The Quasi-Randomized Approach to Uncertain Convex Programs

Many engineering problems can be cast as uncertain convex optimization problems with convex constraints. Although the robust approach and the chance-constrained approach have been introduced, both of them may lead to computationally intractable problems. A computationally tractable randomized approach was proposed recently which resulted in an approximation of original problem and a random solution depending on the random samples. In this paper, a quasi-sampled program is presented which not only inherits the merits of the randomized approach but also obtains a determined solution. Some numerical examples demonstrate the good performance, such as exactness and lower violation probability, of the proposed method.