This thesis deals with the problem of determining an optimal number of spare parts of a weapon system required for a given period of time under the procurement budget constraint.
Two kinds of the objective functions are adopted to develop models, one with minimizing the weighted number of stockouts and the other with maximizing the operational availability.
It is shown that the approach with Lagrange Multiplier or Dynamic Programming can generate an optimal solution of each model with prohibitive computational burden when the number of spare parts under consideration is large.
Recognizing these shortcomings, a heuristic procedure is developed based on Marginal analysis. These three solution procedures are programmed for microcomputer and compared through an example problem.