Performance Evaluation of Various Heuristic Algorithms to Solve Job Shop Scheduling Problem (JSSP)

Abstract

Scheduling is a famous optimization problem that seeks the best strategy of allocating resources over time to perform jobs/tasks satisfying specific criteria. It exists everywhere in everyday life, particularly in manufacturing or industrial applications. An essential class of scheduling problems is a job shop scheduling problem (JSSP), an NP-hard optimization problem. Several researchers have reported the use of heuristic methods to solve JSSP. This paper aims to investigate the performance of various heuristic algorithms to solve JSSP. Firstly, we developed a Genetic Algorithm (GA and compared the performance of some heuristic algorithms, including Particle Swarm Optimization (PSO), Upper-level algorithm (UPLA), Differential-based Harmony Search (DHS), Grey Wolf Optimization (GWO), Ant Colony Optimization (ACO), Bacterial Foraging Optimization (BFO), Parallel Bat Optimization (PBA), and Tabu Search (TS). The experimental results of the 28 benchmark test problems validated that the algorithms, except ACO, can provide the optimal solution of JSSP. PBA delivers the most impressive performance that solves 26 cases optimally, with the average error equal to 0.05%. Among those 28 test problems, TS, DHS, and PBA can solve 26 instances optimally, followed by GA that solves 21 cases.