基于混合概率分布的演化算法在电网负荷恢复中的应用

APPLYING HYBRID PROBABILITY DISTRIBUTION-BASED EVOLUTIONARY ALGORITHM IN LOAD RESTORATION OF POWER SYSTEM

在电网恢复过程中最后的负荷恢复阶段,寻求全局优化的恢复方案通常都是NP完全问题。在综合考虑负荷恢复阶段中各种不同因素的基础上,建立了基于多目标多约束组合优化的数学仿真模型,并将参数可变的混合概率分布演化算法(EABHPD)引入到模型的求解。该算法通过改变分布函数的参数实现了负荷恢复问题求解精度与计算复杂度的折衷,并且避免了演化算法在搜索遍历过程中陷入局部最优解而过早收敛的问题。仿真实验结果表明算法具有良好性能。

As the last stage of load restoration in power restoration process,to search global optimised restoration scheme is usually an NP-hard problem.On the basis of comprehensively considering various factors in the stage of load restoration,a combinatorial optimised mathematical simulation model based on multi-target and multi-constraint is set up,and the evolutionary algorithm based on hybrid probability distribution（EABHPD） with variable parameters is introduced to the solution of the model.The algorithm achieves the trade-off between the desired precision of the solution and the computation cost in resolving the load restoration problem by altering the parameters of distributed function,and avoids the premature convergence problem the common evolutionary algorithms have caused by falling into local optimum during the course of ergodic search.The good performance of this algorithm is manifested by the results of simulation experiment.