A Local Pareto Front Estimation Framework for Multi-Objective Optimization

Advanced manufacturing processes frequently require scheduling, optimization, and decision-making, areas where evolutionary multi-objective optimization (EMO) has demonstrated strong capabilities. The Pareto front estimation has gained significant attention due to its remarkable performance on EMO. Recent studies have proposed few Pareto front (PF) estimation strategies, which could guide the search direction according to the geometric characteristics of the approximated Pareto front. However, when handling the complicated MOPs with irregular Pareto fronts, these strategies encounter significant challenges. This paper proposes a local Pareto front shape estimation approach based on Minimum Manhattan distance, which estimates the geometric information at a fine-grained level. Then an adaptive fitness function based geometric information is used to guide the search direction. Experiments on benchmark problems show that the proposed algorithm is highly competitive compared with state-of-the-art multiobjective evolutionary algorithms.