Research On Inertia Weight Of Particle Swarm Optimization

With the continuous development of society,economy and science,there are often non convex,non differentiable,and even some objective function can not be expressed optimization problems.To solve this kind of optimization problems,the traditional optimization problems are invalid,so the intelligent optimization algorithm appeared.At present,particle swarm optimization is one of the most widely used intelligent optimization algorithms,and inertia weight is an important parameter of particle swarm optimization,so the inertia weight of particle swarm optimization algorithm is one of the main research directions of particle swarm optimization algorithm.In this paper,the inertia weight of particle swarm optimization algorithm is studied,the main work is as follows:1.On the basis of consulting a large number of literatures,this paper reviews the inertia weight improvement strategy of particle swarm optimization algorithm.The inertia weight improvement strategy of particle swarm optimization algorithm is discussed from five aspects: constant inertia weight,random inertia weight,time-varying inertia weight,chaotic inertia weight and adaptive inertia weight.It provides a reference for further research on inertia weight of particle swarm optimization algorithm.2.An adaptive particle swarm optimization algorithm based on inertia weight set is proposed.In this algorithm,we select 12 commonly used particle swarm optimization algorithms with inertia weight,and applied these algorithms to a typical test set.Five particle swarm optimization algorithms are selected by nonparametric test method,and constitute a weight set.Then a concept of K-step evolution degree is defined,according to the evolution degree,an adaptive particle swarm optimization algorithm based on weight set is constructed.Finally,the proposed algorithm is applied to a typical test set,numerical results show that the algorithm is feasible and effective.In addition,based on the nonparametric hypothesis testing method,the statistical results show that the proposed algorithm is superior to other algorithms.3.Inertia weight particle swarm optimization algorithm based on group success rate is proposed.Firstly,a new inertia weight is defined based on group success rate in this algorithm.Then in order to overcome the premature phenomenon of particle swarm optimization,a new particle update strategy with global search is proposed.According to the motion of the particles,based on Metropolis criterion for simulated annealing,different particle update strategies are adopted.Finally,the proposed algorithm is applied to a typical test set,numerical results and statistical analysis show that the proposed algorithm is superior to other algorithms.