Learning-Based Grey Wolf Optimizer for Job Shop Scheduling Problem

This study introduces a novel integration of a discretized Grey Wolf Optimizer (DGWO) with Q-Learning (QL), marking a significant step in optimization techniques to solve Job Shop Scheduling Problem (JSSP). This innovative approach not only tailors the GWO for discrete optimization challenges but also utilizes QL for dynamic parameter tuning, a first in the field of JSSP. The synergy between these two methodologies advances solution quality and computational efficiency to new heights. Rigorous benchmarking against leading algorithms demonstrates the DGWO-QL’s superior performance, achieving the best-known solutions in 32 out of 34 cases. This breakthrough underscores the method’s robustness and its promising applicability to real-world scheduling complexities.