Parallel algorithm for nonlinear network optimization problems and real-time applications

Abstract

In this paper, we propose a parallel algorithm to solve a class of nonlinear network optimization problems. The proposed parallel algorithm is a combination of the successive quadratic programming and the dual method, which can achieve complete decomposition and make parallel computation possible. The proposed algorithm can be applied to solve nonlinear network optimization problems in the smart grid. We have tested the proposed parallel algorithm in solving numerous cases of power flow problems on the IEEE 30-bus system. The test results demonstrate that the proposed parallel algorithm can obtain accurate solution. Additionally, neglecting the data communication time, the proposed parallel algorithm is, ideally, 13.1 times faster than the centralized Newton Raphson's method in solving the power flow problems of the IEEE 30-bus system. © 2012 Springer-Verlag.