泵站优化调度中考虑一次性开机约束的改进遗传算法

泵站优化调度中常规的遗传算法是以每台机组各时段的运行状态作为优化变量,其变量数多,难以收敛。针对泵站中一次性开机的约束条件,提出了以运行起始时刻和运行结束时刻作为优化变量的优化模型,该模型变量少且自然满足一次性引水约束条件,简化了约束处理。针对泵站同型号机组在优化过程中出现的振荡,提出了采用实数编码的带优先级的处理方法,并采用多点交叉、变异处理,提高了收敛速度。

Improved genetic algorithm for one-time start up constraint considered for optimal operation of pumping station

During the optimal operation of pumping station,the operation condition during each operation period of each pump unit is normally taken as the variable of the optimization for the conventional genetic algorithm;however,the variables are too many to be converged for the optimization.Based on the constraints of the one-time start up of pumping station,a optimal model established by taking both the start up time and the shut down time as the optimized variables is presented herein;by which the modeling variables are less and the constraint conditions of the one-time water pumping can be satisfied,and then the constraint processing of the variables are simplified as well.Furthermore,the processing method using the decimal encoding with priorities is put forward for avoiding the oscillations from the pumps of the same type;in which the optimization time is reduced with the multi-point crossover and mutation processing.