Improved Particle Swarm Optimization And Its Application In Job-shop Problems

The rise of artificial intelligence drives the development and progress of intelligent algorithms,and provides a new way to break the limitations of traditional optimization algorithms for solving optimal problems.Particle Swarm Optimization(PSO)provides a new and effective way to solve the optimization problem.The particle swarm optimization algorithm is simple in principle,less in parameters,easy to implement and fast in speed.However,the particle swarm optimization is easy to produce premature convergence,weak local optimization ability,and the basic theory is not complete.This paper introduces the derivation process of particle swarm optimization,emphasizing the principle and execution flow of particle swarm optimization,the control parameters of particles such as position velocity,learning factor and weight value,adding improved strategy to the algorithm,For example,extremum perturbation,Cauchy mutation,different-dimensional learning and so on.In order to verify the advantages of the improved algorithm,simulation experiments are carried out in the aspects of efficiency,accuracy and robustness of the improved algorithm,and it has achieved good results compared with some well-known algorithms.The main points of this paper are summarized as follows: Firstly,a simplified particle swarm optimization with nonlinear extremum perturbation and Cauchy mutation is proposed.The simplified particle swarm removes the velocity variable and the particle update formula changes from the second-order differential equation to the first-order differential equation,promoting the performance of the algorithm.By introducing the non-linear decreasing disturbance operator and Cauchy mutation strategy,the diversity of the population is enhanced,the particles are prevented from falling into the local optimum,and the searching accuracy of the algorithm is improved.Secondly,based on the particle swarm optimization and differential evolution hybrid algorithm,a hybrid particle swarm optimization algorithm based on the different-dimensional variation is proposed.According to the distribution of particles and the strategy of variation of different dimension,dimensionality factor is introduced to ensure that even if some particles fall into the local extreme,they can jump out of circulation in time to improve the population diversity and improve the accuracy and efficiency of the later algorithm.Thirdly,the improved algorithm is used to solve the job shop scheduling problem.The mapping between the particle and the scheduling problem is completed by code/decode operation,and the simulation is performed on the classical dispatch library instance.