Sustainable and reliable design of large-scale complex logistics systems under competition and uncertainties

Wang, Xin

Content Files

WANG-DISSERTATION-2015.pdf

Download Options

Download All Files Download Counts…

Loading…

Download Files

Loading…

Download Counts (All Files)

Loading…

Edit File

Loading…

Permalink

https://hdl.handle.net/2142/88208 Copy

Description

Title

Sustainable and reliable design of large-scale complex logistics systems under competition and uncertainties

Author(s)

Wang, Xin

Issue Date

2015-07-17

Director of Research (if dissertation) or Advisor (if thesis)

Ouyang, Yanfeng

Doctoral Committee Chair(s)

Ouyang, Yanfeng

Committee Member(s)

• Cai, Ximing
• Lim, Michael K
• Work, Daniel B.

Department of Study

Civil & Environmental Eng

Discipline

Civil Engineering

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Date of Ingest

2015-09-29T20:50:09Z

Keyword(s)

• Sustainable Logistics System
• Reliable Facility Location
• Biofuel Supply Chain
• Land Use Competition
• Continuum Approximation

Abstract

Logistics systems generally involve multiple interacting stakeholders who endogenously make decisions based on their individual, sometimes conflicting, objectives. Meanwhile, many of such systems may be disrupted from time to time under extreme threats (e.g., natural or human-induced disasters). These endogenous and exogenous factors often adversely impact system performance and result in significant societal disutility. My dissertation research focuses on developing mathematical models for design and analysis of large-scale logistics systems, especially those under competition and uncertainties. It holistically captures interactions and joint impacts of various objectives in large-scale supply chains, including supply reliability (against disruptions), service competition (against competitors), as well as demand uncertainties.Built upon a general analysis framework, we seek applications and extensions to address concerns of the current renewable energy sector. Starting from a logistics angle, i.e., the biofuel supply chain design, we investigate its profound economic and societal impact. First, We develop game-theoretical models based on Continuum Approximation (CA) to study a reliable competitive location problem where facilities are simultaneously subject to (i) symmetric or leader-follower types of competitions, and (ii) location-dependent probabilistic failures. An optimization model is formulated to capture the symmetric Nash competition between two companies. The goal is to maximize the expected profit (service revenue minus the sum of initial facility construction costs and the expected customer transportation costs) under normal and failure scenarios. Building upon this result, we build a bilevel leader-follower Stackelberg competition model to derive the optimal facility location design when one of the companies has the first-mover advantage over its competitor. Our CA approach is able to effectively solve the models. For special cases, closed-form analytical solutions can be obtained. Numerical experiments with hypothetical data and a case study for competitive biofuel supply chain design in the State of Illinois are conducted. The results revealed managerial insights on how competing companies should optimally plan their facility locations. Then, we propose a systematic optimization framework to analyze how biofuel supply chain decisions are affected by (i) crop yield/supply uncertainty, (ii) refinery disruption risks, and (iii) competition against existing food supply chains. The interactions among the biofuel industry, farmers and food industry are captured by a Stackelberg-Nash game, formulated under a CA scheme. The expected profits of both the farmers and the biofuel industry are evaluated based on probability distributions of crop yield and refinery disruption risks over space. Functional optimization, e.g., variational calculus, is used to derive the equilibrium conditions and suggest numerical algorithms. A series of numerical experiments are conducted for both hypothetical test cases and a Midwest case study to (i) show computational performance and robustness of the modeling approach, (ii) analyze the impacts of system parameters, as well as (iii) draw managerial insights in realistic settings. In addition, we propose a heuristic modeling framework to overcome the challenge that applying CA in solving dynamic facility location problems. First, we formulate a continuous model for the dynamic version by augmenting the time dimension, while relaxing the location consistency constraints. To translate the CA output into a set of discrete facility locations, we extend the disk model (for one static time period) to a tube model (for multiple time periods). Then, the location consistency constraints are enforced through a nonlinear optimization model with penalty terms. Lastly, we propose an iterative tube regulation algorithm to solve the penalty-based optimization problem. We analyze the accuracy and convergence of our modeling framework and conduct numerical experiments to verify its performance. The model and the solution procedure we proposed are very generic and flexible; thus, it can be extended to variants (e.g., incorporating existing facilities at the beginning of the horizon). Finally, we investigate a difficult trilemma: with limited farmland, how does the government stimulate the growth of the biofuel industry while, at the same time, protect food security and preserve environmental sustainability? Our framework is applied to address such multiple cross-interacting systems associating with the biofuel industry development in a broader context. We aim to provide policy guidelines on governmental mandates to induce socially favorable farmland use configurations to support a sustainable bio-economy.

Graduation Semester

2015-8

Type of Resource

text

Permalink

http://hdl.handle.net/2142/88208

Copyright and License Information

Copyright 2015 Xin Wang

Owning Collections