Abstract
Randomized algorithms for optimal power flow problems are presented. A standard optimal power flow problem is to minimize a fuel cost satisfying power flow equations and some inequalities consisting of the limits on control variables and the operating limits of a power system. Then, the robust optimal power flow problem is considered when uncertain active and reactive power is injected into the power system. Although both problems are highly nonconvex, the proposed algorithms always stop within the finite number of iterations and find a suboptimal solution which is a feasible in a probabilistic sense. It is shown that the maximum number of random samples is a polynomial of the size of the problems.
