In labor intensive cellular layouts, well-trained workers constitute the single most important resource. In this paper we formulate and solve a workforce planning problem for general manufacturing systems. More specifically, let J be a given set of jobs. Each job consists of a chain of operations, and each operation must be processed by a specified work center. The production system consists of m work centers. We want to plan production over a horizon of P production cycles, where P is a decision variable. The duration of each production cycle is the same for all cycles, i.e., the production system is paced. Associated with each operation is the number of workers required to execute that operation within a production cycle. We seek a feasible schedule that minimizes a linear function of workforce costs and costs associated with the length P of the production horizon. We show that the problem is polynomially solvable for production systems with only two work centers, and NP-complete for m ≥ 3. We develop heuristic algorithms to solve the problem for arbitrary number of work centers. The performance of our heuristics is tested by means of a lower bounding scheme based on composite relaxation. Extensive computational experiments show that the heuristics exhibit near optimal performance.
