自适应蚁群差分进化算法

针对传统差分进化算法在求解高维复杂问题时存在通用性差、鲁棒性低、收敛速度慢和求解精度低等问题,提出一种基于蚁群算法的自适应多模式差分变异策略.算法在每代进化中,个体根据各变异进化模式上的信息素大小,采用轮盘赌选择策略选择变异算子,并根据各变异算子对优化所做贡献的大小对信息素进行动态更新,贡献大的变异算子可以获得更多被选择的机会,使得各变异算子发挥其最大性能,从而提高算法的收敛速度和通用性.对5个高维的benchmark函数进行算法验证,实验结果表明,该算法很好的提高了差分进化算法的通用性和鲁棒性,有效地克服了收敛速度慢和早熟等问题.

Self-adaptive ant colony differential evolution algorithm

Aiming at the defects of low generality and robustness,slow rate of convergence,low accuracy and easily falling into local optimum in the traditional differential evolution algorithm,a self-adaptive and multi-strategy differential mutation based on ant colony optimization algorithm is proposed for high-dimensional problems.According to the pheromone,the individual selects differential operator with roulette selection operator strategy in each generation,and updates the pheromone dynamically based on the contribution of each mutation evolution model.The model which makes a greater contribution will be chosen.Finally,five high-dimensional benchmark functions are used to test this algorithm.Experimental result indicates that the proposed algorithm effectively avoid the premature phenomenon and the slow convergence velocity,while being highly robust and good generality.