2 options
Using genetic algorithms to determine near-optimal pricing, investment and operating strategies in the electric power industry.
Connect to full text Available online
View online- Format:
- Book
- Thesis/Dissertation
- Author/Creator:
- Wu, Dongjun.
- Language:
- English
- Subjects (All):
- Computer science.
- Operations research.
- Power resources.
- Management.
- 0454.
- 0791.
- 0796.
- 0984.
- Penn dissertations--Managerial science and applied economics.
- Managerial science and applied economicx--Penn dissertations.
- Local Subjects:
- Penn dissertations--Managerial science and applied economics.
- Managerial science and applied economicx--Penn dissertations.
- 0454.
- 0791.
- 0796.
- 0984.
- Physical Description:
- 158 pages
- Contained In:
- Dissertation Abstracts International 58-03A.
- System Details:
- Mode of access: World Wide Web.
- text file
- Summary:
- Network industries have technologies characterized by a spatial hierarchy, the "network," with capital-intensive interconnections and time-dependent, capacity-limited flows of products and services through the network to customers. This dissertation studies service pricing, investment and business operating strategies for the electric power network. First-best solutions for a variety of pricing and investment problems have been studied. The evaluation of genetic algorithms (GA, which are methods based on the idea of natural evolution) as a primary means of solving complicated network problems, both w.r.t. pricing: as well as w.r.t. investment and other operating decisions, has been conducted. New constraint-handling techniques in GAs have been studied and tested. The actual application of such constraint-handling techniques in solving practical non-linear optimization problems has been tested on several complex network design problems with encouraging initial results. Genetic algorithms provide solutions that are feasible and close to optimal when the optimal solution is know; in some instances, the near-optimal solutions for small problems by the proposed GA approach can only be tested by pushing the limits of currently available non-linear optimization software. The performance is far better than several commercially available GA programs, which are generally inadequate in solving any of the problems studied in this dissertation, primarily because of their poor handling of constraints. Genetic algorithms, if carefully designed, seem very promising in solving difficult problems which are intractable by traditional analytic methods.
- Notes:
- Thesis (Ph.D. in Managerial Science and Applied Economics) -- University of Pennsylvania, 1997.
- Source: Dissertation Abstracts International, Volume: 58-03, Section: A, page: 0990.
- Local Notes:
- School code: 0175.
- ISBN:
- 9780591362848
- Access Restriction:
- Restricted for use by site license.
The Penn Libraries is committed to describing library materials using current, accurate, and responsible language. If you discover outdated or inaccurate language, please fill out this feedback form to report it and suggest alternative language.