The problem of when to build a lot of component parts, and how many to include in the lot so as to minimize overall production costs, is a fundamental problem facing all repetitive manufacturing environments. Recently this problem has been addressed through the use of Material Requirements Planning (MRP). However, MRP is not in itself a complete answer, since it does not provide a built in method to determine lot sizes and thus optimize the total cost;A number of authors have studied the problem of developing lot sizes in MRP systems, however a consensus has been reached on very little. One reason that few of the results agree is that they lack a basic examination of the actual problem structure. By examining this structure, several currently unrecognized or misunderstood features can be developed. This dissertation begins by developing these features and studying how they impact upon the lot sizing solutions;These features allow a relatively simple, yet robust, new method of determining lot sizes to be developed. This new method is called the Integer Lot Sizing (ILS) heuristic. While, in some ways, the ILS heuristic is similar to previously explored lot sizing methods, in many aspects the ILS heuristic is unique. By utilizing the newly recognized features of MRP systems, the ILS heuristic develops costs which approach the optimum while utilizing an acceptable amount of computational resources.