摘要
在分析竞争进化算法原理和特点的基础上,针对旅行商问题的求解,提出一种改进的离散竞争进化算法(IDCE),其中采取三项关键策略:根据个体适值排名计算变异次数、实施逆转子变异算子和并行贪心机制执行多次子变异,目的在于提高算法的全局搜索能力和单位时间内的进化效率。IDCE算法跟另两种离散竞争进化算法对于4个对称旅行商问题算例进行了性能对比,实验结果显示,在解的整体水平、最好解质量以及求解效率上,IDCE算法都优于另两种算法。
On the basis of analyzing competition evolutionary （CE） algorithm＇s principle and characters, aiming at solving traveling salesman problems（TSP）, an improved discrete competition evolutionary algorithm（IDCE） was presented, which applied three critical strategies： computing mutation times by individual＇s fitness rank; doing inverse mutation operator; implementing sub-mutations with parallel greed mechanism. Performance comparison of IDCE for four TSP Benchmarks was provided against two other existing DCEs. Experimental results showed that IDCE algortithm outper- formed two other DCEs in terms of overall quality of solutions, best solution obtained and optimization efficiency.
出处
《计算机与数字工程》
2009年第12期34-37,共4页
Computer & Digital Engineering
关键词
竞争进化算法
竞争推动力
旅行商问题
子变异
嵌套策略
competition evolutionary algorithm, competition impetus, traveling salesman problems, sub-mutation, nesting strategy
