改进的遗传算法在中期发电规划中的应用

给出了发电扩展规划的基本数学模型。为了满足日益严峻的电网调峰问题，在发电扩展规划模型中引入了峰谷差约束，然后讨论了应用于发电扩展规划的简单遗传算法的交叉和变异因子固定不变和有时进化缓慢的缺点。提出了一种改进的遗传算法。该算法能根据父代群体的进化质量动态地调整交叉和遗传因子，使进化群体始终探测新的超平面并获得新的最优个体，能克服过早收敛，也不易陷入局部最优解，从而有效地改善了算法收敛速度和最好解。最后给出了上海市2000－2007年的中期发电扩展规划的实际算例，分别用简单遗传算法和改进遗传算法进行了计算。计算结果表明，改进的遗传算法由于能自动调整遗传因子，在进化过程中能激励所有个体协作共同度过进化停滞期而能获得较优解。

APPLICATION OF IMPROVED GENETIC ALGORITHM IN MIDTERM POWER GENERATION PLANNING

A basic mathematical model for least-cost power generation expansion planning is provided firstly. To meet the increasingly austere peak load variation problem, a peak-val- ley load constraint is added in generation expansion planning model, then the deficiencies of simple genetic algorithm (SGA) are discussed. An improved genetic algorithm (IGA) is put forward, which could adjust control parameters dynam- ically and improve the best solution and convergence speed. Both the SGA and IGA approach is applied to a practical problem of midterm planning of Shanghai. The comparison of the results shows that IGA gives a better solution than SGA does.