期刊
JOURNAL OF GLOBAL OPTIMIZATION
卷 43, 期 1, 页码 1-22出版社
SPRINGER
DOI: 10.1007/s10898-008-9280-3
关键词
Order-Value Optimization; Algorithms; Convergence; Robust estimation of parameters; Hidden patterns
资金
- PRONEX-Optimization 76.79.1008-00
- FAPESP [06/53768-0, 05/56773-1, 02-14203-6]
- CNPq/FAPERJ [26/171.164/2003-APQ1]
- Fundacao de Amparo a Pesquisa do Estado de Sao Paulo (FAPESP) [06/53768-0] Funding Source: FAPESP
Given r real functions F (1)(x),...,F (r) (x) and an integer p between 1 and r, the Low Order-Value Optimization problem (LOVO) consists of minimizing the sum of the functions that take the p smaller values. If (y (1),...,y (r) ) is a vector of data and T(x, t (i) ) is the predicted value of the observation i with the parameters , it is natural to define F (i) (x) = (T(x, t (i) ) - y (i) )2 (the quadratic error in observation i under the parameters x). When p = r this LOVO problem coincides with the classical nonlinear least-squares problem. However, the interesting situation is when p is smaller than r. In that case, the solution of LOVO allows one to discard the influence of an estimated number of outliers. Thus, the LOVO problem is an interesting tool for robust estimation of parameters of nonlinear models. When p << r the LOVO problem may be used to find hidden structures in data sets. One of the most successful applications includes the Protein Alignment problem. Fully documented algorithms for this application are available at www.ime.unicamp.br/martinez/lovoalign. In this paper optimality conditions are discussed, algorithms for solving the LOVO problem are introduced and convergence theorems are proved. Finally, numerical experiments are presented.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据