Improved Multi-objective Particle Swarm Optimization And Its Application In Cold Rolling Schedule

With the development of science and technology,the optimization problems encountered by mankind have become more and more complicated,new optimization methods-intelligent optimization algorithms,have become an effective means to solve such problems,and has been in-depth research and widely used to replace the traditional optimization algorithms..As one of the emerging intelligent optimization algorithms,particle swarm optimization(PSO)has attracted some attention because of its fast convergence and simple structure.However,due to the fact that the involved theory is not fully mature,the performance of the algorithm has some shortcomings,especially in multi-objective optimization,the distribution of the solution is poor and easy to fall into the local optimum.To improving the performance of particle swarm algorithm to solve multi-objective optimization using the Pareto dominant ideas,adopt various strategies to improve particle swarm optimization mechanism.Firstly,the heterogeneous learning pattern of population particles is constructed by introducing the cross organization form with global search performance in the genetic algorithm so that the particles in the population can exchange information directly with the elite particles to avoid the singularity of particle learning behavior.Secondly,propose the optimal solution of the individual refinement of the particle and set the individual learning enhancement factor to strengthen the "self-cognition" behavior of the particle to meet the optimal condition of the individual to improve the local exploration ability of the particle group.Then,a kind of external file redundancy set mechanism with mutation property is introduced,its effect on the one hand plays an important role in avoiding the algorithm into the local optimal which is achieved by the interference of the population;On the other hand to increase species diversity which is achieved by the re-maintenance of external files based on mutation generate particles.Finally,the simulation results show that this multi-strategy improved particle swarm algorithm has a good performance in multi-objective optimization problem.In this paper,the improved particle swarm optimization algorithm is applied to the optimization of rolling schedule.By analyzing the mathematical model of rolling process,the multi-objective optimization function of rolling mill rolling schedule is constructed,and the particle swarm optimization algorithm is used to optimize it.The goal of Pareto is to provide some reference for decision makers to make the most reasonable decision.