Low-Complexity Interior Point Algorithms for Stochastic Programming: Derivation Analysis and Performance Evaluation

Abstract

The broad purpose of this project was to investigate low-complexity interior point decomposition algorithms for stochastic programming. A specific objective was to evaluate algorithms using test problems arising from useful applications. The important direct results of this project include: (1) a new test problem collection that includes problem instances from a variety of application areas; (2) a new package of C-routines for converting SMPS input data into data structures more suitable for implementing algorithms; (3) a new software package, CPA, for two-stage stochastic linear programs. The test problems and input conversion routines have been developed in a general manner to be useful to other researchers. CPA includes volumetric center algorithms that proved to be successful in our computational evaluations. To the best of our knowledge, CPA is the only software for stochastic programming that includes volumetric center algorithms. Items (1), (2) and (3) are freely accessible over the Internet. The important theoretical results of this project include: (4) a new characterization of convexity-preserving maps; (5) a new coordinate-free foundation for projective spaces; (6) a new geometric characterization of one-dimensional projective spaces; (7) new algorithms for bound-constrained nonlinear optimization. These theoretical results are likely to be useful in computational optimization in general.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jan 01, 2000
Accession Number
ADA418278

Entities

People

  • K. A. Ariyawansa

Organizations

  • Washington State University

Tags

Communities of Interest

  • Energy and Power Technologies
  • Human Systems
  • Weapons Technologies

DTIC Thesaurus Topics

  • Algorithms
  • Applied Mathematics
  • Computer Programming
  • Computer Science
  • Evolutionary Algorithms
  • Linear Programming
  • Mathematical Programming
  • Mathematics
  • Networks
  • Optimization
  • Probability Distributions
  • Random Variables
  • Simplex Method
  • Standards
  • Students
  • Vector Spaces
  • Websites

Fields of Study

  • Computer science

Readers

  • Computer Science.
  • Operations Research

Technology Areas

  • Space