Abstract
Due to the exclusion of capacity constraint considerations, it is difficult for a traditional MRP calculation to obtain a feasible production plan. Also, the separation of Lot Sizing Decision and Capacity Requirement Planning calculation makes the setup decisions difficult. In practical application, a production plan should includes multiple demand classes such as order boards and forecasts. How to allocate finite resources to meet the demand of different classes is an important issue. This thesis propose an integrated production planning model that not only considers multiple demand classes but also has the ability to deal with the multi-level capacitated lot sizing problem, which involves setup times, setup costs and lead times. Moreover, it can be applied to the problem with general product structures. In fact, this model integrates four modules known as Master Production Scheduling, Material Requirement Planning, Capacity Requirement Planning and Lot Sizing Decision into a single model. Under multiple demand classes and lot sizing decisions, each demand class problem is a Mixed Integer Programming problem with different priority. By sequentially solving each MIP problem according to its priority, we allocate finite manufacturing resources and generate a feasible production plan. This thesis uses three modern searching algorithms, which are tabu search, adaptive simulated annealing (ASA) and adaptive genetic algorithm (AGA), to solve this problem. Experimental designs and statistical methods are used to evaluate and analyze the performance of these three algorithms. As a result, tabu search has the best performance, ASA ranks second, and AGA is the last.