High performance solver for structured optimization problems

• Grant number: 96/02341-2
• Support Opportunities: Research Grants - Research Partnership for Technological Innovation - PITE
• Start date: October 01, 1996
• End date: May 31, 1999
• Field of knowledge: Physical Sciences and Mathematics - Mathematics - Applied Mathematics
• Principal Investigator: Julio Michael Stern
• Grantee: Julio Michael Stern
• Host Institution: Instituto de Matemática e Estatística (IME). Universidade de São Paulo (USP). São Paulo , SP, Brazil
• Company: Unisoma Matemática Ltda
• City: São Paulo

Abstract

The present research project had the objective of developing Mathematical Programming solvers especially adapted for the optimization of RBAF - Row Block Angular Form and NRBAF - Nested Row Block Angular Form. The project adapted and implemented the structural method developed by the researcher responsible for it as part of his thesis for a doctorate at Cornell University. This method is presented in the article “J.M. Stern and S.A. Vavasis, Active Set Methods for Problems in Column Block Angular Form, Com.Appl.Math”, v. 12, p. 199-226, 1994. The solvers were to replace commercial solvers (OSL-IBM) in the MultiFor and DyFor computer programs (software) in UniSoma. This software with an industrial application solves optimization problems of multiple mixtures, with an RBAF structure, and multiperiod optimization problems, with an NRBAF structure. (AU)