凸二次规划问题基于核函数的全牛顿步内点算法

A Full-Newton Step Interior-point Algorithm Based on a Kernel Function for Convex Quadratic Optimization

针对凸二次规划问题,构造了新的核函数。通过构造的核函数来确定搜索方向和逼近度量,接着给出了求解凸二次规划问题的全牛顿步内点算法,最后给出了算法的复杂性界。

For solving convex quadratic optimization, we introduce a new kernel function.Through the new kernel function, we derive a new search direction and proximity measure, then present a full-Netwon step interior-point algorithm. Finally, we prove that the complexity bound of the algorithm.