求解机组组合问题的帝国竞争算法

提出一种用于解决机组组合问题的改进帝国竞争算法（imperialistic competition algorithm，ICA）。种群个体（国家）分为帝国和殖民地2种类型，两者又组成新的帝国，通过帝国吸收殖民地和帝国之间对殖民地的竞争实现寻优过程。在求解机组组合问题时，首先根据波峰波谷所在时刻将日负荷曲线分割成若干小时间段，然后利用ICA依次求解，与原ICA相比，避免了机组分类不合理对于优化结果的影响，同时采用整数编码的国家个体长度相对减小。最后对10~100机6个算例进行仿真分析，结果表明，算法在较少国家个体的情况下保持了较强的搜索能力，可获得较好的计算结果，是一种有效的优化算法。

Imperialistic Competition Algorithm for Unit Commitment Problem

Based on human political and social process of evolution, an improved Imperialistic Competition Algorithm of intelligent methods for unit commitment is introduced in this paper. In ICA the initial population individuals(countries) are divided into two types: imperialists and colonies. The optimal solution is obtained by moving colonies toward their imperialists and imperialistic competition. Firstly, the daily load curve is divided into small time periods according to load peak and valley. Then unit ON/OFF variables are solved by ICA successfully. Compared with original algorithm, the proposed method avoids impact to optimal result by unreasonable units’ classification and shortens individual country coded length. Six test systems with 10~100 units are simulated by proposed algorithm. The results show that proposed method keeps excellent searching ability with less country and be able to obtain good solutions.