| A company in the building products industry decides not to build a finished goods warehouse and, instead, manufactures and ships all products to order. The production and shipping schedules are identical and transportation costs are high. Constraints include order due dates, and production and vehicle capacities. Decisions include vehicle loading and routing, carrier choice, and production date. The objective of the problem is to minimize the combined transportation cost, lateness penalty and any overtime production cost over a rolling horizon.; Two different heuristic approaches, the Route-First-Schedule-Second (RFSS) approach and the Schedule-First-Route-Second (SFRS) approach, are presented to solve the combined production and transportation scheduling problem. The RFSS approach first assigns all the available orders to different vehicle routes, then solves an optimization problem to schedule the production over the planning period. The SFRS approach takes advantage of historical data and produces seed routes in a pre-processing step. The results from the pre-processing step, along with other order information, are used in by optimization problem to select a subset of orders for the current day production. Then a variant of vehicle routing problem is solved for the current production day by using either a recombinant route-first cluster-second algorithm or a generalized assignment approach. Two alternative methods that use centroids to estimate insertion cost are also discussed.; Results from simulation show that substantial improvements over the human expert's solution are possible. Sensitivity analysis of the SFRS algorithm with respect to production unitilization and lateness penalty is also presented as part of the computational results. |
