A Unifying Geometric Solution Framework and Complexity Analysis for Variational Inequalities
Name
OR-276-93.pdf
Size
2.45 MB
Format
Adobe PDF
Checksum (MD5)
e92bfd3744cd61df652b194f10a7b254
Author(s)
Magnanti, Thomas L.
Perakis, Georgia
Date Issued
February 1996
Publisher
Massachusetts Institute of Technology, Operations Research Center
Series/Report no.
Operations Research Center Working Paper;OR 276-93
Abstract
In this paper, we propose a concept of polynomiality for variational inequality problems and show how to find a near optimal solution of variational inequality problems in a polynomial number of iterations. To establish this result we build upon insights from several algorithms for linear and nonlinear programs (the ellipsoid algorithm, the method of centers of gravity, the method of inscribed ellipsoids, and Vaidya's algorithm) to develop a unifying geometric framework for solving variational inequality problems. The analysis rests upon the assumption of strong-f-monotonicity, which is weaker than strict and strong monotonicity. Since linear programs satisfy this assumption, the general framework applies to linear programs.
MIT Department
Massachusetts Institute of Technology. Operations Research Center
Persistent DSpace Link
