Discrete variable optimization design algorithm based on improved marginal optimization

Aiming at the problems of the traditional discrete variable optimization method such as too many times of objective function calculation and poor convergence, a discrete variable optimization design algorithm based on improved marginal optimization learning from marginal optimization theory and pattern search algorithm is designed. Based on the principle of marginal utility optimization, the concept of unit step space is introduced to improve the selection of initial point, marginal increment design, tabu search strategy, and mutation operation is designed to jump out of local optimum. Case studies show that the proposed algorithm can quickly and accurately converge to the local optimal solution, and the satisfactory solution or optimal solution can be obtained with as few objective functions as possible, which is suitable for solving high-dimensional discrete variable optimization problems and simulation optimization problems. © 2021, Editorial Office of Systems Engineering and Electronics. All right reserved.