Research Interests of David Wu

Manufacturing, Production and Logistics Systems

This area of research is motivated by industry interactions over the past ten years with Bethlehem Steel, Lucent Technologies, Ford Motor Company, UPS, PP&L, Unisys and a number of medium size manufacturing companies. While my earlier research focused on planning and scheduling robustness, the industry interaction has broadened my interests considerably. Along with my colleagues we identified a new focus in manufacturing called Manufacturing Logistics, referring to the planning, coordination, and integration issues among manufacturing facilities. My particular interest is in the coordination of production and shipment planning in manufacturing supply chains (specifically, automotive and electronics supply chains). In 1995, I co-founded the Manufacturing Logistics Institute (MLI), a research group conducting academic research on real industry problems. We have been quite successful in getting NSF and private sector supports for these research activities (a MLI brochure summarizing current projects is attached). More excitingly, MLI has assumed a leadership role in charting a national research agenda in manufacturing and logistics. In May 29-30, 1997, MLI hosted an NSF workshop on Manufacturing Logistics where more than 100 academic and industry researchers attended. After lively discussion and many debates, a comprehensive outline of future research directions was put into shape. In 1998, a workshop organized for UPS will take place, focusing on Supply Chain Integration services for their regional and worldwide customers.

Specific Topics:

  • Planning and Scheduling Robustness, Assembly Design
  • Supply Chain Integration via Coordinated Manufacturing Planning
  • Capacity Planning, Configuration, and Allocation for Semiconductor Facilities
  • Demand Characterization in Electronics and Automotive Supply Chains
  • Sipment Planning to Achieve Supply Chain Lead Time Reduction

    This line of research has been supported by several NSF grants, and several grants provided by the industry. I have collaborated many of these activities with Bob Storer and Laura Burke and this has been the subject of a few dissertations by my Ph.D. students Mary J. Meixell, Kadir Ertogral, Kedar Naphade, Jorge Leon, Eui-Seok Byeon, Michael Bartolacci, Jewel Bonser and Michael Chang, and master's students Phil Brennan, Nabeela Al-Refai, Hakan Gobashi and Koichi Tsuruta.

Distributed and Game Theoretic Decision Processes

My research in this area examines theories and analytic tools basic to various industrial decision processes. My focus has been to methods that reconcile centralized, monolithic planning processes with distributed, multi-facet operational environments. This entails theories and models that reconcile a priori stochastic analysis with dynamic recourse policies, reconcile global strategic objectives (e.g., business planning) with local constraints/ preferences (e.g., manufacturing operations), and reconcile aggregated a priori information (e.g., forecast, market analysis) with actual observations over time (e.g., actual demand, sales data). Techniques rooted in graph theory, mathematical programming, game/group decision theory, and robust /stochastic optimization provide me with the primary tools to address these problems. Most research are motivated by experiences from modeling real industrial planning systems.

Specific Topics:

  • Distributed Decision Making
  • Auction and Bidding Models
  • Mapping Distributed and Monolithic Optimization Models
  • Graph-Theoretic Decomposition
  • Lagrangean Methods for Model Preprocessing
  • Robust A Priori Optimization
  • Applications: Supplier Management, Networks Traffic Management, Reconcile Contracted and Short-Term Procurement Decisions (PP&L), Reconcile Centralized Resources with Multi-Facet, Market Specific Requirements (Ford, Lucent), Reconcile Forecasted and Actual Demand Information (Lucent), Reconcile Business Planning with Manufacturing Process Design (Bethlehem Steel)

    This line of research has been supported by an NSF grant and a grant from U.S. Air Force, and several grants provided by the industry. I have collaborated many of these activities with Pat Harker at the University of Pennsylvania and this has been the subject of a few dissertations by my Ph.D. students Erhan Kutanoglu, Suleyman Karabuk, and Hakan Golbasi.

Discrete and Combinatorial Optimization

My focus in this area is to develop search algorithms, decomposition schemes, relaxation methods and other computational techniques that allow the efficient solution of optimization problems. Of special interests are discrete and combinatorial optimization. The research is motivated by special structures often observed in real-world optimization problems, which provide insights for efficient algorithm design. Research has been conducted in construction project management (ATLSS), production scheduling (Neapco, Bethlehem Steel, Tray-Pak), critical resource scheduling (Bethlehem Steel), software project management (UNISYS), VLSI design and routing problems (AT&T), telecommunication network routing, and assembly design and planning problems.

Specific Topic Areas:

  • Specialized Search Algorithms: Problem Space Search, Simulated Annealing, Genetic Algorithms
  • Lagrangean Heuristics
  • Graph Theoretic Algorithms
  • Ordinal Comparison of Algorithms
  • Special Structures in Scheduling and Routing Problems

I have collaborated with Bob Storer on many of these problems. The students I have collaborated with in this line of research includes Kedar Naphade , Mike Bartolacci, I.-K. Park (Ph.D. students), and Ranzo Vaccari, Bhavin Doshi, Jaime Bustos (Master's).