Solving unit commitment problem with parallel computing

Abstract

The Unit Commitment Problem (UCP) involves finding the least-cost dispatch of available generation resources to meet the electricity demands over an electrical grid. It is an important planning problem in the electric utility industry. The purpose of the UCP is to determine when to start up and when to shut down a particular generating unit and how to dispatch the committed units to meet the electricity demands and satisfy the relevant security constraints. This paper presents a procedure for modeling and solving large scale unit commitment problems (UCP) by using a parallel computing approach. We assume that the electricity demands are known, and a predetermined amount of reserve is used to guarantee the security of the system. A parallel computing version of the Lagrangian Relaxation technique is used to solve this problem efficiently. © Springer-Verlag Berlin Heidelberg 2014.