摘要
BFGS算法是解无约束优化问题的公认的最有效的算法之一。针对BFGS算法的Hesse矩阵修正保持正定的特点,利用Chlolesky分解,对于算法中矩阵修正及确定相应的搜索方向的实现作了一定的分析和探讨,并导出了相应的计算公式,使得计算量下降了一个数量级,并且尽可能地保证了修正矩阵的正定性。
BFGS algorithm is one of the most effective methods in solving the non-constrained optimization problems.Based on te positive-definite property of the Hesse matrix in the BFGS algorithm and the Chlolesky decomposition,this paper gives an implementation of the Hesse iteration.It also presents a corresponding formula, which decreases the computing amount by one degree and keeps the iteration matrix be positive-definite.
出处
《南京邮电学院学报（自然科学版）》
2004年第4期59-61,共3页
Journal of Nanjing University of Posts and Telecommunications
