电力系统负荷恢复优化的并行遗传算法实现

对电力系统的负荷恢复问题进行了研究.将该问题建模为一个多约束条件的组合优化问题,根据遗传算法特别适合求解大规模组合优化问题的特点,设计了一种粗粒度并行遗传算法来对此优化问题进行求解.在消息传递类并行软件开发环境提供的基于消息传递的并行虚拟环境下,采用master/slave的并行编程模式,有效地提高了算法的计算速度.将各种约束条件与目标函数融合在一起,建立一种序关系,来处理负荷恢复中的约束条件.求解过程满足系统的约束条件,不会出现系统的越限.算例结果表明,所提出的并行遗传算法不仅可以最大限度地恢复负荷,而且可有效提高算法的计算速度.

Implementation of Load Restoration Optimization for Power System by Parallel Genetic Algorithm

In this paper,the problem of the load restoration was studied and it was modeled as a combinational optimization problem with many constraints.Then,according to the high efficiency of genetic algorithm for solving large-scale combinational optimization problems,a coarse-grain parallel genetic algorithm is presented.In the parallel virtual environment based on message passing,the calculation can be efficiently speeded up by using the master/slave mode of parallel programming.Moreover,by combining the constraints with the objective functions,an order relation is constructed to deal with the constraints in load restoration.As the constraints of load restoration cannot be violated in the solving process,the power system security can be effectively ensured.Simulated results show that the proposed algorithm can effectively speed up the calculation and restart the load as much as possible.