No abstract

No abstract

The pheromone based positive feedback approach of ant algorithm is introduced in evolutionary computation of discrete problem, so as to accomplish the optimization of each allele. It ensures stable converge of the algorithm into global optimum. The optimal cutting problem is studied as an example to analyze the performance of the algorithm. The experimental results show the novel performance of the algorithm in the optimization of discrete problem. Key words: pheromone; Ant System; evolutionary computation CLC number: TP 301. 6The optimization of discrete problem is a kind of complex optimization problem. Because evolutionary computation overcomes the traditional weakness of question-dependence, disability of global optimization, etc, it has been successfully applied to many fields^1-41 . For the optimization of discrete problem, a typical problem of searching optimum in discrete search space, binary-coding approaches usually can obtain better results.M. Dorigo 151 presented Ant Algorithm (AA) by the co-operating phenomenon of self-organized colony. It has been successfully applied to many optimization problems in discrete space. It is the substance of AA that the selected probability distribution for each element in the feasible solutionset will be converged to a constant by positive feedback, and the optimal solution is finally obtained.This paper studies the problem of the selected probability on the gene, and directly discusses the optimization process on the gene level. The algorithms presented in this paper overcomes the defect that traditional evolutionary algorithm only optimize the individual while the optimization for gene is neglected. Our approach obtains better results in optimal-cutting problem.

No abstract

No abstract