Research And Application On Dynamic Multi-objective Particle Swarm Optimization Based On Diversity Strategy

Dynamic multi-objective optimization problem widely exists reality. Research on the optimization algorithm to solve the problem is of great significance to promote the development of this area. And at present, the particle swarm optimization algorithm is widely used in the static optimization problem, because of its high accuracy, fast convergence. But when it was extended to dynamic multi-objective problem solving,on the one hand, the fast convergence of particle swarm optimization can meet urgent requirement of convergence speed in solving dynamic multi-objective optimization problems; On the other hand, the conflicts between objectives, the variability of the problem, and the premature convergence of particle swarm, force algorithm improve the diversity of solution set and environmental adaptation ability, to improve the abilityof describing and tracking dynamic optimal solution set. All above bring new challenges to the design of the dynamic multi-objective particle swarm optimization algorithm.Based on the above facts, in view of the requirements of solution set distribution and environment adaptability, with the designing of a novel diversity strategy based on population distribution information and an adaptive environment treatment strategy,a dynamic multi-objective particle swarm optimization algorithm based on the diversity strategy is proposed, and applied to the dynamic optimization problem solution in power system. Specific work is as follows:1. Study on a novel diversity strategy based on population distribution information. First of all, with the aid of statistical methods to classify evolutionary status; Secondly, diversity introduction strategy based on the population distribution is designed; Thirdly, according to the state of evolutionary design diversity maintaining strategy; Finally, a diversity strategy based on population distribution information is proposed. Experimental results indicate that the new diversity strategy is of significant advantages in avoiding local optimum and improve global diversity.2. Study on a novel dynamic multi-objective particle swarm optimization algorithm based on the new diversity strategy. In view of the dynamic change of optimal solution set in dynamic multi-objective problem. An environmental change monitoring strategy and an adaptive dynamic environment response strategy based on the indicator is designed. Finally, based on the new diversity strategy above, a dynamic multi-objective particle swarm optimization algorithm based on diversity strategy(dMOPSO) is proposed. The experimental results show the obvious performance advantage of dMOPSO in a statistical sense.3. Application of the new algorithm in the practical, dynamic electric power system optimization problem. In order to verify the effectiveness of the algorithm, the dynamic grid system optimization problem is solved in a simulation experiment.Experimental results indicate the effectiveness the proposed algorithm by comparing with other optimization algorithms.