Non-Dominated Sorting Genetic Algorithm Based on Auxiliary Archive Strategy and Difference Learning Operator

Current multi-objective evolutionary algorithm based on domination still encounters several challenges, including bad population convergence, bad population diversity, as well as relatively slow algorithm convergence rate. Therefore, this paper proposes a non-dominated sorting genetic algorithm based on auxiliary archive strategy and difference learning operator (NSGA/AD). Firstly, a difference learning operator based on fuzzy logic (DL/FL) is proposed. This operator can use the features of high-quality solutions to improve the population convergence and accelerate the algorithm convergence rate. Then, a diversity maintenance scheme based on weight vectors (DM/WV) is proposed. This scheme can use weight vectors to improve the population diversity. Then, an auxiliary archive strategy based on losers (AA/L) is proposed. This strategy can avoid the loss of some high-quality solutions and improve the population convergence. Finally, algorithm framework of NSGA/AD algorithm is given. Simulation results show that compared with other advanced algorithms, NSGA/AD algorithm is more excellent and has better competitiveness.