Analysis and evaluation of deterministic approximate models for the stochastic periodic inventory routing problem

Inventory Routing Problem (IRP) combines the two sub problems -inventory management and transport management- into one integrated optimization model, with the objective of reducing costs and assuring the promised service to the retailer. Every retailer has a limited capacity inventory location to store stocks so they are able to satisfy the demand until the next replenishment. The supplier decides the delivery time, the size of delivery and the transportation method while also minimizing the total costs of transportation, backlog and inventory holding costs. The parameters involved in this problem, retailer's demand, delivery time, inventory capacity, vehicle's capacity, etc. are stochastic in nature. It is not easy to predict their exact values when dealing with real life cases, due to various circumstances such as transportation conditions, weather, market fluctuations, etc. Taking the stochasticity of some of these parameters into account, results in more realistic outcomes. The model becomes however more complex. In this dissertation we develop, analyse, and evaluate the Inventory Routing Problem. A periodic version of the Inventory Routing Problem with stochastic demand (SPIRP) is considered in this dissertation. We reformulate the problem as an approximate deterministic model. Following this, different solution methods are presented and compared to obtain optimal results for each case