A Comparative Study of Hybrid Metaheuristics for the Flexible Job Shop Scheduling Problem: From SA-GA Operator Analysis to Multi-Algorithm Evaluation

This paper presents a two-phase comparative study of hybrid metaheuristic algorithms for solving the Flexible Job Shop Scheduling Problem (FJSSP). In Phase 1, we analyze the hybridization of Simulated Annealing (SA) with Genetic Algorithm (GA) operators — examining parameter sensitivity and the effect of selection, crossover, and mutation operator design on solution quality. In Phase 2, we broaden the comparison to include Ant Colony Optimization (ACO), Memetic Algorithms (MA), Harmony Search (HS), and Variable Neighborhood Search (VNS), evaluating hybrid combinations across eight instances of the Brandimarte benchmark dataset (MK01–MK08). Results identify which hybrid pairings consistently achieve competitive makespan minimization across problem instances of varying scale and complexity, and provide practical insights into algorithm selection for production scheduling applications.