Seminar: Price and Quantity Optimization in Multiple Retailer Systems Using Machine Learning by Tuğberk Tunçinan


Department of Industrial Engineering  


Price and Quantity Optimization in Multiple Retailer Systems Using Machine Learning

 Tuğberk Tunçinan



Multiple retailer systems usually have a main supplier and many geographically dispersed retailers each of which faces external random demand.  Since the demand in each retailer is not known exactly, pricing decisions and the number of products sent to retailers is important in optimizing the revenue and customer satisfaction.

The solution procedure we proposed consists of obtaining a demand model depending on price, retailer characteristics and other relevant attributes using machine learning techniques. Then, the demand predictions of this model are used as parameters in profit maximization problem to decide on the total amount of items to be supplied, the allocation of these items to retailers, and the price for each period. We introduce quadratically constrained program (MIQCP) equivalent of the NLP profit maximization model under certain conditions. The computational study on randomly generated test instances show that the proposed formulation improves the solution quality within the allocated time limit.


Tuğberk Tunçinan is a PhD candidate in the Department of Industrial Engineering at Boğaziçi University, Istanbul. He holds M.Sc degree in Industrial Engineering, and B.S. degree in Mechanical Engineering from Istanbul Technical University. His research focuses on linear and combinatorial optimization, developing mixed-integer programming approaches to machine learning problems and data mining. At Istanbul Technical University, he is currently employed as a research assistant. He also focuses on building high performance, interactive data-driven internet applications, using web technologies. He founded Fiyat Ritmi with the vision of providing full price and product transparency in e-commerce.


Date: Friday, May 28, 2021

Time: 15:00-16:00.

Online Seminar Link:

Meeting ID: 934 3051 2540

Passcode: 335899