Title: Tractable Approximations to Robust Conic Optimization Problems (paper by D. Bertsimas, M. Sim)
| Speaker: | Matthew William Slavin |
| Affiliation: | University of Waterloo |
| Room: | MC 5479 |
Abstract: We review the paper listed in the title of this talk. In the paper, Bertsimas and Sim propose a relaxed robust counterpart for general conic optimization problems that a) preserves the computational tractability of the nominal problem; specifically the robust conic optimization problem retains its original structure (i.e. robust LPs remain LPs, robust SDPs remain SDPs, etc.), and b) provides a guarantee on the probability that the robust solution is feasible when the uncertain coefficients obey i.i.d normal distributions).
We will review the construction of Bertsimas and Sim’s proposed robust counterpart in detail for LP, QCQP, and SDP problems in detail, and review their probability bounds for these problems as much as time will allow.