Efficient computation of spectral bounds for Hessian matrices on hyperrectangles for global optimization

Efficient computation of spectral bounds for Hessian matrices on hyperrectangles for global optimization
复制标题

高效计算超矩形上 Hessian 矩阵的谱界以实现全局优化

DOI:
10.1007/s10898-013-0099-1
复制
发表时间:
2012
影响因子:
1.8
通讯作者:
M. Mönnigmann
M. Mönnigmann
中科院分区:
数学3区
文献类型:
--
作者:
M. Schulze Darup;M. Kastsian;S. Mross;M. Mönnigmann

文献摘要

参考文献

被引文献

相似文献

通过对从基准全局优化问题中提取的1522个目标函数和约束函数进行分析,比较了两种已建立的计算超矩形上Hessian矩阵谱界的方法和一种新的计算方法。评估了谱界的紧密性和三种方法的计算工作量,这些方法适用于可以写成codelists的函数。具体来说,我们将得到的特征值界与Gershgorin圆准则的区间变分进行了比较(Adjiman et al. in computer Chem Eng 22(9): 1137-1158, 1998;格什戈林在伊兹夫。Akad。诺克SSSR,爵士。fizmat。Hertz (IEEE Trans auto Control 37:53 - 535, 1992)和Rohn’s (SIAM J Matrix Anal, 15(1): 175-184, 1994)的区间矩阵紧界方法,以及最近提出的Hessian矩阵特征值算法(Mönnigmann in SIAM J. Matrix Anal)。应用学报,32(4):1351-1366,2011),刻意避免区间Hessians的计算。在大约15%、61%和24%的示例中,特征值算法分别提供了比Gershgorin圆准则的区间变量更紧、同样紧和更不紧的边界。赫兹和罗恩的方法得到的边界总是与格什戈林圆准则的边界一样紧或更紧,在96%的情况下与特征值算法的边界一样紧或更紧。在4%的例子中,特征值算法得到的边界比Hertz和Rohn方法更紧。这个结果是令人惊讶的,因为赫兹和罗恩的方法为区间矩阵提供了严格的边界。特征值算法在这些情况下提供了更严格的边界,因为它不是基于区间矩阵的。
We compare two established and a new method for the calculation of spectral bounds for Hessian matrices on hyperrectangles by applying them to a large collection of 1,522 objective and constraint functions extracted from benchmark global optimization problems. Both the tightness of the spectral bounds and the computational effort of the three methods, which apply tofunctionsthat can be written as codelists, are assessed. Specifically, we compare eigenvalue bounds obtained with the interval variant of Gershgorin’s circle criterion (Adjiman et al. in Comput Chem Eng 22(9):1137–1158, 1998; Gershgorin in Izv. Akad. Nauk SSSR, Ser. fizmat. 6:749–754, 1931), Hertz (IEEE Trans Autom Control 37:532–535, 1992) and Rohn’s (SIAM J Matrix Anal Appl 15(1):175–184, 1994) method for tight bounds of interval matrices, and a recently proposed Hessian matrix eigenvalue arithmetic (Mönnigmann in SIAM J. Matrix Anal. Appl. 32(4): 1351–1366, 2011), which deliberately avoids the computation of interval Hessians. The eigenvalue arithmetic provides tighter, as tight, and less tight bounds than the interval variant of Gershgorin’s circle criterion in about 15, 61, and 24 % of the examples, respectively. Hertz and Rohn’s method results in bounds that are always as tight as or tighter than those from Gershgorin’s circle criterion, and as tight as or tighter than those from the eigenvalue arithmetic in 96 % of the cases. In 4 % of the examples, the eigenvalue arithmetic results in tighter bounds than Hertz and Rohn’s method. This result is surprising, since Hertz and Rohn’s method provides tight bounds for interval matrices. The eigenvalue arithmetic provides tighter bounds in these cases, since it is not based on interval matrices.
DOI: --
发表时间: 2011
影响因子: 1.5
作者:
M. Mönnigmann
通讯作者: M. Mönnigmann
DOI: --
发表时间: 2012
期刊: Optim. Methods Softw.
影响因子: --
作者:
H. Schichl;M. C. Markót
通讯作者: M. C. Markót
DOI: 10.1007/978-3-663-08773-1_12
发表时间: 1997
影响因子: 4.5
作者:
N. Tönshoff
通讯作者: N. Tönshoff
具有有效 Hessian 矩阵特征值界限的正不变性测试
DOI: --
发表时间: 2008
期刊:
影响因子: --
作者:
M. Mönnigmann
通讯作者: M. Mönnigmann
对全局优化和约束满足代码进行基准测试
DOI: --
发表时间: 2002
期刊: Global Constraint Optimization and Constraint Satisfaction
影响因子: --
作者:
O. Shcherbina;A. Neumaier;Djamila Sam;Xuan;Tuan
通讯作者: Tuan