Return to search

Managing Stochastic Uncertainty in Dynamic Marketplaces

Firms' operations management decisions are often complicated by various types of uncertainties, ranging from micro level customer behavior to macro level economic conditions. Operating in the presence of uncertainties and volatilities is a challenging task, one that requires careful mathematical analysis and tailored treatment based on the uncertainty's characteristics. In this thesis we provide three distinct studies on managing stochastic uncertainty in dynamic marketplaces. The first study considers agents' dynamic interactions in a large matching market. A pair needs to inspect for their compatibility in order to form a match. We study a type of market failure called 'information deadlock' that may arise when pairs are only willing to inspect their most preferred prevailing partner. Under information deadlock, a large fraction of agents wait in the market for long (if not forever) in spite of there being opportunities remaining in their consideration sets. Using advanced tools in statistical physics and random graph theory, we derive how the size of the deadlock is affected by the market's primitives. We also show that information deadlock is prevalent in a wide range of markets.

Our second study tackles a service firm's problem of choosing between a safe service mode and a risky service mode when serving a customer who might probabilistically churn. One key behavioral feature of the customer that we consider is named recency bias --- his happiness with the firm (that crucially determines his churn risk at the time) depends more heavily on his more recent experience. We show, by solving a stochastic control problem, that the firm should be risk-averse when the customer is marginally satisfied and risk-seeking when the customer is marginally unsatisfied. The optimal sandwich policy can significantly outperform the naive myopic policy in terms of customer lifetime value. Our third study deals with a dual sourcing problem under fluctuating economic conditions. We model this via an underlying Markov modulated state-of-the-world which affects the two suppliers’ cost structures, capacity limits and demands. We develop two approaches to show how the optimal combined ordering strategy from the two suppliers, along with a salvaging policy, can be efficiently computed, and characterize the relatively simple structure of the optimal policies. Interestingly, we find that the firm can, by exploiting the dual sourcing options, benefit from increased environmental volatilities that affect the suppliers’ cost structures or capacity limits.

Identiferoai:union.ndltd.org:columbia.edu/oai:academiccommons.columbia.edu:10.7916/d8-7kwh-tx47
Date January 2021
CreatorsLu, Jiaqi
Source SetsColumbia University
LanguageEnglish
Detected LanguageEnglish
TypeTheses

Page generated in 0.0076 seconds