A general pattern search algorithm for electric power system unit commitment problems

作者：Li, Jinghua; Wei, Hua

来源：Transactions of China Electrotechnical Society, 2009, 24(6): 121-128.

• Discrete variables
• Filter algorithm
• Mixed variables
• Pattern search algorithm
• Unit-commitment

Unit commitment belongs to mixed variables programming which is difficult to find the optimal solution in mathematics. This paper presents a general pattern search algorithm with mixed variable (GPSMV) to solve unit commitment problems at the first time. The proposed algorithm can solve problems for which the objective function is nonlinear, nonconvex, nondifferentiable, stochastic, or even discontinuous. GPSMV guarantees global convergence and it only needs values of objective function and barrier functions consisted by constraint condition while discards the information of their derivative. During the process of optimization, the discrete variables can be deal directly while not dividing original problem to be discrete and continuous part. Simulations are executed on six systems of 10-100 units and 26 units in 24 time intervals, and the results verify the effectiveness of the proposed algorithm.

• 出版日期2009-6
• 单位广西大学