Uniform experimental design method with constrained regions based on adaptive neighborhood search

For the problem of constrained or irregular regions in uniform experimental design with constrained regions, a uniform experimental design method with adaptive neighbourhood search based on heuristic information named as neighborhood search uniform design (NSUD) was proposed. In the NSUD, to improve the overall uniformity of the population, each individual adjusted the neighborhood search direction and search step size based on the current population distribution information. For the strong constraint problem and equation problem, the differential algorithm was applied to help find the search direction for feasible solutions, which could guarantee the feasibility of solutions during the neighborhood search process. In the experiment, eight test problems for experimental design were used to compare with other experimental design algorithms based on heuristic information. Experimental results show that the proposed method can achieve better results in terms of solution set quality and convergence speed. © 2022, Editorial Board of Journal of Huazhong University of Science and Technology. All right reserved.