Niching Particle Swarm Optimization Algorithms For Multi-Modal And Dynamic Problems

In real life,many complex engineering problems have some characteristics in common,such as non-linear,large-scale,multi-modal and dynamic.Solving such problems by using traditional evolutionary algorithms have some limitations,so it is necessary to improve the traditional evolutionary algorithms to meet the needs of solving such complex problems.The particle swarm algorithm(PSO)is inspired by the observation of the behavior of the flocking animals in the biological world to find food,and the proposed computational model has features of simplicity,robust,fast search speed,and high precision.This paper describes in detail the origin,structural characteristics of the PSO algorithm,as well as its development and related applications.Multimodal optimization problems are very common in practical applications.This type of problem often contains several optimal solutions(global and local).In order to solve such problems,a large number of scholars have added niche technologies to improve on the basis of evolutionary algorithms.In the third chapter,the mathematic model of multimodal optimization is described in detail.Some classical niche technologies are introduced.The related research of PSO algorithm in multimodal optimization is described recently.The linear degressive inertia weight coefficient was introduced to improve the star topology structure and the ring topology PSO algorithm.The simulation experiments of the improved two PSO algorithms using 15 complex multi-modal test functions proved the advantages of the relevant algorithms..In practical applications,many problems to be optimized will change dynamically with time.Therefore,dynamic optimization problems are very important and very difficult research hotspots.Group intelligent evolutionary algorithm is a feasible solution to optimize this kind of problem.In this paper,several improvements are proposed based on the introduction of the PSO algorithm of the linear topological weight coefficient ring topology.Firstly,the dynamic detection example is added to detect whether the external environment changes,and secondly,once the environmental change is detected,the corresponding response mechanism is added.One response mechanism is a niche technology based on a clearing mechanism,and one is a niche technology based on an adaptive congestion degree.The simulation experiments of the improved algorithm before and after the improvement were carried out through relevant dynamic test functions.The experimental results show that the improved algorithm performs better than the improved one and has a certain degree of effectiveness.The PSO algorithm that incorporates the adaptive crowding degree response mechanism is among the three.The best performance.