Date of Award

Fall 2014

Degree Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Industrial Engineering

First Advisor

Abhijit Deshmukh

Committee Chair

Abhijit Deshmukh

Committee Member 1

J. George Shanthikumar

Committee Member 2

Andrew Lu Lui

Committee Member 3

Julian Romero

Abstract

A flexible system is defined as one that can change the entity's stance, capability or status reacting to a change of the entity's environment. Flexibility has gathered the attention of academic researchers and industry practitioners as an efficient approach to cope with today's volatile environment. As the environments become more unpredictable and volatile, it is imperative for a flexible system to respond quickly to a change in its circumstance. How much flexibility is embedded into the system also has a critical impact on the long-term effectiveness of the flexible system. Moreover, this research focuses on the strategic environment where a decision maker's behavior influences other decision makers' and vice versa. ^ The primary objectives of this dissertation are developing a concrete framework for designing a flexible system by considering the exercise delay as a measure of flexibility and investigating the rational behaviors of decision makers who operate flexible systems under strategic environments. The general approach employed to develop the theoretical models for this dissertation includes optimal control theory, non-linear optimization, stochastic differential equation and game theory. ^ The first part of this research studies the optimal decisions on a flexible system with exercise delay within stochastic environments by postulating two level decisions, operational level and design level decisions. The operational level problem is modeled as a delayed optimal stopping time problem, and this research provides a comprehensive profile of the optimal operational policies according to the parameters representing the market conditions and characteristics of the alternative and designed features of the flexible system. In addition, the profile elucidates the interdependence between the operational level decision and the design level decision separating the entire domain of the design problem into sub-regions. This research effort finds that the design problem is decomposable with well-behaved non-linear optimization problems, and provides illustrative examples to show the usefulness of the developed framework. ^ The second part of this research concentrates on strategic environments which force a decision maker to cope with both exogenous uncertainty and endogenous interactions among decision makers. As the strategic environment, a duopoly market share competition is postulated where the total market profit is regarded as the underlying uncertainty. The player retaining an exclusive patent is regarded as a player competing in the market with a flexible system that does not have exercise delay, and the other competitor is interpreted as a player operating a flexible system with exercise delay. The open loop and closed loop information structures are considered for each model. The results showed that the open loop equilibrium is unique dominant strategy equilibrium. An interesting implication of the open loop equilibria is that the profitability of the flexible option decides the role of its owner in the duopoly market competition. This research finds that the closed loop equilibrium has two distinctive forms. When the asymmetry of exercise delay is large, the closed loop equilibrium is identical to the open loop equilibrium. On the other hand, if the asymmetry provides only a small enough advantage to the player who has a flexible option without exercise delay, the rational behaviors of the players are complicated in the closed loop equilibrium. The first insight from the closed loop equilibrium with large asymmetry is that the closed loop information structure hastens the execution of flexible options, and it results in lower payoffs to both of the players. Second, the role of each player is determined not only by the characteristics of the flexible options but also by the value of stochastic factor. Third, even the player with a competitive disadvantage from the asymmetry has a positive chance to be the leader of the market. ^ This research contributes to the area within industrial engineering and operations research by improving the current theoretical achievement of flexibility. The accomplishments of this work provides insights to various domains those would benefit from enhanced flexibility in the decision making process.

Share

COinS