A quadratic association vector and dynamic guided operator search algorithm for large-scale sparse multi-objective optimization problem

Qinghua Gu,Yixiao Sun,Qian Wang,Lu Chen
DOI: https://doi.org/10.1007/s10489-023-04500-z
IF: 5.3
2023-03-08
Applied Intelligence
Abstract:There are many large-scale sparse multi-objective optimization problems in real life, but it is difficult to solve such problems due to the high dimension of decision variables and the excessive sparseness of Pareto optimal solutions. To address these issues, this paper proposes a quadratic association vector and dynamic guided operator search algorithm for large-scale sparse multi-objective optimization problems. Firstly, according to the individual distribution in the decision space, the reference vector is used to quadratically associate the individual, aiming to find better solutions during the initialization process. Secondly, the dynamic guided operator is introduced into the crossover mutation. The population is guided to evolve toward the sparse optimal solution by flipping the real and binary variables according to the dynamic guided operator. Thirdly, individuals are selected into the mating pool for crossover mutation based on the decision variables of the population individuals, with the aim of further enhancing the convergence of the algorithm. The experimental results on eight benchmark problems show that the algorithm obtains the best comprehensive performance on 66.7% of the test problems. The proposed algorithm is superior to the existing algorithm in terms of effectiveness.
computer science, artificial intelligence
What problem does this paper attempt to address?