----------------------------------------------------------------------------------------------------

        Department of Systems Engineering and Engineering Management

                            The Chinese University of Hong Kong

----------------------------------------------------------------------------------------------------

Date: 10:00 - 11:30 on 28 August (Thursday)

Venue: ERB 513, The Chinese University of Hong Kong

Title: Multi-product Newsvendor with Indifferent Demand

Speaker: Gar Goei Loke, Department of Management and Marketing, Business School, Durham University

Abstract: In this work we study the multi-product newsvendor model with

indifferent demand. We define indifferent demand as the situation where

customers demand a set of products and are satisfied by a single product

from this set. Such form of demand flexibility can occur in shared

products or service allocation contexts. We model the demand as a

2^N-dimensional random variable over the space of submodular set

functions on N products. We exactly characterize the optimal solution.

Specifically optimal inventory levels should be decided greedily in

decreasing order of price and cost only affects the specific inventory

levels. Furthermore these inventory levels are the inverse distribution

functions of the critical quantile as in the traditional newsvendor

problem however under a distribution that adjusts for potential

shortfall as a result of subsequent products and aggregating over the

different combination of product sets that contain the relevant product.

Our work opens the door to stochastic optimization over submodular

random variables.

Biography: Dr. Gar Goei Loke is an Associate Professor in the Department of

Management and Marketing, Business School at Durham University. His

research focuses on decision-making under uncertainty, and developing

models, frameworks, methods and algorithms that help decision-makers go

from data to decisions. In his earlier stream of research, he has

applied techniques in robust optimization to the solution of

optimization problems in queueing networks. More recently, he is

developing and proposing new ways to integrate machine learning and

optimization harmoniously. His research is primarily applied to business

areas such as service operations management, supply chain management,

healthcare operations management, and energy and water. His research has

been published in journals such as Operations Research and Manufacturing

& Service Operations Management.