The joint optimization of a replacement policy and inventory control policies using response surface methodology

This thesis considers the joint optimization of a replacement policy and inventory control policies. For the replacement policy the modified block replacement policy (MBRP) is used. As inventory policy the (s,S) policy is used. Joint ordering policies and separate ordering policies for both the orders for failure replacement and orders for block replacement are compared. Two new ordering policies are proposed and compared with two existing policies. Response surface methodology (RSM) is used with experimental design in order to determine the response surface function, which gives the relationship between the average costs and the input parameters. Using RSM for finding the optimal parameter settings led to some difficulties. The common used fractional factorial design did not give a model with high explanatory power and one of the main parameters did not have a significant influence on the costs. The results show that the proposed policies are an improvement, since they are more practical to implement than the existing policies and perform the same or only slightly poorer.