Swarm Intelligence Research And Its Application On Constraint Optimization

Nature is magic.Creatures adapt themselves to changing environments through continuous evolution,it's fascinating.A series of intelligent optimization algorithms based on multi-population are proposed by computer workers,which provide efficient,simple and robust solutions for various optimization problems.Particle Swarm Optimization is an important branch of Swarm intelligence.Its simple concept,easy implementation and fast convergence speed make it more and more popular and widely used in science and engineering.However,particle swarm optimization algorithm has some limitations when dealing with problems,such as:the algorithm has premature convergence and is easy to fall into local extremum.Constrained optimization is a widespread problem in the fields of economy and engineering.When solving a problem,the constraint conditions make the solving process difficult.Evolutionary algorithm has the characteristics of wide applicability and strong robustness,so this paper tries to apply the difference algorithm to solve constrained optimization problems.The main research work of this paper is as follows:(1)In all PSO variants,the relative position relation between the global best position and the personal best position has important influences on the performance of algorithms.In order to make full use of their relative positions,the relative positions of the two difference vectors are determined by the inner product of vectors,and different particle velocity updating strategies are adopted accordingly.The improved particle swarm optimization algorithm is divided into two aspects:vector level and component level.The effectiveness of the inner product based particle swarm optimization is verified by simulation and comparison test.(2)In order to solve the constrained optimization problem,the second order differential evolution(SODE)is applied.Then,the feasibility is verified by simulation experiment,that is,the algorithm formed by the combination of second order differential evolution and ?constrained method can effectively solve the constraint optimization problem.(3)In addition to the study of the classical DE/rand/1 and DE/best/1 evolution of the second order differential evolution algorithm,this paper also studies the theoretical analysis of DE/rand-to-best/1,DE/current-to-best/1,and DE/best/2 after the difference strategy changes into the second order difference.Two difference strategies,DE/current-to-best/1 and DE/best/2,were selected for the simulation experiments to verify the effectiveness of the second order difference idea.