Improved Particle Swarm Optimization Algorithm For Complex Optimization Problems

Complex optimization is an important problem in engineering application and scientific research.With the increasing complexity of practical engineering application and the increasing requirement of decision makers for the fidelity of simulation models,simulation experimental models become more and more complex,larger and larger,and the computational cost is more expensive.Therefore,efficient algorithms for solving complex optimization problems are urgently needed.Therefore,this paper proposes two improved particle swarm optimization algorithms for complex optimization problems.The main work is as follows:1 ? First,based on decomposition strategy,an improved particle swarm optimization algorithm based on decomposition is proposed to solve complex large-scale multi-objective optimization problems.Firstly,decomposition strategy and social learning are introduced into individual learning process.Secondly,for each individual and its neighboring individuals,the distance between the direction of the weight vector and the reference point and the distance to the weight vector are calculated,and they are sorted.Individuals realize granules by learning any individual near the reference point and all individuals close to the weight vector.Then an improved particle swarm optimization(PSO)algorithm is proposed to solve complex large-scalemulti-objective optimization problems.Finally,a 500-dimensional and1000-dimensional test comparison is made on five ZDT test functions.The results show that the improved PSO has better convergence and uniformity of distribution.2?Based on the strategy of setting ideal points,an improved particle swarm optimization algorithm is proposed to solve complex many objective optimization problems.Firstly,on the basis of decomposition strategy,the ideal point on the weight is found.Secondly,for each individual and its neighboring particles,the distance between the ideal point on the weight vector and the direction of the weight vector to the ideal point is calculated,and they are sorted.Individuals learn from the individual near the ideal point on the corresponding weight vector to update the position.Then,the position is proposed.An improved particle swarm optimization algorithm is used to solve complex many objective optimization problems.Finally,the DTLZ test function is used to test and compare 3,5,8,10,15 and 20 targets.The results show that the improved algorithm is effective in solving many objective optimization problems and has better optimization performance than other algorithms.