A comparative study of heuristic algorithms to solve maintenance scheduling problem

Purpose - The purpose of this paper is to compare the effectiveness of two meta-heuristics in solving the problem of scheduling maintenance operations and jobs processing on a single machine. Design/methodology/approach - The two meta-heuristic algorithms, tabu search and simulated annealing are hybridized using the properties of an optimal schedule identified in the existing literature to the problem. A lower bound is also suggested utilizing these properties. Finding - In a numerical experimentation with large size problems, the best-known heuristic algorithm to the problem is compared with the tabu search and simulated annealing algorithms. The study shows that the meta-heuristic algorithms outperform the heuristic algorithm. In addition, the developed meta-heuristics tend to be more robust against the problem-related parameters than the existing algorithm. Research limitations/implications - A future work may consider the possibility of machine failure along with the preventive maintenance. This relaxes the assumption that the machine cannot fail but it is rather maintained preventively. The multi-criteria scheduling can also be considered as an avenue of future work. The problem can also be considered with stochastic parameters such that the processing times of the jobs and the maintenance related parameters are random and follow a known probability distribution function. Practical implications - The usefulness of meta-heuristic algorithms is demonstrated for solving a large scale NP-hard combinatorial optimization problem. The paper also shows that the utilization of the directed search methods such as hybridization could substantially improve the performance of a meta-heuristic. Originality/value - This research highlights the impact of utilizing the directed search methods to cause hybridization in meta-heuristic and the resulting improvement in their performance for large-scale optimization.