Compressed Genetic Algorithm

Worasait Suwannik, Naris Kunasol, and Prabhas Chongstitvatana

Proc. of Northeastern Computer Science and Engineering Conference, Thailand, 2005

Abstract

Genetic Algorithm is a search method that imitates natural evolution. Genetic Algorithm searches the solution using a population of binary-number chromosomes. When the size of a chromosome is increasing, the search time will be longer. This is because the search space is rapidly increasing. This paper reports an improvement of the efficiency of Genetic Algorithm by reducing the size of a chromosome with data compression. The proposed method is called "Compressed Genetic Algorithm" (CGA). The evolution process in CGA manipulates the compressed chromosome. When CGA is used to solve a problem, its chromosome will be shorter than an ordinary chromosome in Genetic Algorithm. The experiment is carried out with Onemax problem of the size 128 bits. The result shows that CGA is faster than Genetic Algorithm upto 805 times. For the robot arm control problem, CGA is faster than Genetic Algorithm upto 4.5 times.