A Speedy Algorithm For Determining Personalized Discounts For Products Approaching Their End-of-Life Stage

Open Access
Inochanon, Keerati
Graduate Program:
Industrial Engineering
Master of Science
Document Type:
Master Thesis
Date of Defense:
Committee Members:
  • Dr Soundar Kumara, Thesis Advisor
  • Soundar Rajan Tirupatikumara, Thesis Advisor
  • decision support system
  • personalized discount
  • inventory policy
  • dynamic pricing
  • expert system
  • fuzzy inference rules
  • data mining
Revenue maximizing entities in a supply chains such as suppliers, distributors, and retailers are often concerned with how they would quickly move products out of their inventory while still making profit. In this thesis, we look at a scenario in which a supplier is launching a new product and thus has a need to quickly empty out the inventory for the existing product that will be replaced, under the constraint that the revenue from the sales must at least cover the cost of the items. We present a method to determine an optimal discount for each customer to increase the likelihood that the customer will accept the offer and maximize the revenue. In order to achieve this, past purchase behaviors of customers are examined. Data mining techniques such as sequential pattern mining and clustering are used to mine customer behavior data. The amount of discount to be given for each customer is based on a price model that we created in combination with fuzzy modeling. Based on past purchase behaviors and past discount responses, fuzzy inference rules are created and a knowledge base is formed to make discount decisions. An example of how the algorithm is used with RFM (Recency, Frequency, and Monetary) features are presented. We show that the use of fuzzy modeling provides good results while also maintains unambiguity through its linguistic rules.