MASTER SCHEDULING AND COMPONENT LOT SIZING DECISIONS IN RESOURCE-CONSTRAINED MATERIAL REQUIREMENTS PLANNING ENVIRONMENTS

The study formulates and investigates an integrated model for making managerial decisions on master production scheduling, component lot sizing, and resources and capacity planning in material requirements planning (MRP) environments. MRP, due to its dependent demand concept and the ability to organize large amount of data in a manufacturing system, has been getting wide acceptance in industry. However, due to their recent origins, MRP systems are used largely as sophisticated information systems. The crucial decisions on master schedule, lot sizes, and resources and capacity planning are still made based upon past experience, managerial judgment and trial-and-error approaches.;Test problems are generated by controlling ten experimental variables representing shop and cost characteristics. Different decision methods such as the capacity insensitive method, decoupled decision method, and the proposed heuristic procedures are compared. Optimal solutions are also obtained for smaller problems in the study. Cost structure variables, especially capacity change cost (overtime cost), setup time as percentage of total capacity and routing complexity, are the significant variables in differentiating the above decision methods. The decoupled decision method shows significant improvement over the capacity insensitive method, but it still falls short of the optimal solutions. On the other hand, the heuristic procedures do very well and give nearly optimal solutions.;This study formulates a unified model for the above decisions and suggests heuristic procedures for solving the model. The model is based upon MRP logic, but the MRP explosion process is carried out in terms of decision variables of the master schedule. This is made possible by making an assumption that assembly and subassembly production is done lot-for-lot. Major trade-off between setup cost and inventory carrying cost is required only at component level. This is a realistic assumption, as has been suggested by many practitioners. The above formulation results in a mathematical model for the decision environments. The heuristic procedure proposed iterates between master schedule decisions and lot sizing decisions until no more improvement in the objective function is possible. Master schedule decision model is a linear program, whereas the component lot sizing model is solved by the column generation technique.