一种带有随机变异的动态差分进化算法  

Dynamic differential evolution algorithm with random mutation

作　　者：高岳林[1] 刘俊梅[1] 

机构地区：[1]北方民族大学信息与系统科学研究所,银川750021

出　　处：《计算机应用》2009年第10期2719-2722,共4页journal of Computer Applications

基　　金：国家自然科学基金资助项目(60962023);国家社会科学基金资助项目(60962006);宁夏自然科学基金资助项目(NZ0848)

摘　　要：提出一种带有随机变异的动态差分进化算法。在这个算法中,两种不同的变异策略DE/rand/1和DE/best/1通过线性递减加权组合策略产生新的变异策略,以便动态利用DE/rand/1和DE/best/1的优点,并且引入一种指数递增交叉概率算子、线性递减缩放因子和一种随机变异机制以进一步提高算法的全局寻优能力。通过四个标准测试函数的测试结果表明,该算法是一种收敛速度快、求解精度高、鲁棒性较强,更适合求解高维复杂的全局优化问题。The dynamic Differential Evolution （DE） algorithm with random mutation was proposed. In this algorithm, the mutation strategies of DE/rand/l and DE/best/1 were combined by linear dec, reased weight convex combination strategy to produce a new mutation strategy so as to dynamically .use the advantages of DE/rand/1 and DE/best/1. In order to improve the global optimization abilily of DE algorithm, exponent increased crossover probability operator, linear decreasing scaling factor amt random mutation mechanism were introdunced. The test results on the four standard test functions show that the new algorithm has fast convergence, high accuracy and more robustness, more suitable to solve high-dimensional complex global optimization problems.

关 键 词：全局优化 差分进化算法 加权策略 指数递增交叉概率 随机变异 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]