Abstract
In order to effectively solve the dynamic multi-objective optimization problem, a new dynamic multi-objective optimization algorithm based on prediction strategy is provided in this paper. The algorithm detects changes in the environment by recalculating individuals. At the same time, the prediction model is established based on individuals of the first two generations, which is used to generate the new individual. In order to improve the diversity, Cauchy Mutation is used in the algorithm. Then, there are the non-dominated sort and tournament selection to deal the individual. The proposed algorithm is validated on several typical test functions. Meanwhile, the algorithm is compared to DNSGA-A. The experimental results show that the rate of convergence gets improved and the population solution is closer to the real solution. The algorithm do well in resolving the basic dynamic multi-objective function.