C. Domínguez Sánchez, J. M. Morales González, S. Pineda, Á. Porras Cabrera
Chance-Constrained Programming (CCP) is one of the main methods to deal with uncertainty when the feasibility of all the constraints is not required. It allows to obtain solutions specifying a desired acceptable level of reliability (or a tolerance to risky outcomes) by means of a probability, leading to less conservative solutions at the expense of models with non-convex feasible regions. CCPs find applications in power systems, supply chain problems and portfolio optimization problems, among others.
In general, the probability distribution is unknown or difficult to deal with. The Sample Average Approximation (SAA) allows for a deterministic reformulation of the problem assuming a finite discrete distribution. The resulting model is equivalent to a MIP. In this work, we propose preprocessing techniques and valid inequalities to tighten the resultant MIPs, and we apply our results to the Optimal Power Flow problem.
Keywords: Combinatorial Optimization, Chance-constrained problems, Probabilistic constraints, Mixed-integer optimization, Optimal Power Flow
Scheduled
Integer and Combinatorial Optimization
June 10, 2022 4:00 PM
Grade Hall
