Loading...

A Markov Decision Processes Model for Supplier Selection in Supply-Chain Management Under Uncertainty

Milani, Hamed | 2021

171 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 54710 (01)
  4. University: Sharif University of Technology
  5. Department: Industrial Engineering
  6. Advisor(s): Haji, Alireza
  7. Abstract:
  8. There are several types of uncertainty in supply chains that cause some difficulties in the planning of a supply chain. Uncertainty on the supply side can cause interruptions in the quality, quantity, or delivery time of orders. To avoid such interruptions, retailers usually contract with two different suppliers for a common good. In this setting, one of the suppliers has higher reliability and higher price, while the other supplier offers a lower price but has a lower level of reliability. In other words, to ensure a lower level of uncertainty, retailers have to pay more. This is one of the main tradeoffs in the dual sourcing strategy. In this thesis, we consider a two-echelon supply chain in which there is a retailer with a dual sourcing strategy that has a central distribution center and several branches in the first layer, and in the second layer, there are two different suppliers with different characteristics. The first supplier is called reliable and the second one is called unreliable, with a high rate of reliability and price and a low rate of reliability and price respectively. Our goal is to decide about the portion or quantity of our orders from each supplier. For this aim, we consider a stochastic inventory control problem with a single product, periodic review, and discounted infinite planning horizon. In this problem, the product demand is random and stationary, and supply is uncertain. This problem is modeled as a Markov Decision Processes with discrete-time. We introduce a new Markov Chain model with three states to control the unreliable supplier’s more complicated behaviors. In our model, other than one and zero supply, we consider another situation in which unreliable suppliers deliver only a portion of the order to the retailer. The optimal solution of this MDP has obtained by the Value Iteration algorithm. In the numerical result, we analyze the impact of reliability level on the optimal solution, and also the impact of cost parameters on total cost improvement. We observe that when reliability level of unreliable supplier is higher, decreasing unit cost of reliable supplier has better impact on total cost improvement
  9. Keywords:
  10. Supply Chain Management (SCM) ; Supplier Selection ; Stochastic Process ; Dynamic Programming ; Markov Decision Making ; Uncertainty

 Digital Object List

 Bookmark

No TOC