| Production scheduling is one of the most studied areas in the applied Operation Research literature. Aggregate level production scheduling has also found a large audience and been extensively researched because of its essential importance on the management of the manufacturing plans. In this study, the main concern is how the capacity management decisions or, equivalently, facility sharing decisions among the products are made, interacting with scheduling and lot sizing decisions. Inspiration for this concern stems from the gap between MRP and capacity requirements planning.; The increasing popularity of Just-In-Time systems and recent developments in manufacturing technology have led many production planning practitioners to ignore the setup cost and time parameters in scheduling problems. They have focused instead on the details of the capacity management in multi-level inventory systems. With a similar motivation, a solution to the capacitated multi-level production scheduling problem is attempted in this study. A hierarchical order between the facilities is established by the assumptions concerning interaction between facilities and products. A Dual Directed Algorithm to Modify Production Schedule (DDAMPS) taking advantage of the Leontief substitution system inherent in the model is introduced. The preliminary test results show that DDAMPS is very effective in terms of the quality of the solution and its solution time, especially when the lead times are ignored. An analysis for over 600 test problems is presented. Inspired by Leontief matrix properties, we show how alternative process plans may be added to the production scheduling model as substitute activities.; Even though setup times and related costs ideally should be eliminated, in many manufacturing systems these are still important factors to consider. For these cases, several decomposition approaches are considered. It is concluded that a possible algorithm taking advantage of both the primal and dual structures of the problem, like Cross decomposition, might be very beneficial to solve the model since both structures have some special features which make their solutions easier. |
