016 (2016)
20. Mirowski, P., Chen, S., Ho, T.K., Yu, C.-N.: Demand forecasting in smart grids.
Bell Labs Tech. J. 18(4), 135–158 (2014)
21. Wijaya, T.K., Vasirani, M., Humeau, S., Aberer, K.: Cluster-based aggregate fore-
casting for residential electricity demand using smart meter data. In: IEEE Inter-
national Conference on Big Data (Big Data)
22. Yang, J., Liu, J., Fang, Z., Liu, W.: Electricity scheduling strategy for home energy
management system with renewable energy and battery storage: a case study. IET
Renew. Power Gener. 12(6), 639–648 (2018)
23. Merrikh-Bayat, F.: A numerical optimization algorithm inspired by the strawberry
plant, arXiv preprint arXiv:1407.7399 (2014)
24. Waterloo North Hydro Inc. Residential Charges and Rates. Accessed 20 May 2016.
https://www.wnhydro.com/en/yourhome/time-of-use-rates.asp
