基于免疫算法的机组组合优化方法

机组组合是改善传统电力系统运行经济性和电力市场出清的重要手段。基于群体进化的智能优化算法存求解过程中存在计算效率低和易于早熟收敛等缺点。提出机组组合的免疫算法，利用免疫算法保持种群多样性的内在机制和免疫记忆特性改进既有的智能优化方法。新算法扩展了约束处理技术，能更好地对可行解空间搜索，采用一种由后向前、由前及后、双向迂回推进的精简程序改善个体可行解的局部最优性，同时利用优先级顺序法产生能较好反映问题先验知识的初始种群。典型算例证实新算法能获得更优的结果，具有更快的收敛速度，且在系统规模扩大时有大致线性的计算复杂性，是一种新的高效的机组组合智能优化算法。

Immune algorithm based unit commitment optimization method

Unit commitment is an important means to improve operational economic efficiency in traditional power systems and a critical market clearing tool of electricity market. Colony evolution based intelligent optimization algorithms have deficiencies of poor computational efficiency and are prone to premature during the solution process. An immune algorithm based unit commitment was presented and the aforementioned faults were chastened effectively by the inherent mechanism of keeping diversity among populations and immune memory of the immune algorithm. Furthermore, several important measures were adopted to achieve better performance. Firstly, the previous constraint satisfaction technique was generalized to obtain more solutions to be explored in the feasible solution space. Secondly, a bi-directional unit decommitment procedure was developed to improve local optimization performance of the individuals. Finally, initial population was generated by the priority list method, which was suitable to reflect prior knowledge about the problem. A typical system shows that the newly developed algorithm obtains better results and faster convergence speed, requires approximately linear computational complexity when the system is scaled up, and is qualified to be a new effective algorithm for solving unit commitment problem.