Mathematical Model for Solving Petroleum Replenishment and Routing Problem
Hsu, Yan Cheng
MetadataShow full item record
The dissertation aims to relax two common assumptions often considered while planning the distribution of commodities: (1) the customers' decisions are completely decoupled from those of the transporter and (2) there are always feasible dispatching and routing strategies that satisfy all the given demands and time requirements. The first part presents an inventory model that seeks to identify the optimal order quantities and desired time windows considering several scenarios for customers. The total cost considered for the customers includes ordering, holding, shortage, and send-back cost. The time windows and order quantities are determined by calculating the expected total cost based on the values of time window, delivery lead time, and stock out time. The second part focuses on the design of the daily distribution and replenishment operations of petroleum products over a weekly planning horizon by taking into account the perspectives of both the transporter and its customers. Due to the variability of the customers’ demands and expected time windows based on the order quantities and time windows selected by the inventory model in the first part, an approach that considers the possibility of having late deliveries is proposed. Then, the optimization models are designed to balance the late deliveries among the customers in order to mitigate the overall customer dissatisfaction. The third part of the dissertation develops a methodological framework that integrates the customer behavior within the transporter’s decision-making process for optimizing the product distribution logistics. The blocked time windows method is proposed to alleviate the difficulties of making routing plans. A column generation solution approach is utilized to solve the integrated routing problem. The sensitive analysis and case study are conducted to demonstrate insights and impacts of the proposed method. The results of this study have the potential to tackle distribution problems of several other commodities.