A Kind of Knowledge Evolution Algorithm and Its Application

摘要

Some traditional algorithms would easily trap into the drawback of local optimum when solving all kinds of optimization problems. To solve this problem, this paper proposes a kind of knowledge evolutionary algorithm. This algorithm forms a knowledge space on the basis of local solutions in each swarm space by using surmise operation and rebut operation in the knowledge evolution algorithm. And then updates the knowledge space through co-evolution to find out the optimum solution. The baseline data of function optimization problems, the knapsack problems and the permutation flow shop scheduling problems arc tested by this algorithm. By contrasting with the algorithms such as genetic algorithm as well as the standard PSO. it attests the feasibility and effectiveness of the proposed algorithm.