Faculty Advisor

Martin, William J.

Abstract

Semidefinite programming is a type of convex optimization that aims to optimize a linear function, the trace of the product of a matrix and the variable matrix X, while subject to nonlinear constraints. In a semidefinite program (SDP), the decision matrix X is required to be positive semidefinite. We examine an interior point method for solving SDPs and explore its application to the Quadratic Assignment Problem (QAP), an NP-hard problem used to assign n facilities to n locations, minimizing the quadratic objective function of the product of distances between locations and flow between facilities. Using a relaxation of the QAP formulation into an SDP, we solve QAP relaxations using the NEOS solver, a web service for numerical optimization, with the interior point method described above.

Publisher

Worcester Polytechnic Institute

Date Accepted

April 2019

Major

Mathematical Sciences

Project Type

Major Qualifying Project

Accessibility

Unrestricted

Advisor Department

Mathematical Sciences

Share

COinS