Matthew Fontana

 

Robust optimization has been in the center of much research activity in optimization. We develop a new approach to multi-objective robust optimization, and provide fully polynomial-time approximation schemes when the deterministic multi-objective problem is easy to solve.

In addition, we consider the robust single-objective optimization problem under a certain kind of uncertainty, which can be framed as a deterministic bi-objective optimization problem. Building on earlier work by Bertsimas and Sim (2004), we develop an algorithm to solve this problem, and report on its computational performance. In particular, we apply the algorithm to solve the robust shortest-path problem and present additional results from computational experiments.

This is joint work with Dimitris Bertsimas and Thomas Magnanti.