Distributionally robust joint chance constraints with second-order moment information (Q1942277)

The chance constraint stochastic programming is considered assuming that only the first- and second-order moments of the random parameters and their supports are given. The authors show that the worst-case conditional value at risk approximation of robust individual constraints with quadratic constraint functions is exact. Moreover, such an approximation can be efficiently computed. For the case of joint chance constraints, the tightness of the proposed approximation is evaluated. An algorithm based on this approximation is implemented and tested on an applied problem related to the dynamic water reservoir control for the hydro-power generation.