A dynamic multi-objective optimization evolutionary algorithm based on classification of environmental change intensity and collaborative prediction strategy

The dynamic multi-objective optimization evolutionary algorithm (DMOEA) has garnered widespread attention due to its superiority in solving dynamic multi-objective optimization problems (DMOPs). Existing DMOEAs do not judge the intensity of environmental changes after they have been detected, which may lead to incorrect evolutionary directions of the population. To address this issue, this study proposes a DMOEA based on the classification of environmental change intensity and collaborative prediction strategy. Firstly, the algorithm optimizes the static optimization process, thereby determining the relative position of individuals in the objective space and enhancing the accuracy of environmental change detection. Upon detecting an environmental change, the algorithm proposes a method based on mutual information to further classify the intensity of the environmental change, and guides the particle swarm to adopt different velocity update methods for evolution based on the classification results. Secondly, a collaborative prediction strategy is proposed to ensure that the predicted population is closer to the Pareto optimal solution Set (PS) in the new environment. Lastly, a dual individual screening strategy is employed to select superior individuals from both the predicted population and the population before the environmental change to form the initial population in the new environment. Comparative experiments with advanced DMOEAs on 20 different types of test functions demonstrate the superiority of the proposed algorithm in solving complex DMOPs.