An efficient active-set approach is presented for both nonnegative and general linear programming by adding varying numbers of constraints at each iteration. Computational experiments demonstrate that the proposed app...An efficient active-set approach is presented for both nonnegative and general linear programming by adding varying numbers of constraints at each iteration. Computational experiments demonstrate that the proposed approach is significantly faster than previous active-set and standard linear programming algorithms.展开更多
Posterior constraint optimal selection techniques (COSTs) are developed for nonnegative linear programming problems (NNLPs), and a geometric interpretation is provided. The posterior approach is used in both a dynamic...Posterior constraint optimal selection techniques (COSTs) are developed for nonnegative linear programming problems (NNLPs), and a geometric interpretation is provided. The posterior approach is used in both a dynamic and non-dynamic active-set framework. The computational performance of these methods is compared with the CPLEX standard linear programming algorithms, with two most-violated constraint approaches, and with previously developed COST algorithms for large-scale problems.展开更多
We describe a new active-set, cutting-plane Constraint Optimal Selection Technique (COST) for solving general linear programming problems. We describe strategies to bound the initial problem and simultaneously add mul...We describe a new active-set, cutting-plane Constraint Optimal Selection Technique (COST) for solving general linear programming problems. We describe strategies to bound the initial problem and simultaneously add multiple constraints. We give an interpretation of the new COST’s selection rule, which considers both the depth of constraints as well as their angles from the objective function. We provide computational comparisons of the COST with existing linear programming algorithms, including other COSTs in the literature, for some large-scale problems. Finally, we discuss conclusions and future research.展开更多
In this paper we propose a new model for segmentation of an image under some geometrical constraints in order to detect special regions of interest.Our work is based on the recent work by Gout et al.[Numer.Algorithms,...In this paper we propose a new model for segmentation of an image under some geometrical constraints in order to detect special regions of interest.Our work is based on the recent work by Gout et al.[Numer.Algorithms,39(2005),pp.155-173 and 48(2008),pp.105-133]using geodesic active contours models,by combining it with the idea of a piecewise constant Mumford-Shah model as with the non-selective Chan-Vese segmentation.Numerical tests show that our method is more robust than the previous works.展开更多
颗粒粒度反演需要求解第一类Fredholm积分方程,此问题是动态光散射中的难点之一,其中,双峰颗粒的反演更是亟待解决的问题.为保证反演结果的非负性,采用了trust region reflective Newton和active set算法实现的非负Tikhonov,非负TSVD算...颗粒粒度反演需要求解第一类Fredholm积分方程,此问题是动态光散射中的难点之一,其中,双峰颗粒的反演更是亟待解决的问题.为保证反演结果的非负性,采用了trust region reflective Newton和active set算法实现的非负Tikhonov,非负TSVD算法对双峰颗粒数据进行了反演.结果表明采用前者实现的非负Tikhonov和非负TSVD不能区别间隔粒径较近双峰,而采用后者实现的非负Tikhonov和非负TSVD能区别出.展开更多
窄带法是水平集图像分割的一种常见的加速方法.传统窄带仍然存在冗余的计算区域;传统窄带法与LATE (Local Approximation of Taylor Expansion)水平集模型结合时,图像分割效率反而可能下降.针对这些问题,本文提出了一种基于LATE水平集...窄带法是水平集图像分割的一种常见的加速方法.传统窄带仍然存在冗余的计算区域;传统窄带法与LATE (Local Approximation of Taylor Expansion)水平集模型结合时,图像分割效率反而可能下降.针对这些问题,本文提出了一种基于LATE水平集图像分割模型的矩形窄带法.在每次LATE水平集迭代之前,对水平集做如下窄带处理.首先找出水平集的所有过零点;然后对过零点做活动约束,剔除不活动的过零点,有效缩小窄带范围;再对活动约束的过零点生成矩形窄带;对重叠的矩形窄带进行合并优化,使得矩形窄带总面积尽可能小.最后,在矩形窄带范围内求解水平集微分方程,更新水平集,完成本次迭代.在水平集演化的不同阶段,对传统窄带法的窄带面积与本文矩形窄带面积进行了比较.随着迭代次数增加,矩形窄带面积与传统窄带法的窄带面积之比逐渐减小到0,说明矩形窄带法有效地减少了冗余计算量.针对不同程度的灰度不均匀图像,本文方法与LATE方法、结合LATE模型的直接窄带法、以及结合LATE模型的DTM窄带法进行了比较.直接窄带法和DTM窄带法的分割速度反而慢于LATE方法.对灰度严重不均匀的图像,直接窄带法和DTM窄带法的分割质量受到了较大影响.本文方法在保持较好分割效果的条件下,分割速度快于LATE方法.本文的矩形窄带方法有效地降低了算法复杂度,提高了图像分割效率.展开更多
文摘An efficient active-set approach is presented for both nonnegative and general linear programming by adding varying numbers of constraints at each iteration. Computational experiments demonstrate that the proposed approach is significantly faster than previous active-set and standard linear programming algorithms.
文摘Posterior constraint optimal selection techniques (COSTs) are developed for nonnegative linear programming problems (NNLPs), and a geometric interpretation is provided. The posterior approach is used in both a dynamic and non-dynamic active-set framework. The computational performance of these methods is compared with the CPLEX standard linear programming algorithms, with two most-violated constraint approaches, and with previously developed COST algorithms for large-scale problems.
文摘We describe a new active-set, cutting-plane Constraint Optimal Selection Technique (COST) for solving general linear programming problems. We describe strategies to bound the initial problem and simultaneously add multiple constraints. We give an interpretation of the new COST’s selection rule, which considers both the depth of constraints as well as their angles from the objective function. We provide computational comparisons of the COST with existing linear programming algorithms, including other COSTs in the literature, for some large-scale problems. Finally, we discuss conclusions and future research.
文摘In this paper we propose a new model for segmentation of an image under some geometrical constraints in order to detect special regions of interest.Our work is based on the recent work by Gout et al.[Numer.Algorithms,39(2005),pp.155-173 and 48(2008),pp.105-133]using geodesic active contours models,by combining it with the idea of a piecewise constant Mumford-Shah model as with the non-selective Chan-Vese segmentation.Numerical tests show that our method is more robust than the previous works.
文摘颗粒粒度反演需要求解第一类Fredholm积分方程,此问题是动态光散射中的难点之一,其中,双峰颗粒的反演更是亟待解决的问题.为保证反演结果的非负性,采用了trust region reflective Newton和active set算法实现的非负Tikhonov,非负TSVD算法对双峰颗粒数据进行了反演.结果表明采用前者实现的非负Tikhonov和非负TSVD不能区别间隔粒径较近双峰,而采用后者实现的非负Tikhonov和非负TSVD能区别出.
文摘窄带法是水平集图像分割的一种常见的加速方法.传统窄带仍然存在冗余的计算区域;传统窄带法与LATE (Local Approximation of Taylor Expansion)水平集模型结合时,图像分割效率反而可能下降.针对这些问题,本文提出了一种基于LATE水平集图像分割模型的矩形窄带法.在每次LATE水平集迭代之前,对水平集做如下窄带处理.首先找出水平集的所有过零点;然后对过零点做活动约束,剔除不活动的过零点,有效缩小窄带范围;再对活动约束的过零点生成矩形窄带;对重叠的矩形窄带进行合并优化,使得矩形窄带总面积尽可能小.最后,在矩形窄带范围内求解水平集微分方程,更新水平集,完成本次迭代.在水平集演化的不同阶段,对传统窄带法的窄带面积与本文矩形窄带面积进行了比较.随着迭代次数增加,矩形窄带面积与传统窄带法的窄带面积之比逐渐减小到0,说明矩形窄带法有效地减少了冗余计算量.针对不同程度的灰度不均匀图像,本文方法与LATE方法、结合LATE模型的直接窄带法、以及结合LATE模型的DTM窄带法进行了比较.直接窄带法和DTM窄带法的分割速度反而慢于LATE方法.对灰度严重不均匀的图像,直接窄带法和DTM窄带法的分割质量受到了较大影响.本文方法在保持较好分割效果的条件下,分割速度快于LATE方法.本文的矩形窄带方法有效地降低了算法复杂度,提高了图像分割效率.
