International Journal of Electrical Engineering & Technology (IJEET)

Volume 7, Issue 2, March-April, 2016, pp.69–78, Article ID: IJEET_07_02_008

Available online at
http:// http://www.iaeme.com/IJEET/issues.asp?JType=IJEET&VType=7&IType=2
ISSN Print: 0976-6545 and ISSN Online: 0976-6553
Journal Impact Factor (2016): 8.1891 (Calculated by GISI) www.jifactor.com
© IAEME Publication

APPLICATION OF MODIFIED (PSO) AND

SIMULATED ANNEALING ALGORITHM
(SAA) IN ECONOMIC LOAD DISPATCH
PROBLEM OF THERMAL GENERATING
UNIT
Youssef Nagem Amhamad and Jyoti Shrivastava
Department of Electrical Engineering, SSET, SHIATS, Allahabad

ABSTRACT
This paper deals with the problem of economic load dispatch (ELD) in
thermal generating unit. The main issue of generating unit to minimize the
cost of generation so modified Particle Swarm Optimization (PSO) method is
proposed for solving this issue. The modified PSO method was developed
through simulation of a simplified social system and has been found to be
robust in solving continuous nonlinear optimization problems in terms of
accuracy of the solution and computation time The proposed algorithm is
applied for the ELD of six unit thermal plant systems and the performance of
the proposed modified PSO method is compared with the Simulated Annealing
Algorithm (SAA). All results obtained through MATLAB Simulink software.
The comparison of results shows that the proposed modified PSO method was
indeed capable of obtaining higher quality solutions efficiently for ELD
problems within less computation time.
Key words: Economic Load Dispatch (ELD), Particle Swarm Optimization
(PSO), Simulated Annealing Algorithm (SAA), MATRIX LABORATORY
(MATLAB).
Cite this Article: Youssef Nagem Amhamad and Jyoti Shrivastava,
Application of Modified (PSO) and Simulated Annealing Algorithm (SAA) In
Economic Load Dispatch Problem of Thermal Generating Unit. International
Journal of Electrical Engineering & Technology, 7(2), 2016, pp. 69–78.
http://www.iaeme.com/IJEET/issues.asp?JType=IJEET&VType=7&IType=2

1. INTRODUCTION
The Economic Load Dispatch (ELD) problem is one of the fundamental issues in
power system operation. The main objective is to reduce the cost of energy production
taking into account the transmission losses. While the problem can be solved easily if

http://www.iaeme.com/IJEET/index.asp 69 editor@iaeme.com
 Youssef Nagem Amhamad and Jyoti Shrivastava

the incremental cost curves of the generators are assumed to be monotonically

increasing piece-wise linear functions, such an approach will not be workable for
nonlinear functions in practical systems. In the past decade, conventional optimization
techniques such as lambda iterative method, linear programming and quadratic
programming have been successfully used to solve power system optimization
problems such as Unit commitment, Economic load dispatch, Feeder reconfiguration
and Capacitor placement in a distribution system. For highly non-linear and
combinatorial optimization problems, the conventional methods are facing difficulties
to locate the global optimal solution. Recently there is an upsurge in the use of
modern evolutionary computing techniques in the field of power system optimization.
Particle Swarm Optimization (PSO), first introduced by Kennedy and Eberhart, is one
of the modern heuristic algorithms. It was developed through simulation of a
simplified social system, and has been found to be robust in solving continuous non-
linear optimization problems. The PSO technique can generate high-quality solutions
within shorter calculation time and stable convergence characteristics than other
stochastic methods. All the particles in PSO are kept as members of the population
through the course of a run (a run is defined as the total number of generation of the
evolutionary algorithms prior to termination). It is the velocity of the particle which is
updated according to its previous best position of its companions.
The proposed method results have been compared with the Simulated Annealing
Algorithm (SAA). Simulated Annealing (SA) has been proved to be effective and
quite robust in solving the optimization problems. SA can provide near global
solutions and can also handle effectively the discrete control variables. SA does not
stick into local optima because SA begins with many initial points and search for the
most optimum in parallel. SA considers only the pay-off information of objective
function regardless whether it is differentiable or continuous. Consequently, the most
realistic cost characteristic of power plants can be formulated . Simulated Annealing
(SA) is a stochastic optimization technique which is based on the process of annealing
in Thermodynamics proposed by Kirkpatrick. Mathematical model of simulated
annealing describes how the molecules of liquidated metal move freely with respect to
each other and by gradually cooling (thermodynamic process of annealing) thermal
mobility are lost. The atoms start to get arranged and finally form crystals, having the
minimum energy which depends on the cooling rate.

2. MATHEMATICAL MODEL
A. Objective function
The objective of the economic load dispatch is to minimize the generating cost based
on the premise that constraints are satisfied. Coal consumption (standard coal) is
selected as the optimization objective in order to emphasize the main aspects, simplify
the mathematical model and make the problem comparable. Then the mathematical
description of ELD's objective function is:
The objective of the ELD problem is to minimize the total fuel cost.
Mathematically it can be represented as
Minimize Ct = (1)

Ct= (2)
Where Ct Fuel cost of the system.

http://www.iaeme.com/IJEET/index.asp 70 editor@iaeme.com
 Application of Modified (PSO) and Simulated Annealing Algorithm (SAA) In Economic
Load Dispatch Problem of Thermal Generating Unit

Ci Fuel cost of the generating unit of the system. and are cost coefficients of
generator i.
Output power generation of unit i.
The ELD problem is subjected to the following constraints, the power balance
equation,
= (3)
The total Transmission loss,
= (4)
In addition, power output of each generator has to fall within the operation limits
of the generators as shown below,
≤ (5)
In the power balance criterion, an equality constraint must be satisfied, as shown
in equation (3). The generated power should be the same as the total load demand plus
total line
Losses .The generating power of each generator should lie between maximum and
minimum limits represented by equation (5), where Pi is the power of generator i (in
MW); n is the number of generators in the system; PD is the system’s total demand
(in MW); PL represents the total line losses in (MW) and min Pi and max Pi are,
respectively, the output of the minimum and maximum operation of the generating
unit i in (MW).

3. OVERVIEW OF PSO
PSO, as an optimization tool, provides a population-based search procedure in which
individuals called particles change their position (states) with time. In PSO system
particles fly around in a multi- dimensional search space. During flight, each particle
adjusts its position according to its own experience and the experience of
neighbouring particles, making use of the best position encountered by it and
neighbours. The swarm direction of a particle is defined by the set of particles
neighbouring the particle and its history experience. Instead of using evolutionary
operation to manipulate the individuals, like in other evolutionary computational
algorithms, each individual in PSO flies in the search space with a velocity which is
dynamically adjusted according to its own flying experience and its companions
flying experience.
Let x and v denote a particle co-ordinate (position) and its corresponding flight
speed (velocity) in a search space respectively. Therefore, each ith particle is treated
as a volume less particle, represented as xi= (xi1, xi2 …xid) in the d -dimensional
space. The best previous position of the ith particle is recorded and represented as
pbesti=(pbesti1,pbesti2,…….. pbestid).The index of the best particle among all the
particles is treated as global best particle, is represented as gbestd. The rate of velocity
for particle ‘i’ is represented as vi= (vi1, vi2……...vid). The modified velocity and
position of each particle can be calculated using the current velocity and the distance
from pbestid to gbestd as shown in the following formulas,

(7)
(8)

http://www.iaeme.com/IJEET/index.asp 71 editor@iaeme.com
 Youssef Nagem Amhamad and Jyoti Shrivastava

In the above equation, C1 has a range (1.5, 2), which is called self-confidence
range; C2 has a range (2, 2.5), which is called swarm range.
The parameter Vd max determines the resolution, or fitness, with which regions are
to be searched between the present position and the target position .If Vd max is too
high, particles may fly past good solutions. If Vd max is too small, particles may not
explore sufficiently beyond local solutions. In many experiences with PSO, Vd max
was often set at 10-20% of the dynamic range on each dimension.
The constants C1and C2 pull each particle towards pbest and gbest positions. Low
values allow particles to roam far from the target regions before being tugged back.
On the other hand, high values result in abrupt movement towards, or past, target
regions. Hence, the acceleration constants C1 and C2 are often set to be 2.0 according
to past experiences. Suitable selection of inertia weight ‘ ω ’ provides a balance
between global and local explorations, thus requiring less iteration on average to find
a sufficiently optimal solution. As originally developed, ω often decreases linearly
from about 0.9 to 0.4 during a run. In general, the inertia weight w is set according to
the following equation,

ɷ= (9)
where ω - inertia weight factor
ωmax - maximum value of weighting factor
ωmin - minimum value of weighting factor
itmax - maximum number of iterations
it - current number of iteration

4. APPLICATION OF PSO METHOD TO ECONOMIC LOAD

DISPATCH
In population based optimization algorithm, there is an avid necessity of improving
the performance of existing algorithm. This can be implemented by two means. Either
the basic operators of algorithm should be redesigned or proper tuning of adjustable
parameters should be done. In proposed variant of PSO, proper tuning of adjustable
parameters like w, c1 and c2 are done so that it can reach on optimal solution as early
as possible. In existing PSO, the values of adjustable parameters like w,c1 and c2 are
independent from the values of gbest and pbest. This is the main reason why PSO
converges very slow toward optimal solution. These values are may remain fixed or
may vary according to the number of generation. In proposed algorithm, a relationship
has been established between adjustable parameters and the values of gbest and pbest
and the values of w, c1 and c2 are set accordingly. The value of c1 has been set to
pbestval-out /gbestval and c2 has been set to repmat (gbestval,ps,1)-out /gbestval so
that particles may exploit good solutions as early as possible. Moreover to improve
convergence rate of the algorithm a very high inertia weight equivalent to gbestval-
pbestavg/gbestval has set. These values motivate particles to exploit solution around
good regions and capture optimal solution as early as possible. The PSO algorithm
was utilized mainly to determine the optimal allocation of power among the units,
which were scheduled to operate at the specific period, thus minimizing the total
generation cost.

http://www.iaeme.com/IJEET/index.asp 72 editor@iaeme.com
 Application of Modified (PSO) and Simulated Annealing Algorithm (SAA) In Economic
Load Dispatch Problem of Thermal Generating Unit

A. Calculation process of the proposed method

This paper presents a quick solution to the constrained ELD problem using the PSO
algorithm to search optimal or near optimal generation of each unit. The sequential
steps of the proposed PSO method are given below.
Step 1: Initialize randomly the individuals of the population according to the limit of
each unit including individual dimensions, searching points, and velocities. These
initial individuals must be feasible candidate solutions that satisfy the practical
operation constraints.
Step 2: To each chromosome of the population the dependent unit output Pd will be
calculated from the power balance equation and Bmn coefficient matrix.
Step 3: Calculate the evaluation value of each individual Pgi, in the population using
the evaluation function C given by (2).
Step 4: Compare each individual’s evaluation value with its pbest. The best
evaluation value among the pbests is denoted as gbest.
Step 5: Modify the member velocity v of each individual Pg, according to equation
(7)
Step 6: Check the velocity components constraint occurring in the limits from the
following conditions,
If ,

,
Where

Step 7: Modify the member position of each individual Pg according to (8)

Step 8: If the evaluation value of each individual is better than previous pbest, the
current value is set to be pbest. If the best pbest is better than gbest, the value is set to
be gbest.
Step 9: If the no.of iterations reaches the maximum, Go to step 10.Otherwise, go to
step 2.
Step 10: The individual that generates the latest gbest is the optimal generation power
of each unit with the minimum total generation cost.

5. OVERVIEW OF SIMULATED ANNEALING ALGORITHM

(SAA)
Simulated Annealing (SA) algorithm is a nature-inspired method which is adapted
from process of gradual cooling of metal in nature. In the metallurgical annealing
process, a solid is melted at high temperature until all molecules can move about
freely and then a cooling process is performed until thermal mobility is lost. The
perfect crystal is the one in which all atoms are arranged in a low level pattern, so
crystal reaches the minimum energy. It is basically a stochastic optimization
technique which is based on the principles of statistical engineering. The search for
global minima of a multidimensional function is quite a complex problem especially
when a big number of local minima correspond to the respective function. The main
purpose of the optimization is to prevent hemming about to local minima. The

http://www.iaeme.com/IJEET/index.asp 73 editor@iaeme.com
 Youssef Nagem Amhamad and Jyoti Shrivastava

originality of the SA method lies in the application of a mechanism that guarantees

the avoidance of local minima.

A. The Process of Annealing in Thermodynamics

At high temperature, the metal is in liquid stage. The molecules of liquidated metal
move freely with respect to each other, via gradual cooling (thermodynamic process
of annealing) thermal mobility is lost. The atoms start to get arranged and finally form
crystals, having the minimum energy which depends on the cooling rate. If the
temperature is reduced at a very fast rate, the crystalline state transforms to an
amorphous structure, a meta-stable state that corresponds to a local minimum of
energy. Annealing process of metal influences SA algorithm. If the system is at a
thermal balance for given temperature T, then the probability PT(s) that it has a
configuration s depends on the energy of the corresponding configuration E(s), and is
subject to the Boltzmann distribution:

(10)

Where, k is the Boltzmann constant and the sum includes all possible states
W. Metropolises were the first to suggest a method for calculating a distribution of a
system of elementary particles (molecules) at the thermal balance state. Let the
system has a configuration g, which corresponds to energy E(g). When one of the
molecules of the system is displaced from its starting position, a new state σ occurs
which corresponds to energy E(σ). The new configuration is compared with the old
one. If E(σ)≤ E(g) , then the new state is accepted. If E(σ)>E(g), then the new state is
accepted with probability :

(11)

Table 1 Connection between Thermodynamic and Combinatorial Optimization:

Thermodynamics simulation Combinatorial Optimization

System state Feasible Solutions

Energy Cost

Change of state Neighboring Solutions

Temperature Control Parameter

Frozen state Heuristic Solution

B. Control parameters of SA algorithm

For the successful application of the SA algorithm is the annealing schedule is vital,
which refers to four control parameters that directly influence its convergence (to an
optimized solution) and consequently its efficiency. The parameters are the following:
a) Starting Temperature
b) Final Temperature
c) Temperature Decrement
d) Iterations at each Temperature

http://www.iaeme.com/IJEET/index.asp 74 editor@iaeme.com
 Application of Modified (PSO) and Simulated Annealing Algorithm (SAA) In Economic
Load Dispatch Problem of Thermal Generating Unit

a) Starting Temperature The starting temperature must be set to a big enough value,
in order to make possible a big probability of acceptance for non optimized solutions
during the first stages of the algorithm’s application. However, if the value of the
starting temperature gets too big, SA algorithm becomes non-effective because of its
slow convergence and in general, the optimization process degenerates to a random
walk. On the contrary, if the starting temperature is low then there is a greater
probability of achieving local minima. There is no particular method for finding the
proper starting temperature that deals with the entire range of problems.
b) Final Temperature During the application of the SA algorithm it is common to let
the temperature fall to zero degrees. However, if the decrement of the temperature
becomes exponential, SA algorithm can be executed for much longer time. Finally,
the stopping criteria can either be a suitable low temperature or the point when the
system is “frozen” at current temperature.
c) Temperature Decrement Since the starting and final temperatures have been
defined, it is necessary to find the way of transition from the starting to the final
temperature. The way of the temperature decrement is very important for the success
of the algorithm suggested the following way to decrement the temperature:
T (t) =d/log(t) (12)
Where d is a positive constant.
An alternative is the geometric relation: T(t)=a.t (13)
Parameter a, is a constant near 1. In effect, its typical values range between 0.8
and 0.99.

d) Iterations at each Temperature

For increased efficiency of the algorithm, the number of iterations is very important.
Using a certain number of iterations for each temperature is the proper solution. The
temperature decrement should take place at a really slow pace that can be expressed
as
T(t)=t/(1+βt) (14)
Where, β takes a very low value.

SA Algorithm Implementation of ELD Problems

Step 1: Initialization of temperature, T, parameter α (α is a constant smaller than but
close to 1), and maximum number of generations. Find, randomly, an initial feasible
solution, which is assigned as the current solution Si and perform ELD in order to
calculate the total cost.
Step 2: Set the iteration counter to i=1
Step 3: Find a neighbouring solution Sj through a random perturbation of the counter
one and calculate the new total cost, Fcost.
Step 4: If the new solution is better, we accept it, if it is worse, we calculate the
deviation of cost ΔS=Sj-Si and generate a random number uniformly distributed over
Ω∊ (0, 1).

If ≥ Ω∊ (0, 1). Accept the new solution Sj to replace Si.

http://www.iaeme.com/IJEET/index.asp 75 editor@iaeme.com
 Youssef Nagem Amhamad and Jyoti Shrivastava

Step 5: If the stopping criterion is not satisfied, reduce temperature using parameter
α:
T(t) =α. t and return back to Step 2.

6. RESULT
To verify the feasibility of the proposed modified PSO method, six generating unit
and ten generating unit has been taken into consideration.
Case 1: for 6 generating uni:tThe result of Proposed method and Simulated
Annealing (SA) algorithm is compared in the following tables(1) :-
PD = 700 PL P1 P2 P3 P4 P5 P6 Cost
i
(MW) (MW) (MW) (MW) (MW) (MW) (MW) (MW) (Rs/Hr)
M PSO 19.42 28.103 10.000 118.28 118.58 232.05 212.42 36912.22 6000
SAA 19.43 28.29 10.00 118.9 118.6 230.7 212.7 36912.52 242216

Figure 1.Convergence characteristics of 6 unit system (PD=700MW)in case (pso).

Response
x 10 of the System between the Cost vs the the number of sessions
5

2.5

2
cost

1.5

0.5

0
0 0.5 1 1.5 2 2.5 3
T--(number of sessions) x 10
5

Figure 2 Convergence characteristics of 6 unit system (PD=700MW) in case (SAA).

http://www.iaeme.com/IJEET/index.asp 76 editor@iaeme.com
 Application of Modified (PSO) and Simulated Annealing Algorithm (SAA) In Economic
Load Dispatch Problem of Thermal Generating Unit

Case 2:for 10 Generating unit The result of Proposed method and Simulated
Annealing (SA) algorithm is compared in the following tables(2) :-

PD = 2000 PL P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 Cost

i
(MW) (MW) (MW) (MW) (MW) (MW) (MW) (MW) (MW) (MW) (MW) (MW) (Rs/Hr)
Modified
87.017 55 80 106.86 99.315 82.932 83.853 300 339 470 470 111266.1 5000
PSO

SAA 83.236 54.99 77.658 104.96 103.55 160.00 187.28 240.27 264.63 440.05 449.80 114383.8 28523

From the comparative analysis of above table it is clear that the number of
iteration and the cost of modified PSO is less than the number of iteration and the cost
of SAA method.

7. CONCLUSION
In this paper, the proposed modified PSO method was successfully employed to solve
the ELD problem with all the constraints and compared with the Simulated Annealing
algorithm (ASS). The proposed method has been demonstrated to have superior
features including high quality solution, stable convergence characteristics, and less
computation time. Many non-linear characteristics of the generators can be handled
efficiently by the proposed method. The comparison of results for the test cases
clearly shows that the proposed method was indeed capable of obtaining higher
quality solution efficiently for ELD problems.

ACKNOWLEDGEMENT
The authors are thankful to the authorities of Sam Higginbottom Institute of
Agriculture, Technology and Sciences, Allahabad for providing all facilities to
complete this work.

REFERENCES
[1] Bakirtzis, V. Petridis, and S. Kazarlis, Genetic algorithm solution to the economic
dispatch problem, IEE proceedings on Gen., Transm. Dist., vol. 141, No.4, pp.
377-382, July 1994.
[2] D. C.Walters and G. B. Sheble, Genetic algorithm solution of economic dispatch
with valve point loading, IEEE Trans. on Power Systems, vol.10, No.8, pp. 1325-
1332, Aug. 1993.
[3] R. C. Eberhart and Y. Shi, Comparison between genetic algorithms and particle
swarm optimization, Proceedings of IEEE Int. Conference on Evolutionary
Computation, pp.611–616, May 1998.
[4] M. Sudhakaran, Dr S M R Slochanal, Integrating Genetic Algorithms and Tabu
Search for Emission and Economic Dispatch Problems”, IE (I) Journal –EL, vol
86, pp.39-43, June 2005.
[5] Nidul Sinha, Chakrabarti.R, and Chattopadhyay.P.K, Fast Evolutionary
Computing Techniques for Short-Term Hydrothermal Scheduling, IEEE Trans.
on Power Systems, vol.18, No.1, pp. 214-220, February 2003.
[6] Y. Shi and R. Eberhart, A modified particle swarm optimizer, Proceedings of
IEEE Int. Conference on Evolutionary Computation, pp.69-73, May 1998.

http://www.iaeme.com/IJEET/index.asp 77 editor@iaeme.com
 Youssef Nagem Amhamad and Jyoti Shrivastava

[7] P. J. Angeline, Using selection to improve particle swarm optimization,

Proceedings of IEEE Int. Conference on Evolutionary Computation, pp. 84-89,
May 1998.
[8] P. Venkatesh, Dr P S Kannan, M Sudhakaran, Application of Computational
Intelligence to Economic Load Dispatch, IE (I) Journal–EL Vol 81, pp.39-43,
September 2000.
[9] H. Yoshida, K. Kawata, Y. Fukuyama, S. Takayama, and Y. Nakanishi, A
particle swarm optimization for reactive power and voltage control considering
voltage security assessment, IEEE Trans. on Power Systems, vol.15, No.4, pp.
1232–1239, Nov. 2000.
[10] Zwe-Lee Gaing, Particle Swarm Optimization to solving the Economic Dispatch
Considering the Generator Constraints, IEEE Trans. On Power Systems, Vol.18,
No.3, pp. 1187-1195, August 2003.
[11] Bo Lu and Mohammad Shahidehpour, Unit Commitment with Flexible
Generating Units, IEEE Trans. on Power Systems, Vol.20, No.2, pp. 1022-1034,
May 2005.
[12] Yong Fu, Mohammad Shahidehpour and Zuyi Li, Security Constrained Unit
commitment with AC constraints, IEEE Trans. on Power Systems, Vol.20, No.3,
pp. 1538-1550, August 2005.
[13] Kamlesh Kumar Vishwakarma, Hari Mohan Dubey, Simulated Annealing Based
Optimization for Solving Large Scale Economic Load Dispatch Problems.
International Journal of Engineering Research & Technology (IJERT) Vol. 1
Issue 3, May - 2012 ISSN: 2278-0181.
[14] Ashish Dhamanda and A.K. Bhardwaj, Automatic Generation Control of Thermal
Generating Unit by Using Conventional and Intelligent Controller. International
Journal of Electrical Engineering & Technology, 5(10), 2014, pp. 56–64.
[15] Sanjay Mathur, Shyam K. Joshi and G.K. Joshi, Development of Controller For
Economic Load Dispatch by Generating Units Under Varying Load Demands.
International Journal of Electrical Engineering & Technology, 4(4), 2013, pp.
159–171.
[16] Mr.Vijay Kumar, Dr.Jagdev Singh, Dr.Yaduvir Singh, Dr.Sanjay Sood, Design &
Development of Genetic Algorithms for Economic Load Dispatch of Thermal
Generating Units. International Journal of Computer Engineering & Technology,
3(1), 2012, pp. 59–75
[17] M. Venkatesh, Ramakrishna Raghutu. Economic Load Dispatch Using Simulated
Annealing Algorithm. International Research Journal of Engineering and
Technology (IRJET) e-ISSN: 2395-0056 Volume: 02 Issue: 03 | June-2015.

http://www.iaeme.com/IJEET/index.asp 78 editor@iaeme.com