Abstract

Reacting to customer demand changes by frequently re-planning and re-scheduling the production plans is one of the managerial perspectives to improve customer satisfaction enabled through new information technologies such as Radio Frequency Identification (RFID). However, this may cause nervousness in supply chain operations, resulting in unexpected operational costs, and increased lead time due to congestion in the nodes of the supply chain. The objective of this thesis is to develop a decision support system to identify the lot sizing and batching decisions by considering the congestion effects resulting from uncertainties in a supply chain environment. A Mixed Integer Linear Programming (MILP) model is developed to determine lot sizing decisions for a supply chain. The developed MILP model considers the effect of congestion at the supplier nodes using queuing models. In addition, instability metrics are proposed to measure the stability of supply chain lot sizing decisions. The output of the lot sizing decisions is tested with the proposed metrics in a simulation environment by considering various uncertainty levels. A sensitivity analysis is conducted in order to demonstrate the impact of batching and supplier capacity decisions under high, medium, and low demand variability. The results show that increasing the supplier capacity by a small increment has a significant improvement on the total cost. Moreover, considering the congestion effect into their MRP schedule increases the overall service level. The benefits of incorporating the congestion effects have also been demonstrated by the proposed stability metric.