Application of Multi-objective Particle Swarm Optimization Based on Adaptive Dynamic Reorganization in Teaching Data Library

Dynamic Multi-objective optimization (MOO) problems are common in reality. The research on optimization algorithms for solving such problems has important research and practical significance to promote the development of this field. At present, Particle Swarm Optimization (PSO) is widely used in static optimization problems because of its advantages of high precision and fast convergence. However, when it is extended to solve dynamic multi-objective problems, on the one hand, the advantage of fast convergence of the algorithm meets the urgent need for convergence speed of dynamic MOO problems; On the other hand, the conflict between objectives, the variability of problems and the precocity of PSO force the algorithm to improve the population diversity and environmental adaptability, so as to improve the depiction and tracking ability of the dynamic change optimal solution set. However, the existing strategies focus on single performance, only analyze the target space, and lack the ability to quantify the degree of environmental change, all these bring new challenges to the design of dynamic multi-objective PSO (MOPSO). Based on the above facts, a MOPSO based on adaptive network and dynamic congestion distance is proposed.When the number of external population exceeds the population size, the objective function space can be evenly divided into grids with the same interval, the number of particles in each grid can be counted, and then the particle density can be estimated to limit the size of external files; Then the variance information of particles is introduced, and an algorithm based on dynamic crowding distance is designed, which avoids the problem of deteriorating the distribution of solutions by eliminating all individuals with small crowding distance at one time, and improves the diversity of solutions.