Two-layer Progressive Hedging-Benders Decomposition Algorithm for Stochastic Unit Commitment

Abstract

With the large-scale development of renewable energy generation, the uncertainty of power system has increased significantly. In order to make proper day-ahead scheduling decisions, the stochastic unit commitment (SUC) problem considering reserve capacity and line power flow limit is studied, and multiple independent random scenarios are used to represent the uncertainty of load and renewable generation in the system. In order to solve this large-scale problem caused by too many scenarios, a two-layer PH-BD decomposition algorithm is designed to further improve the computing speed. The outer Progressive hedging (PH) algorithm is used to decompose the scenarios to realize parallel solution of the corresponding problems in each scenario. The inner Benders decomposition (BD) algorithm is used to decompose the corresponding problems of each scenario into the main problem without line constraints and the subproblem with line constraints. The calculation results of IEEE 118-bus system show that the calculation speed of the proposed algorithm is faster and the solution time is shorter than that of the traditional PH algorithm.