基于智能混合算法的经济调度问题及其参数调整方法  

Hybrid AC and SQP Algorithm and Its Parameters Adjustment for Economic Dispatch Problem with Valve-point Effect

作　　者：罗中良[1] 易明珠[2] 周有平[1] 刘小勇[3] 

机构地区：[1]佛山科学技术学院电气工程及自动化系,广东佛山528000 [2]中山大学岭南学院,广东广州510275 [3]西安交通大学自动控制系,陕西西安710049

出　　处：《中山大学学报（自然科学版）》2008年第1期33-37,共5页Acta Scientiarum Naturalium Universitatis Sunyatseni

基　　金：国家自然科学基金资助项目(60605009)

摘　　要：带阀点效应的经济调度问题,属非凸集的发电成本最小化问题,具有不连续、不可微分且易陷入局部最优解。为了获得整体近似最优解,提出采用蚁群算法,搜索发电机可运行状态的最优组合,对蚁群算法的数学模型进行分析,讨论算法参数的调整方法,给出具有普遍意义的收敛性定理。在此求解过程中,以每只蚂蚁来表示符合限制条件的某个可运转状态的发电机组合并以序列二次规划法来求解传统的经济调度问题。通过三部机组的数值模拟,验证其正确有效性。The economic dispatch （ED） problem with valve-point effect is generally formulated as a non-smooth, non-differential and cost-minimizing problem. Due to its non-convex and highly nonlinear properties, the solution entraps in a local minimum easily. For the purpose of discovering the near-globally optimal solution, a hybrid approach of ant colony （AC） algorithm and sequential quadratic programming （SQP） was proposed. An artificial ant is realized as a certain feasible operating set of units and then SQP is utilized to solve the traditional ED. Finally, tests are conducted to show that the proposed approach is very effective in solving the ED problem with valve-point effect.

关 键 词：蚁群算法 序列二次规划法 经济调度 阀点效应 

分 类 号：TP392[自动化与计算机技术—计算机应用技术]