Professional Documents
Culture Documents
International Journal of Computing Science and Communication Technologies, VOL. 4, NO. 1, July 2011. (ISSN 0974-3375)
Integer
linear
I. INTRODUCTION
Particle swarm optimization (PSO) is a population
based optimization technique, which is an alternative tool
to genetic algorithm (GAs) and other evolutionary
algorithms (EAs) and gained lots of attention in recent
years. PSO is a search technique with reduced memory
requirements, computationally effective and easier to
implement as compared to EAs. In 1995 Kennedy and
Eberhart introduced the PSO as a new heuristic method
[1]. The idea is based on the simulation of the social
behavior of bird flocking and fish schooling. Initially PSO
was designed for continuous optimization problems, but
later a wide variety of challenging engineering and science
applications came into being. Also has a more global
searching ability at the beginning of the run and has
greater local search ability near the end of the run [2].
A linear or non- linear optimization problem, with or
without constraints, in which some or all decision
variables are restricted to have integer values is known as
a Mixed Integer Optimization Problem (MIOP). Such
problems frequently arise in various application fields
such as process industry, finance, engg. design,
Management Science, portfolio selection, automobile
engg., aircraft design and VLSI manufacturing.
The general Mathematical model of an MIOP is:
Min
Subject to:
663
xiL
y
L
i
xiu , i 1,2,
xi
, n1 .
u
i
yi
x1 , x 2 ,
, x n1
y1 , y 2 ,
, y n2 .
, n2 .
acceleration coefficients,
and 2 are random numbers,
uniformly distributed in [0, 1]
is the current position of the
th
particle.
the
th
is the position of
particle achieved based on its own experience.
TECHNIA
International Journal of Computing Science and Communication Technologies, VOL. 4, NO. 1, July 2011. (ISSN 0974-3375)
is
the
and
2
4)
(5)
Eberhart and Shi [4] compared inertia weight PSO
and constricted PSO. It can be seen that equation (3) is
equivalent to equation (4) if the inertia weight w is set to
be
, and
meet
the
conditions
. The PSO algorithm with the
constriction factor can be considered as a special case of
the algorithm with inertia weight since the three
parameters are connected through equation (5). Eberhart
and Shi [4] also showed that constriction factor with
Vmax= Xmax, where Xmax is the upper bound of the
decision variables, provides good results.
The PSO algorithm is shown below
For t = 1 to the max. Bound of the no. of iterations,
For i = 1 to the swarm size.
For j = 1 to the problem dimensionality.
Apply the velocity update equation (1)
664
Using PSO
S
NR
10
25
20
50
25
40
25
70
80
50
50
25
25
50
400
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
ANE
SR (%)
3232
5207
6060
60560
142
152
4532
2803
235
700
770
162
3205
3245
2380
100
100
90
70
100
100
100
80
100
100
100
100
100
100
100
Using RCGA
ANE
SR
172
62
18608
10933
671
84
7447
3571
258
171
299979
77
78
2437
1075
84
85
43
95
100
100
59
41
93
100
71
99
100
92
100
S=Swarm Size,
NR= No. of runs,
ANE=Average no. of function evaluations in
each run.
SR = Rate of success =Percentage of successful runs
to total runs.
IV. CONCLUSION
The PSO algorithm with inertia weight w has been
used for solving constrained integer and mixed integers
optimization problems. The performance of the PSO
algorithm has been compared with RCGA [21] on a set of
15 test problems on the basis of success rate and no. of
functions evaluations. Our results show that the PSO
algorithm out performs RCGA algorithm in most of the
cases. In future we intend to apply PSO algorithm to solve
the larger real life optimization problems.
ACKNOWLEDGEMENT
One of the authors (Ashok Pal) would like to thank
Dr. J.C. Bansal, Assistant Professor, Department of
Mathematics, IIITM-Gwalior (India) for his valuable
guidance in the development of PSO codes for solving the
problems.
TECHNIA
International Journal of Computing Science and Communication Technologies, VOL. 4, NO. 1, July 2011. (ISSN 0974-3375)
V. APPENDIX
s.t.
5.5(1 x1 ) 7 x 2
6 x3
x1 (2 x1 1)
0.9(1 exp( 0.5 x 2 )
50
1 ( x1 (2 x1 1)
0.8(1 exp( 0.4 x3 )
s.t.
.
The global optimal solution is
= (1, 3.514, 0; 99.245209).
Problem 8: This problem is taken from [13]. It is also
given in [10, 11].
1
4
2
2
s.t.
.
The global optimal solution is
= (14.095, 0.84296; -6961.741616).
Problem 5: This problem is taken from [14].
Min
(integers).
s.t.
The
(Integers).
= (2, 0,
5; -68).
Problem
6:
This
problem
represents
a
quadratic capital budgeting problem, taken from [15] and
it is also given in [11].
Min f
665
global
optimal
solution
is
max f
TECHNIA
International Journal of Computing Science and Communication Technologies, VOL. 4, NO. 1, July 2011. (ISSN 0974-3375)
Min
i 1
x3
Where
s.t.
= (0, 1, 1, 1, 0, 1, 1, 0; 0.94347).
Problem 10: This problem is taken from [9] and is
also given in [16].
Min
s.t.
,
1
2
1
3
2
s.t.
(integers).
The global optimal solution is (
= (1, 1, 1, 1, 2; 8).
Problem 11: This problem is taken from [9] and is
also reported in [16].
Min
s.t.
(integers).
The global optimal solution is
= (16, 22, 5, 5, 7; 807).
Problem 15: This problem is taken from [20] and is
also given in [12].
Maxf ( x ) 215 x1 116 x 2 670 x3 924 x 4 510 x5
600 x6 424 x7 942 x8 43x9 369 x10 408 x11
52 x12 319 x13 214 x14 851x15 394 x16 88x17
124 x18 17 x19 779 x 20 278 x 21 258 x 22 271x 23
281x 24 326 x 25 819 x 26 485 x 27 454 x 28
297 x 29 53x30 136 x31 796 x32 114 x33 43x34
80 x35 268 x36 179 x37 78x38 105 x39 281x 40 ;
s.t.
Integers.
The
(
global
optimal
solution
is
) = (0, 2, 4, 0, 2, 1, 4; 14).
(integers).
The global optimal solution is
42.632).
11x9 11x10
x16
666
2 x11
2 x18 3 x19
2 x 23
2 x 24
2 x30
x31 9 x32
10 x37
= (1, 3; -
x17
8 x38
x 25
6 x39
4 x 20
2 x 26
x 27
x33 9 x34
x 40
2 x15
7 x 21 6 x 22
8 x 28 10 x 29
2 x35
25000;
4 x36
TECHNIA
International Journal of Computing Science and Communication Technologies, VOL. 4, NO. 1, July 2011. (ISSN 0974-3375)
15 x9
8 x10 16 x11
7 x16
2 x17
2 x18
4 x19 3x 20
2 x21 13 x 22
8 x 23
2 x 24 3x 25
4 x 26 3x 27
2 x 28
x 29
x34 8 x35
6 x36
10 x30
3x37
x12
6 x31 3x32
4 x38
6 x39
4 x33
2 x 40
3 x9
7 x10
2 x11 16 x12
8 x16
9 x17
7 x18
3 x 23 14 x 24
28 x13
3x13
3x14
2 x32
8 x33
3 x34
x37
2 x38
6 x39
5 x 40
25000;
20
xi
Where
2 x35
[10]
9 x15
x 22
6 x 27 16 x 28
x31
xi
[8]
[9]
6 x19 16 x 20 12 x 21
7 x 25 13x 26
[7]
7 x15
25000;
2 x30
0
2 x14
[11]
3x 29
7 x36
[12]
99; i 1,2,..............20;
99; i
21,22,............40;
[13]
( i =1,
[14]
[15]
REFERENCES
[1]
[16]
[4]
J. Kennedy, and R.
Proceedings IEEE International Conference Neural Networks, 4,
1942 1948, 1995.
M.S. Arumugam, M.V.C. Rao, and A.W.C.
effective particle swarm optimization like algorithm with
extrapolation
308 320, 2009.
Y. Shi, and R.C. Eberhart,
Proceedings of the IEEE International Conference on Evolutionary
Computation, Piscataway, NJ: IEEE Press, 69 73, 1998.
R.C. Eberhart, and Y.
[5]
[2]
[3]
[6]
mic
667
[17]
[18]
[19]
[20]
[21]
[22]
M.
Congress on Evolutionary
Computation,
Washington
D.C.,
p.
1951 1955, 1999.
M. Clerc, and J.
and convergence in a multiTransactions
on
Evolutionary
Computation,
6,
58 73, 2002.
H.M. Salkin, Integer Programming, Eddison Wesley Publishing
Com., Amsterdam, 1975.
C.A.
Floudas,
Nonlinear
Mixed-integer
Optimization.
Fundamentals and Applications, Oxford University Press, New
York, USA, 1995.
M.F. Cardoso, R.L. Salcedo, S.F. Azevedo, D. Barbosa, A
simulated annealing approach to the solution of minlp problems,
Computers and Chemical Engineering 21, p. 1349 1364, 1997.
C. Mohan, H.T. Nguyen, A controlled random search technique
incorporating the simulating annealing concept for solving integer
and mixed integer global optimization problems, Computational
Optimization
and
Applications,
14,
p. 103 132, 1999.
L.P. Costa, E. Oliveria, Evolutionary algorithms approach to the
solution of mixed integer non-linear programming problems,
Computers and Chemical Engineering, 21, p. 257 266, 2001.
Y.X. Li, M. Gen, Nonlinear mixed integer programming problems
using genetic algorithm and penalty function, in: Proceeding of the
IEEE International Conference on Systems, Man and Cybernatics,
vol. 4, p. 2677 2682, 1996.
G.R. Kocis, I.E. Grossmann, Global optimazation of nonconvex
mixed- integer nonlinear programming (minlp) problems in process
synthesis, Industrial & Engineering Chemistry Research, 27, p.
1407 1421, 1998.
H.T. Nguyen, Some Global Optimization Techniques and Their Use
in Solving Optimization Problems in Crisp and Fuzzy
Environments, Ph.D. Thesis, Department of Mathematics,
University of Roorkee, Roorkee, India, 1996.
O. Berman, N. Ashrafi, Optimization models for reliability of
modular software systems, IEEE Transactions on Software
Engineering, 19 p. 11 19, 1993.
M.S. Bazaraa, H.D. Sherah, C.M. Shetty, Nonlinear Programming:
Theory and Algorithms, second ed., John Wiley and Sons, Asia,
2004.
D.M. Himmelblau, Applied Nonlinear Programing, McGraw Hill,
New York, USA, 1972.
W. Conley, Computer Optimization Techniques, Petrocelli Books,
Newjersy, USA, 1984.
K. Deep, K.P. Singh, M.L. Kansal, C.
algorithm for solving integer and mixed integer optimization
Applied Mathematics and Computation, 212(2), p. 505
518, 2009.
K. Deb, An efficient constraint handling method for genetic
algorithms, Computer Methods in Applied Mechanics and
Engineering, 186, p. 311 338, 2000.