基于约束松弛变量策略的中心校正内点法

Centrality correction interior point method based on constrained slack variables strategy

不等式约束互补对之间的尺寸比例是影响内点法收敛性能的一个重要因素。为避免互补对差异过大,提出一种基于约束松弛变量策略的中心校正法。该法通过投影的方法来限制松弛变量的取值范围,彻底解决了因互补对比例相差太大而造成的数值问题。将所提方法应用到考虑静态电压稳定的有功出力方式优化中,优化结果表明即使在当前不等式约束变量远离其上、下边界的条件下,所提方法也能可靠收敛,极大地提高了内点法的实用性。

The dimension scale of complementary pairs is one of important factors affecting the convergence of interior point method. In order to avoid large discrepancy among complementary pairs, a centrality corrections method based on constrained slack variables strategy is proposed. In the proposed method, the maximum of slack variables is curtailed by projection method, by which the numerical problem caused by large discrepancy in complementary pairs is settled fully. The proposed method is applied to generation rescheduling considering static voltage stability and the optimization results indicate that the proposed method has a sound convergence performance even when the current inequality constraint variables are far away from their limits, enhancing the practicability of the interior point algorithm greatly.