求解无约束优化问题的仿水循环算法

Water Cycle-like Algorithm for Unconstrained Optimization Problem

针对无约束优化问题,根据自然界水循环过程,提出一种仿水循环算法。其中包括汇流、分流、下渗、蒸发、降雨等粒子选优步骤,通过判断种群数量、粒子质量和位置,对种群和粒子进行相应调整,智能且动态地适应当前搜索的要求,同时采用新的相对重力粒子寻优机制,计算粒子相对重力的方向和大小,引导粒子持续向更优的位置移动。理论分析与仿真结果均表明,该算法能加快种群迭代速度,提高粒子搜索精度,防止粒子陷入局部最优。

According to unconstrained optimization problem,inspiring from water cycle process,a new global optimization algorithm called Water Cycle-like Algorithm（WCA） is presented.Five optimal selected processes of particles,such as confluence,diffluence,infiltration,evaporation and rainfall,are given.Through measure particles＇ population,location and weight,these processes adjust particles and population to adapt the need of current searching,intellectually and dynamically.A new optimal searching mechanism of relative gravity is proposed too,which uses relative gravity to guide particles to a better location.Theoretical analysis and simulation results prove that the algorithm can increase iteration speed,enhance search accuracy,prevent the situation that particles fall into local best.