A Novel Hybrid Method for DAP: Differential Evolution with Variable Neighborhood Search

Abstract

This research investigates MOPFSP-SDST, an advanced and highly computational scheduling difficulty in real-world manufacturing systems. It examines how it correlates with multi-objective permutation flow shops. LS-MOVNS stands for "Learning and Swarm-based Multi-objective Variable neighbourhood Search." It is a better metaheuristic method that combines evolutionary swarm search and adaptive local search techniques to address this Problem. The two main improvements have been discussed: a partial neighbourhood assessment framework that reduces the computational expenses by analysing only a particular portion of the neighbourhood, and an adaptable neighbourhood series selection procedure that rapidly chooses the most beneficial neighbourhood order depending on past performance rates. These improvements aim to make searches more effective and productive by finding a better balance between exploration and exploitation. Particularly in medium to large problem sizes, experimental tests in benchmark instances show that LS-MOVNS frequently outperforms current modern algorithms in convergence and diversity. The results verify the long-term reliability, scalability, and practical applicability of LS-MOVNS for resolving challenging multi-objective scheduling issues.