Previous [ 1] [ 2] [ 3] [ 4] [ 5] [ 6] [ 7] [ 8] [ 9] [ 10] [ 11] [ 12]

@

Journal of Information Science and Engineering, Vol. 20 No. 5, pp. 1019-1034 (September 2004)

A New Efficient Encoding Mode of Genetic Algorithms for
the Generalized Plant Allocation Problem

Young-Chang Hou and Ying-Hua Chang
Department of Information Management
National Central University
Chungli, 320 Taiwan

This study proposes a novel, efficient means of encoding genetic algorithms to solve the generalized plant allocation problem. The problem relates to allocating products across plants to minimize a total cost function. The proposed encoding method can reduce the search space of solutions more efficiently than the penalty encoding method does. The new encoding method thus exhibits higher performance. It need involve only a few more generations to yield sufficiently good solutions when the number of plants is increased. The penalty encoding method, however, requires many more generations to yield the same solutions. Additionally, a new simultaneous crossover and mutation operation is proposed to enable the new method of encoding chromosomes to run correctly following standard genetic algorithm procedures. In addition to the mathematical certification, the performance of this approach is evaluated using some test problems of various sizes. Solutions obtained by this approach are always efficient.

Keywords: genetic algorithms, efficient encoding scheme, plant allocation, combinatorial optimization problems, penalty encoding

Full Text () Retrieve PDF document (200409_12.pdf)

Received November 4, 2002; revised February 14, 2003; accepted May 15, 2003.
Communicated by Hsu-Chun Yen.