Joint Location-Two-Echelon-Inventory Supply chain Model with Stochastic Demand and Environmental Considerations

  • Malek Rasmi Abu Alhaj

Student thesis: Master's Thesis


Supply chain management studies the flow of materials/products from suppliers to customers in order to minimize costs and increase profitability. These channels interconnect manufacturers, distributors, retailers, and/or customers, and allow for the transport of raw materials, products, and in some cases recyclable waste. The design of a supply chain is affected by a number of factors, including product lifetime, geography/location, storage feasibility, legal frameworks, and environmental considerations. Given the increasing importance of environmental considerations, it is vital that we consider approaches to incentivizing emissions reduction in supply chains. This thesis consists of three parts. In the first part, we investigate a joint location-inventory supply chain problem. The problem consists of one plant, multiple distributors, and multiple retailers. Products flow from a plant to distribution centers (DCs), and from DCs to retailers. In order to minimize costs, the number of DCs, the assignment of retailers to DCs, and the amount of holding inventory are carefully chosen. We present an extension to the deterministic multi-echelon joint location-inventory model that adds uncertainty by including a new variable that reflects the probability of different demand scenarios. Different demand scenarios are considered to understand the tradeoffs between demand fluctuations and risk. In the second part of the thesis, we extend our model to take into account carbon emissions and environmental considerations. This model is similar to the previous model, but includes an extra term that reflects the cost of carbon emissions associated with operating distribution centers and shipping. Due to the complexity of the models we develop, the third part of this thesis is concerned with developing efficient solution algorithms that can solve such difficult problems in reasonable computational times. We develop a Genetic Algorithm (GA) to solve these problems, and test it using several chromosome representations and different probabilities for mutation and crossover, in addition to different evaluation functions. We validate the accuracy of our GA on small problems that have been solved to optimality using GAMS.
Date of Award2012
Original languageAmerican English
SupervisorAli Diabat (Supervisor)


  • Stochastic analysis

Cite this