Investigation of the Computational Utility of Geometric Programming Formulation and Algorithms.
Abstract
This project proposes to investigate the effectiveness of prototype and generalized geometric programming (GP) developments in the routine solution of inequality constrained algebraic programs. The research seeks to elucidate under what conditions which of the equivalent prototype GP problem forms: primal, convexified primal, transformed primal, and, or transformed dual, as well as their generalized GP analogous are best solved and under what conditions. Moreover, direct and sequential strategies will be examined to identify the most efficient strategies for the solution of GGP's. The computational tests in the study will be performed using the leading specialized GP and GGP codes as well as a selection of general nonlinear programming codes. Statistical tests will be employed to deduce suitable measures of prototype and generalized GP problem difficulty. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 1981
- Accession Number
- ADA098245
Entities
People
- G. V. Reklaitis
Organizations
- Purdue University