基于自适应免疫算法和预测-校正内点法的无功优化

针对电力系统无功优化问题，将自适应免疫算法（adaptive immtme algorithm，AIA）和预测-校正内点法相结合，提出了一种新的混合优化算法．先利用AIA进行大范围全局寻优，找到候选最优点，把它作为内点法的初始可行点，再通过预测-校正内点法在初始可行点的邻域内进行局部的确定性搜索，提高解的精度和速度；在此基础上，根据对偶间隙的变化过程，提出了对中心参数及相应障碍参数的改进选择方法，有效地避免了数值振荡，使计算精度及收敛速度均得到明显改善．将上述方法用于IEEE14节点系统，计算时间为2．0s，优化后网损下降2．27％；而用于IEEE118节点系统，计算时间为322s，优化后网损下降14．29％．这表明本文所提出的算法在计算速度和精度上较其他方法均有明显改进．

Reactive Power Optimization Based on the Adaptive Immune Algorithm and Predictor-Corrector Interior Point Method

A new method, which combines the adaptive immune algorithm （AIA） with predictor-corrector interior point method, was presented for the reactive power optimization of power system. First, the candidate optimal points, which were taken as the initial points of interior point method, were found by use of the global optimization capability of AIA. Then in the neighborhood of candidate optimal points local optimal search was performed by the predictor-corrector interior point method to further improve the accuracy of the solution. Moreover, according to the variation of the duality gap during the iterating process, a new selecting method for central parameter and relevant obstacle parameter was proposed to avoid the numerical vibration and improve the convergence rate and accuracy. Finally, the new method was applied to IEEE14 system and IEEE118 system. For IEEE14 system, the calculation time was 2.0 s and minimum power loss was reduced by 2.27% ; for IEEE118 system, the calculation time was 322 s and the minimum power loss was reduced by 14.29%. All the results demonstrate that the new method effectively improves the speed and accuracy of calculation compared with other methods.