实数编码的遗传算法在厂级负荷优化分配中的应用(英文)

Application of Real-code Genetic Algorithm to Economic Load Dispatch in Power Plants

提出了一种改进的实数编码遗传算法来解决厂级负荷优化问题。选取了简单且符合电厂实际的数学模型和机组煤耗特性的拟合方法。在遗传操作上,突破了传统的选择一交义—变异模式,采用了融入交叉和变异的双个体组进化来进行选择操作,以及固定替代的变异操作。针对负荷优化问题高维、约束条件多、对实时性要求高的特点,文中提出了对仞始群体的随机产生过程加入边界约束的方案。迭代过程结束后,在小邻域内用枚举法搜索更优的值,进一步提高了算法的准确性。仿真结果证明.改进后的遗传算法简单、高效、全局搜索能力强,具有较高的实用价值。

An improved real-code genetic algorithm was presented to solve economic load dispatch in power plants. The simple and factual mathematical model and the fitting method of coal consumption curve were utilized. The improved genetic algorithm, which blends crossover and mutation method in the two-individual team selection along with fixed substitute mutation, breaks through the traditional selection-crossovermutation mode in genetic operations. Considering the high dimension, multiple constraints and real time specialty in the economic load dispatch, a scheme to restrict random process in creating initial population has been put forward. The precision can be improved further by searching better results in a small range after the interacting process of genetic algorithm. Simulation has shown that the improved genetic algorithm, which has the characteristics like simple, efficient and strong ability in global search, has enormous value in application.