Aggregate planning in crew-loaded production environments

Virtually all existing optimization techniques for the aggregate planning problem incorporate continuous decision variables. However, predetermined shift settings in crew-loaded production operations often require the decision variables to be discrete values. Consequently, currently available optimization methods are not appropriate for discrete production environments. This paper describes an optimal mixed-integer programming formulation for the aggregate planning problem that explicitly specifies discrete variables, thus resulting in feasible workforce levels for crew-loaded production environments. The mixed integer model is compared to the LDR [3, Holt, the Production Switching Heuristic [5, Mellichamp and Love, Mgmt Sci. 24,1242–1250 (1978)] and then to the PSR [4, Oliff and Leong, Decision Sci. 18, 582–597 (1987)]. An application at Owens-Coming Fiberglas is also presented.