NEXT GENERATION OF UNIT COMMITMENT MODELS

[Mary Steffens]

********************************************************************
*                           DIMACS                                 *
* Center for Discrete Mathematics and Theoretical Computer Science *
*   A National Science Foundation Science and Technology Center    *
********************************************************************
  
	    NEXT GENERATION OF UNIT COMMITMENT MODELS
  			  A DIMACS Workshop
                   September 27 - 28, 1999
       DIMACS Center, Rutgers University, Piscataway, NJ

Organizers:

Michael H. Rothkopf, Rutgers University, RUTCOR
rothkopf at rutcor.rutgers.edu

Benjamin F. Hobbs, Johns Hopkins University
bhobbs at jhu.edu

Hung-Po Chao, Electric Power Research Institute
hchao at epri.com

Co-Sponsored by DIMACS and the Electric Power Research Institute

Presented under the auspices of the Special Year on Large
Scale Discrete Optimization.


For many years, the electric power industry has been using methods 
to solve the unit commitment problem. As the industry is in the 
process of being transformed into one that relies increasingly 
on auction markets, issues arise about the adequacy of the 
traditional unit commitment methods.  This conference will 
explore the technology and needs of the next generation of 
computer models for aiding unit commitment decisions in 
competitive markets.

The unit commitment problem can be defined as the scheduling of 
production from power generation units over a daily to weekly 
time horizon in order to accomplish some objective.  The problem 
solution must respect both generator constraints (such as ramp 
rate limits and minimum up or down times) and system constraints 
(reserve and energy requirements and, potentially, transmission 
constraints).  The objective function should account for costs 
associated with energy production and start-up and shut-down 
decisions, along with possible revenue or customer impacts of
those decisions.  The resulting problem is a large scale 
nonlinear integer program.

Because of the problem's size and complexity, and large
economic benefits that could result from its improved solution,
considerable attention has been devoted to algorithm development. 
However, the solution approach that has been most successful, 
and which is most widely used at present, is Lagrangian 
relaxation. Lagrangian relaxation has proven useful for quick 
development of good, if not optimal, generator schedules.  
However, recent improvements in integer programming codes and 
other algorithms suggest that it may be possible to
find better solutions more rapidly. Due to major structural 
changes in the industry, a reconsideration of how unit commitment 
models are solved and the purposes for which they are used is 
needed.  The goal of this workshop is to bring together those 
who know the problem and what is needed with those who know what 
improvements in algorithms are really possible.
 
Presentations by top experts in algorithms and utility systems
operations will be made. Then working groups will convene to 
define the new roles that unit commitment models may play
and promising directions in algorithm development and application. 
The goal of this workshop is to ultimately result in better 
algorithms yielding more value to their users.

Those interested in giving a talk should send a title and abstract
(full papers will be accepted if available) in plain text to Michael
Rothkopf (rothkopf at rutocr.rutgers.edu) by July 30, 1999.  There is a
$40/day, $5/day for students/postdocs registration fee for this
workshop. Information on registration, accommodations, and travel 
can be found at:

http://dimacs.rutgers.edu/Workshops/NextGeneration/

Advisory Committee: 
Ross Baldick, University of Texas, Austin; Robert Bixby, Rice 
University; Sebastian Ceria, Columbia University; George Gross, 
University of Illinois; Karla L. Hoffman, George Mason University; 
Ellis Johnson, Georgia Institute of technology; Khai Le, ABB; Fred 
Lee, University of Oklahoma; Richard O'Neill, Federal Energy 
Regulatory Commission; Shmuel Oren, University of California, 
Berkeley; Andras Prekopa, Rutgers University; Fred Roberts, Rutgers 
University; S. Mohammed Shahidehpour, Illinois Institute of Technology; 
David Shanno, Rutgers University; Gerald Sheble, Iowa State University;
David Sun, ALSTOM ESCA; Samer Takriti, Enron Capital and Trades


If you have further questions about the meeting, please contact:

Ben Hobbs
DoGEE
313 Ames Hall
The Johns Hopkins University
410 516-4681
410-516-8996 fax
bhobbs at jhu.edu

Michael Rothkopf
Rutgers University
(732)445-5632
rothkopf at rutcor.rutgers.edu

This message was sent to the Acord_List list of INFORMS Online.
Only members of the list can post to the list.  Send postings to
acord_list at mail.informs.org.  Moderator:
informs_subdiv at mail.informs.org