A List of Responders
Virtual Manufacturing Background Project
Dr. Norman M. Sadeh, Research Scientist and Co-Director, Intelligent Coordination and Logistics Laboratory, The Robotics Institute, School of Computer Science, Carnegie Mellon University sadeh@cs.cmu.edu Dr. Stephen F. Smith, Senior Research Scientist and Director, Intelligent Coordination and Logistics Laboratory, The Robotics Institute, School of Computer Science, Carnegie Mellon University sadeh@cs.cmu.edu
For the past several years, our laboratory has been involved in the
development of innovative production planning, scheduling and control
technologies, which we generically refer to as "Knowledge-Based Production
Management" techniques. Key distinguishing characteristics of these new
techniques include:
- Finite capacity scheduling models that optimize production
objectives while accounting for the full complexity of the
production environment, enabling more realistic planning (e.g.,
inventory management, order promising, integration with process
planning, etc.), and improved manufacturing performance,
- Incremental techniques that support rapid revision of production
schedules, enabling efficient response to changing circumstances
(e.g. changes in demand, machine breakdowns, etc.), integration
with real-time control systems, flexible interactive user
manipulation of schedules and exploration of alternative scenarios,
- Integration of finite capacity scheduling with higher-level
planning decisions (e.g. order promising, inventory management,
overtime and subcontracting decisions) and implementation of
integrated production management strategies.
- Promotion of encapuslation and re-use of component services,
enabling rapid adaptation to accommodate changing application
requirements (e.g. new products/processes, market shifts) and
different application environments.
- Emphasis on multi-plant coordination and supply chain management
issues through development of decision support tools to help
enterprises in (1) rapidly establishing efficient supply chains for
new products/projects, (2) efficiently coordinating material flows
across these chains, and (3) assessing the impact of various
supplier/customer agreements.
The power of these concepts and techniques has been demonstrated through
development of a progression of software systems and extensive comparative
performance analyses with conventional state-of-the-art techniques. Various
software systems and tools resulting from this research have been transferred
or are being transferred to various industrial organizations, including
Westinghouse, Intel, IBM, Mitsubishi, GE, McDonnell Douglas, Raytheon and
United Technologies, and many other scheduling systems now in commercial and
government use are directly based on our techniques.
Our laboratory is currently working on several major virtual/agile
manufacturing efforts, including an Agile Manufacturing Initiative
project aimed at developing an integrated process planning/production
scheduling shell in collaboration with Raytheon and under ARPA
sponsorship.
Applications of these technologies in virtual manufacturing include:
1. VISUALIZATION: Visualization concerns are an integral
part of our work in planning and coordination technologies for
virtual manufacturing. This includes development of innovative
visualization technologies to support mixed-initiative planning
and scheduling functionalities (e.g. visualization of inefficiencies
or opportunities for improvement in an existing production schedule,
visual evaluation of schedules along multiple dimensions), visualization
of different schedule execution modes under different sources of executional
uncertainty, and tools to visualize supply chain performance under
different coordination protocols.
2. ENVIRONMENT CONSTRUCTION TECHNOLOGIES:
A major objective of our work is to develop VM environments that
support the rapid development, refinement and dynamic maintenance
of high-quality high fidelity production plans and schedules both
within single manufacturing facilities as well as across multiple
manufacturing entities.
3. MODELING TECHNOLOGIES:
A key aspect of our research has been the development of modular/customizable/
re-usable planning and coordination software systems. These concepts are
emphasized in our work within the context of the Ditops and Micro-Boss
scheduling systems and in our multi-agent modeling and simulation tools
for supply chain analysis.
4. REPRESENTATION:
related to 3.
5. META-MODELING:
We are currently working with Raytheon on the development of an integrated
process planning/production scheduling shell for agile manufacturing.
Two important aspects of this work include compatibility with existing
standards (e.g. STEP) and support for integration with legacy systems.
6. INTEGRATING INFRASTRUCTURE & ARCHITECTURE:
*Development of a blackboard architecture to support integration of
of process planning and production scheduling. This includes
work to support coordination with other sites within the supply
chain (e.g. order promising, re-ordering decisions, etc.)
*Supply Chain Modeling and Analysis: development of muti-agent
modeling and simulation tools to help companies compare alternative
supply chain configurations, evaluate the impact of different
partnership agreements and different coordination policies.
7. SIMULATION:
*Simulation of shop floor operations under different sources of
uncertainty (e.g. machine breakdowns, dynamic order arrivals, etc.).
Simulation complements our finite capacity scheduling
technologies and provides support for what-if evaluation of
different strategic, tactical and operational alternatives
(e.g. different supply agreements, overtime alternatives,
different schedule execution policies, etc.)
*Simulation is also a central component of our work in supply chain modeling
and analysis where it is used to evaluate the impact of (1) different
supply chain configurations, (2) different buyer-supplier agreements and
(3) different coordination protocols (e.g. quick response
protocols), on the performance of the overall supply chain as well as
on the performance of individual entities within the supply chain.
8. METHODOLOGY:
Our work in planning and coordination combines a number of
methodologies and problem solving paradigms. Our lab has developed
a number of (opportunistic) constraint-directed scheduling search
procedures to support rapid development and dynamic revision of high
quality production schedules. We have developed blackboard-based architectures
to support integration of multiple heuristics and/or systems, enabling solution
development and revision from multiple problem-solving perspectives.
We have developed innovative optimization techniques including
adaptive simulated annealing procedures. We are developing
multi-agent modeling and simulation technologies for supply
chain analysis, etc.
9. INTEGRATION OF LEGACY DATA:
Our work in integrated process planning and production scheduling: See
above
10. MANUFACTURING CHARACTERIZATION:
See above
11. VERIFICATION, VALIDATION & MEASUREMENT:
12. WORKFLOW:
Our work in supply chain analysis includes comparison of different
supply chain coordination protocols.
13. CROSS-FUNCTIONAL TRADES:
e.g. our work in integrated process planning/production scheduling
2) a list of relevant references;
Nakakuki, Yoichiro, and Norman Sadeh. Increasing the Efficiency of
Simulated Annealing Search by Learning to Recognize (Un)Promising Runs.
Proceedings of the Twelfth National Conference on Artificial Intelligence
(AAAI-94), 1994, pp. 1316-1322.
Norman Sadeh. Micro-Opportunistic Scheduling: The Micro-Boss Factory
Scheduler. In Intelligent Scheduling, Morgan Kaufmann, 1994, Chap. 4.
Norman M. Sadeh. Micro-Boss: Towards a New Generation of Manufacturing
Scheduling Shells. Proceedings of the ARPA/Rome Laboratory Knowledge-Based
Planning and Scheduling Initiative, Tucson, AZ, Februrary, 1994, pp. 191-203.
Norman M. Sadeh, Tom Laliberty, Stephen F. Smith and Robert Bryant.
An Integrated Process Planning/Production Scheduling Shell for Agile
Manufacturing. Working Paper. The Robotics Institute, Carnegie Mellon
Univertsity, Pittsburgh, PA 15213-3891.
Sadeh, Norman, and Yoichiro Nakakuki. "Focused Simulated Annealing Search:
An Application to Job Shop Scheduling". Annals of Operations Research (1995).
To appear in issue on 'Metaheuristics in Combinatorial Optimization.'.
Smith, S.F., and N.M. Sadeh. Knowledge-based Production Management. .
Tutorial SP4 - Tenth National Conference on Artificial Intelligence (AAAI-92).
Katia Sycara, Stephen Roth, Norman Sadeh, and Mark S. Fox. "Resource
Allocation in Distributed Factory Scheduling". IEEE EXPERT 6, 1 (February
1991), 29-40.
Swaminathan, J., N.M. Sadeh, and S.F. Smith. Impact of Supplier
Information on Supply Chain Performance. Robotics Institute, Carnegie Mellon
University, Pittsburgh, PA 15213., 1994.
Swaminathan, J., S.F. Smith and N.M. Sadeh. Modeling the Dynamics of
Supply Chains. Proceedings of the AAAI-94 SIGMAN Workshop, Seattle, WA,
August, 1994, pp. 113-122.
3) The URL address for our lab's WWW home page is:
http://cs.cmu.edu:8001/afs/cs.cmu.edu/project/ozone/www/icllab.html
A List of Responders
Virtual Manufacturing Background Project
Virtual Manufacturing Home Page