多策略的郊狼优化算法

郊狼优化算法（coyote optimization algorithm，COA）是最近提出的一种群智能优化算法，具有独特的搜索结构和较好的优化性能。为了进一步提高COA的优化性能，提出了一种多策略的郊狼优化算法（multi-strategy COA，MSCOA）。首先，对于组内最优郊狼，采用一种全局最优郊狼引导的成长策略提高其社会适应能力，对于组内最差郊狼，采用一种最优郊狼引导强化策略强化最差郊狼的能力；其次，对于组内其他郊狼采用一种动态调整信息交流的组内成长策略提升组内郊狼之间的信息共享程度，并将这种组内成长策略与一种改进的迁移策略融合，更进一步提升搜索能力；最后采用动态分组策略减少参数手动设置，提高算法的可操作性。以上多种策略的使用更好地平衡了探索与开采，使算法的性能最大化。大量来自CEC2014测试集的复杂函数实验结果表明，与COA相比，MSCOA具有更强搜索能力、更快的运行速度和更高的搜索效率，与其他优秀优化算法相比，具有更明显的优势。

Multi-strategy coyote optimization algorithm

Coyote optimization algorithm(COA) is a swarm intelligent algorithm proposed recently, which has a unique search structure and good optimization performance. In order to improve COA further, this paper proposed a multi-strategy COA(MSCOA). Firstly, the best coyote in a group used a global best coyote-guided growth strategy to improve its social adaptation, the worst coyote in the group used the best coyote guidance reinforcement strategy to strengthen the worst coyote''s search ability. Secondly, the other coyotes in the group adopted a group growth strategy that dynamically adjusted information interchange to improve information sharing degree among the coyotes in the group, and MSCOA crossed this growth strategy with an improved migration strategy to further improve search ability. Finally, MSCOA adopted a dynamic grouping strategy to reduce the manual setting of parameters and improve operability. Hybridizing the above strategies can keep better balance exploration and exploitation to maximize the performance. A large number of experimental results on complex functions from the CEC2014 test set show that MSCOA has stronger search ability, faster running speed, and higher search efficiency than COA, and has more obvious advantages over quite a few other excellent algorithms.