Restricting Mutation Method for Binary Genetic Algorithms

Genetic algorithm (GA) is an adaptive algorithm which can be applied on the most kinds of functions to find extremums. This algorithm has an acceptable speed in comparison with the other existing methods for finding minimum or maximum of a function. GA is one of methods in finding extermums by intelligence. Some other methods like PSO are introduced for improving GA . A PSO tries to steer genes toward the elite gene by giving velocity coefficient to the genes which pull them toward the global extermum. Some other variants of GA are introduced for speed up the GA by changing mutation operator in GA like Bacterial Evolutionary Algorithm (BEA) and Jumping gene algorithms. In BEA method, the algorithm tries to find the best gene by applying mutation on all bits of the genes and comparing all results in each iteration. Jumping gene method adds a new operator after crossover on parents and cuts or copies one part of a parent chromosome in other parent and tries to make attribute of elite parents in others. In this paper we introduced new method for changing the place of applying mutation on genes for speed up GA method in the aspect of finding global minimum faster and using minimum recalling of the cost function. This method is compared with introduced methods and results show the capability of this method in finding global minimum better than others

Keywords:

genetic algorithm , mutation , global extremum finding , restricting mutation method