An Improved Multi-Objective Hybrid Algorithm for Solving Job Shop Scheduling Problem

The Job Shop Scheduling Problem (JSSP) continues to represent a challenge for researchers from all over the world who are trying to find optimal solutions. The current trend is represented by combining several algorithms in hybrid methods that offer an increased quality of results. This paper proposes a hybrid artificial-based algorithm that aims to minimise the total work flow time, which makes the problem more difficult mathematically speaking but more interesting from a practical point of view. This algorithm exploits a multiple-refined random generation by using an adapted genetic algorithm that becomes the start generation of the hybrid algorithm. In addition, the proposed hybrid algorithm aims to combine particle swarm optimisation and simulated annealing advantages. A complex heuristic function is implemented in order to evaluate each individual solution as accurately as possible. A substantial experimental study across several classic benchmarks was performed in order to demonstrate the algorithm performance; the results are cross compared with other algorithms, and conclusions were drawn. The experimental study confirms the performance and effectiveness of the proposed algorithm, thus providing a significant contribution to the field of JSSP optimisation.