Resource planning for just-in-time make-to-order environments: A scalable methodology using tabu search

This paper develops a two-phase tabu search-based methodology for detailed resource planning in make-to-order production systems with multiple resources, unique routings, and varying job due dates. In the first phase rather than attempting to construct a good feasible plan from scratch, we define a novel approach to resource planning that computes an infeasible but optimal plan, uses it as the initial resource plan, and then makes the necessary modifications to the times of individual tasks to create a feasible finite-capacity plan. In the second phase we search for alternate finite-capacity plans that have decreased earliness, tardiness and lead time. To reduce earliness as well as tardiness, just-in-time philosophical elements are weaved into the construction of the initial solution, the neighborhood structure and the selection criteria. Computational experiments reveal that the tabu search-based methodology is more effective and reliable for resource planning than an exact approach using binary integer linear programming, which struggles to find a good solution in a reasonable amount of time even for trivially small instances. It also outperforms heuristic methods commonly used in practice for resource planning that sort jobs according to priority and load them onto resources one at a time