Robust optimization for the Segment Routing Traffic Engineering Problem

Segment routing, a modern protocol enhancing traffic engineering, introduces flexibility by enabling traffic to take detours. This talk tackles the challenge of optimizing routing in the face of uncertain traffic distribution. Unlike traditional methods relying on a single traffic matrix, our approach considers an infinite set of matrices defined by linear constraints. Our goal is to optimize routing under the worst-case scenario within this set. Through innovative formulations, our results showcase a substantial speed improvement in optimization compared to traditional methods, offering an efficient alternative to exploring all extreme points in the matrix set.