作　　者： (王俊）; (徐建中）; (徐雷）;

机构地区： 南京理工大学,南京210094

出　　处： 《计算机与数字工程》 2017年第8期1465-1468,共4页

摘　　要： 针对混合蛙跳算法易陷入局部最优的情况,提出一种基于讨论机制的混合蛙跳算法。改进算法在原混合蛙跳算法基础上加入了讨论机制,并根据迭代次数动态调整讨论次数,增加算法后期变异性,避免陷入局部最优。论文选择了3个常用的测试函数进行测试,与基本蛙跳算法比较,在计算精度和求解成功率上均有很大的提高。 A discussion mechanism is put forward based shuffled frog leaping algorithm in view of the situation that the shuf-fled frog leaping algorithm is easy to fall into local optimal situation. The scheme increases the variability of the algorithm,and avoids falling into local optima by adding a discussion mechanism,and the number of the discussion is adjusted dynamically. The algorithm is tested by three classical test functions,the result shows that compared with the basic shuffled frog leaping algorithm, the proposed algorithm improves the optimization capability and the success rate greatly.

关 键 词： 混合蛙跳算法 讨论机制 动态调整 计算精度 成功率