Optimization Over Combinatorial Optimization Polytopes

Abstract

This is a renewal proposal for ONR contract N00014-14-1-0072 on developing efficient techniques for solvingcombinatorial optimization problems. The proposed research considers two unrelated directions. One is on the large class of incremental optimization problems, in which the goal is to choose in which order to make decisions so as to maximize the cumulative benefit over all time periods. This is a very important planning problem for disaster recovery, among other applications. The proposed research in this direction aims to provide powerful techniques for solving these problems efficiently, either exactly or approximately. The second direction aims to apply hyperbolic optimizationfor the solution of combinatorial optimization problems. Hyperbolic optimization is a very large class of efficientlysolvable convex optimization problems that includes the well-known classes of linear programming and semi-definite programming, but which appears to have more expressive power.

Document Details

Document Type
DoD Grant Award
Publication Date
Jan 04, 2017
Source ID
N000141712177

Entities

People

  • Michael Goemans

Organizations

  • Massachusetts Institute of Technology
  • Office of Naval Research
  • United States Navy

Tags

Readers

  • Distributed Systems and Data Platform Development
  • Operations Research