Comparisons of different techniques in semidefinite programming for hydrothermal coordination problem

Abstract

Hydrothermal coordination problem (HTC) is to minimize fuel cost of thermal units under a series of constraints by means of optimization techniques. Semidefinite programming (SDP) is an emerging powerful tool to do so. This paper presents four different programming skills used in transforming HTC into its corresponding SDP models. Through a HTC test case, the computational efficiencies and the results of these skills, which are implemented on a convex solver CVX, are compared and analyzed. The conclusion shows that there exists a best way to express HTC as SDP model without exploring the sparsity of large dimensional matrices produced during the transformation.