Simulated Annealing and Tabu Search for Solving the Single Machine Scheduling Problem

Abstract

AbstractThis paper presents a comparative study of two metaheuristic optimization algorithms for solving the Total Weighted Tardiness problem in its single-machine mode, which represents the most frequent scheduling and sorting problems occurring in the industrial environment. The metaheuristics evaluated in this study were Tabu Search, and Simulated Annealing, because they have shown satisfactory results in this kind of problems. The performance of each algorithm was evaluated by means of the total tardiness and the execution time using instances of 40, 50, and 100 jobs extracted from the OR-Library. The Simulated Annealing algorithm was found to be the most efficient method, being the one that found the best solutions in comparison to Tabu Search, nonetheless, Tabu Search found the results in the shortest time. A difference approximated of twenty-four units between Simulated Annealing and Tabu search was found in the total tardiness value when a set of 125 instances were executed. On the other hand, Tabu Search required only the 13% of the time execution required by Simulated Annealing.