求解离散无功优化问题的精确连续化方法

提出了一种求解离散无功优化问题的新算法。该方法对离散变量进行二进制编码,从而将离散变量转化为若干0-1变量的线性组合,并将二进制变量约束转化为等价的互补约束,再利用非线性互补函数将互补约束转化为等价的非光滑方程,对其进行光滑化处理后,将原问题转化成可微的非线性规划问题,并采用非线性原对偶内点算法求解。此外,还对离散变量编码的原理以及应用原对偶内点法中遇到的求导难点进行了介绍,并通过几个不同规模试验系统的计算分析,验证该方法能有效处理离散变量,而且具有良好的收敛性和精确性。

Accurate Serialization Method for Solving Discrete Reactive Power Optimization

This paper puts forward a new algorithm for solving discrete reactive power optimization. This algorithm trans- forms discrete variables into several linear combination of 0-1 variables by using binary coding, changes binary variable con- straints into equivalent complementary constraints and transforms these complementary constraints into equivalent non- smooth equation by using nonlinear complementary function. After dealing with the equation, this algorithm changes the primal problem into differentiable nonlinear programming problem and solves it by using nonlinear primal-dual interior point method. Furthermore, it introduces principles of discrete variable coding and derivation difficulties in application of primal： dual interior point method. By calculation and analysis of several test systems, it proves that this method can effectively treat discrete variables and is provided with good convergence and accuracy.