In this paper,a new trust region algorithm for unconstrained LC^1 optimization problems is given.Compare with those existing trust regiion methods,this algorithm has a different feature:it obtains a stepsize at each i...In this paper,a new trust region algorithm for unconstrained LC^1 optimization problems is given.Compare with those existing trust regiion methods,this algorithm has a different feature:it obtains a stepsize at each iteration not by soloving a quadratic subproblem with a trust region bound,but by solving a system of linear equations.Thus it reduces computational complexity and improves computation efficlency,It is proven that this algorithm is globally convergent and locally superlinear under some conditions.展开更多
In this paper, we present a new line search and trust region algorithm for unconstrained optimization problems. The trust region center locates at somewhere in the negative gradient direction with the current best ite...In this paper, we present a new line search and trust region algorithm for unconstrained optimization problems. The trust region center locates at somewhere in the negative gradient direction with the current best iterative point being on the boundary. By doing these, the trust region subproblems are constructed at a new way different with the traditional ones. Then, we test the efficiency of the new line search and trust region algorithm on some standard benchmarking. The computational results reveal that, for most test problems, the number of function and gradient calculations are reduced significantly.展开更多
A trust region algorithm for equality constrained optimization is given in this paper.The algorithm does not enforce strict monotonicity of the merit function for every iteration.Global convergence of the algorithm is...A trust region algorithm for equality constrained optimization is given in this paper.The algorithm does not enforce strict monotonicity of the merit function for every iteration.Global convergence of the algorithm is proved under the same conditions of usual trust region method.展开更多
A trust region algorithm for equality constrained optimization is proposed, which is a nonmonotone one in a certain sense. The augmented Lagrangian function is used as a merit function. Under certain conditions, the g...A trust region algorithm for equality constrained optimization is proposed, which is a nonmonotone one in a certain sense. The augmented Lagrangian function is used as a merit function. Under certain conditions, the global convergence theorems of the algorithm are proved.展开更多
The image restoration problems play an important role in remote sensing and astronomical image analysis. One common method for the recovery of a true image from corrupted or blurred image is the least squares error (L...The image restoration problems play an important role in remote sensing and astronomical image analysis. One common method for the recovery of a true image from corrupted or blurred image is the least squares error (LSE) method. But the LSE method is unstable in practical applications. A popular way to overcome instability is the Tikhonov regularization. However, difficulties will encounter when adjusting the so-called regularization parameter α. Moreover, how to truncate the iteration at appropriate steps is also challenging. In this paper we use the trust region method to deal with the image restoration problem, meanwhile, the trust region subproblem is solved by the truncated Lanczos method and the preconditioned truncated Lanczos method. We also develop a fast algorithm for evaluating the Kronecker matrix-vector product when the matrix is banded. The trust region method is very stable and robust, and it has the nice property of updating the trust region automatically. This releases us from tedious finding the regularization parameters and truncation levels. Some numerical tests on remotely sensed images are given to show that the trust region method is promising.展开更多
A class of nonmonotone trust region algorithms is presented for unconstrained optimizations. Under suitable conditions, the global and Q quadratic convergences of the algorithm are proved. Several rules of choosing tr...A class of nonmonotone trust region algorithms is presented for unconstrained optimizations. Under suitable conditions, the global and Q quadratic convergences of the algorithm are proved. Several rules of choosing trial steps and trust region radii are also discussed.展开更多
In this paperwe present a nonmonotone trust region algorthm for general nonlinear constrained optimization problems.The main idea of this paper is to combine Yuan''''s technique[1]with a nonmonotone me...In this paperwe present a nonmonotone trust region algorthm for general nonlinear constrained optimization problems.The main idea of this paper is to combine Yuan''''s technique[1]with a nonmonotone method similar to Ke and Han[2].This new algorithm may not only keep the robust properties of the algorithm given by Yuan,but also have some advantages led by the nonmonotone technique.Under very mild conditions,global convergence for the algorithm is given.Numerical experiments demostrate the efficency of the algorithm.展开更多
In this note, we consider the following constrained optimization problem (COP) min f(x), x∈Ωwhere f(x): R^n→R is a continuously differentiable function on a closed convex set Ω. Forthe constrained optimization pro...In this note, we consider the following constrained optimization problem (COP) min f(x), x∈Ωwhere f(x): R^n→R is a continuously differentiable function on a closed convex set Ω. Forthe constrained optimization problem (COP), a class of nonmonotone trust region algorithmsis proposed in sec. 1. In sec. 2, the global convergence of this class of algorithms isproved. In sec. 3, some results about the Cauchy point are provided. The展开更多
In this note, the following unconstrained nonsmooth optimization problem is considered where f(x):R^n→R is only a locally Lipschitzian function. Many papers appear on the convergence properties of the trust region al...In this note, the following unconstrained nonsmooth optimization problem is considered where f(x):R^n→R is only a locally Lipschitzian function. Many papers appear on the convergence properties of the trust region algorithm to solve several different particular nonsmooth problems. Dennis, Li and Tapia proposed a general trust region model by using regular functions. They proved the global convergence of the general trust region model under some mild conditions which are shown to be satisfied by many trust region algorithms including smooth one. Qi and Sun provided another trust region model展开更多
The trust region approach has been extended to solving nonlinear constrained optimization. Most of these extensions consider only equality constraints and require strong global regularity assumptions. In this paper, a...The trust region approach has been extended to solving nonlinear constrained optimization. Most of these extensions consider only equality constraints and require strong global regularity assumptions. In this paper, a trust region algorithm for solving general nonlinear programming is presented, which solves an unconstrained piecewise quadratic trust region subproblem and a quadratic programming trust region subproblem at each iteration. A new technique for updating the penalty parameter is introduced. Under very mild conditions, the global convergence results are proved. Some local convergence results are also proved. Preliminary numerical results are also reported.展开更多
In this paper, we present a new trust region algorithm for a nonlinear bilevel programming problem by solving a series of its linear or quadratic approximation subproblems. For the nonlinear bilevel programming proble...In this paper, we present a new trust region algorithm for a nonlinear bilevel programming problem by solving a series of its linear or quadratic approximation subproblems. For the nonlinear bilevel programming problem in which the lower level programming problem is a strongly convex programming problem with linear constraints, we show that each accumulation point of the iterative sequence produced by this algorithm is a stationary point of the bilevel programming problem.展开更多
A trust region algorithm is proposed for solving bilevel programming problems where the lower level programming problem is a strongly convex programming problem with linear constraints. This algorithm is based on a tr...A trust region algorithm is proposed for solving bilevel programming problems where the lower level programming problem is a strongly convex programming problem with linear constraints. This algorithm is based on a trust region algorithm for nonsmooth unconstrained optimization problems, and its global convergence is also proved.展开更多
AbstractAn interior trust-region-based algorithm for linearly constrained minimization problems is proposed and analyzed. This algorithm is similar to trust region algorithms for unconstrained minimization: a trust re...AbstractAn interior trust-region-based algorithm for linearly constrained minimization problems is proposed and analyzed. This algorithm is similar to trust region algorithms for unconstrained minimization: a trust region subproblem on a subspace is solved in each iteration. We establish that the proposed algorithm has convergence properties analogous to those of the trust region algorithms for unconstrained minimization. Namely, every limit point of the generated sequence satisfies the Krush-Kuhn-Tucker (KKT) conditions and at least one limit point satisfies second order necessary optimality conditions. In addition, if one limit point is a strong local minimizer and the Hessian is Lipschitz continuous in a neighborhood of that point, then the generated sequence converges globally to that point in the rate of at least 2-step quadratic. We are mainly concerned with the theoretical properties of the algorithm in this paper. Implementation issues and adaptation to large-scale problems will be addressed in展开更多
In this paper we present a filter-trust-region algorithm for solving LC1 uncon- strained optimization problems which uses the second Dini upper directional derivative. We establish the global convergence of the algori...In this paper we present a filter-trust-region algorithm for solving LC1 uncon- strained optimization problems which uses the second Dini upper directional derivative. We establish the global convergence of the algorithm under reasonable assumptions.展开更多
文摘In this paper,a new trust region algorithm for unconstrained LC^1 optimization problems is given.Compare with those existing trust regiion methods,this algorithm has a different feature:it obtains a stepsize at each iteration not by soloving a quadratic subproblem with a trust region bound,but by solving a system of linear equations.Thus it reduces computational complexity and improves computation efficlency,It is proven that this algorithm is globally convergent and locally superlinear under some conditions.
文摘In this paper, we present a new line search and trust region algorithm for unconstrained optimization problems. The trust region center locates at somewhere in the negative gradient direction with the current best iterative point being on the boundary. By doing these, the trust region subproblems are constructed at a new way different with the traditional ones. Then, we test the efficiency of the new line search and trust region algorithm on some standard benchmarking. The computational results reveal that, for most test problems, the number of function and gradient calculations are reduced significantly.
文摘A trust region algorithm for equality constrained optimization is given in this paper.The algorithm does not enforce strict monotonicity of the merit function for every iteration.Global convergence of the algorithm is proved under the same conditions of usual trust region method.
基金Project supported by the National Natural Science Foundation of China and Postdoctoral Foundation of China.
文摘A trust region algorithm for equality constrained optimization is proposed, which is a nonmonotone one in a certain sense. The augmented Lagrangian function is used as a merit function. Under certain conditions, the global convergence theorems of the algorithm are proved.
基金supported by the National Natural Science Foundation of China(Grant Nos.19731010 and 10231060)the Knowledge Innovation Program of CAS+1 种基金was supported by SRF for ROSS,SEM partially supported by the Special Innovation Fund for graduate students of CAS.
文摘The image restoration problems play an important role in remote sensing and astronomical image analysis. One common method for the recovery of a true image from corrupted or blurred image is the least squares error (LSE) method. But the LSE method is unstable in practical applications. A popular way to overcome instability is the Tikhonov regularization. However, difficulties will encounter when adjusting the so-called regularization parameter α. Moreover, how to truncate the iteration at appropriate steps is also challenging. In this paper we use the trust region method to deal with the image restoration problem, meanwhile, the trust region subproblem is solved by the truncated Lanczos method and the preconditioned truncated Lanczos method. We also develop a fast algorithm for evaluating the Kronecker matrix-vector product when the matrix is banded. The trust region method is very stable and robust, and it has the nice property of updating the trust region automatically. This releases us from tedious finding the regularization parameters and truncation levels. Some numerical tests on remotely sensed images are given to show that the trust region method is promising.
文摘A class of nonmonotone trust region algorithms is presented for unconstrained optimizations. Under suitable conditions, the global and Q quadratic convergences of the algorithm are proved. Several rules of choosing trial steps and trust region radii are also discussed.
基金This work was done when the author was studying in the State Key Laboratory of Scientific and Engi- neering Computing, Institute of Computational Mathematics and Scientific/Engineering Computing, Chinese Academy of Sciences, P. O. Box 2719, Beijing 10008
文摘In this paperwe present a nonmonotone trust region algorthm for general nonlinear constrained optimization problems.The main idea of this paper is to combine Yuan''''s technique[1]with a nonmonotone method similar to Ke and Han[2].This new algorithm may not only keep the robust properties of the algorithm given by Yuan,but also have some advantages led by the nonmonotone technique.Under very mild conditions,global convergence for the algorithm is given.Numerical experiments demostrate the efficency of the algorithm.
基金Project supported by the National Natural Science Foundation of China and Postdoctoral Foundation of China.
文摘In this note, we consider the following constrained optimization problem (COP) min f(x), x∈Ωwhere f(x): R^n→R is a continuously differentiable function on a closed convex set Ω. Forthe constrained optimization problem (COP), a class of nonmonotone trust region algorithmsis proposed in sec. 1. In sec. 2, the global convergence of this class of algorithms isproved. In sec. 3, some results about the Cauchy point are provided. The
文摘In this note, the following unconstrained nonsmooth optimization problem is considered where f(x):R^n→R is only a locally Lipschitzian function. Many papers appear on the convergence properties of the trust region algorithm to solve several different particular nonsmooth problems. Dennis, Li and Tapia proposed a general trust region model by using regular functions. They proved the global convergence of the general trust region model under some mild conditions which are shown to be satisfied by many trust region algorithms including smooth one. Qi and Sun provided another trust region model
基金Chinese NSF grants 19525101, 19731001, and by State key project 96-221-04-02-02. It is also partially supported by Hebei provi
文摘The trust region approach has been extended to solving nonlinear constrained optimization. Most of these extensions consider only equality constraints and require strong global regularity assumptions. In this paper, a trust region algorithm for solving general nonlinear programming is presented, which solves an unconstrained piecewise quadratic trust region subproblem and a quadratic programming trust region subproblem at each iteration. A new technique for updating the penalty parameter is introduced. Under very mild conditions, the global convergence results are proved. Some local convergence results are also proved. Preliminary numerical results are also reported.
基金Supported by the National Natural Science Foundation of China(No.11171348,11171252 and 71232011)
文摘In this paper, we present a new trust region algorithm for a nonlinear bilevel programming problem by solving a series of its linear or quadratic approximation subproblems. For the nonlinear bilevel programming problem in which the lower level programming problem is a strongly convex programming problem with linear constraints, we show that each accumulation point of the iterative sequence produced by this algorithm is a stationary point of the bilevel programming problem.
文摘A trust region algorithm is proposed for solving bilevel programming problems where the lower level programming problem is a strongly convex programming problem with linear constraints. This algorithm is based on a trust region algorithm for nonsmooth unconstrained optimization problems, and its global convergence is also proved.
基金Research partially supported by the Faculty Research Grant RIG-35547 and ROG-34628 of the University of North Texas and in part by the Cornell Theory Center which receives major funding from the National Science Foundation and IBM Corporation with ad
文摘AbstractAn interior trust-region-based algorithm for linearly constrained minimization problems is proposed and analyzed. This algorithm is similar to trust region algorithms for unconstrained minimization: a trust region subproblem on a subspace is solved in each iteration. We establish that the proposed algorithm has convergence properties analogous to those of the trust region algorithms for unconstrained minimization. Namely, every limit point of the generated sequence satisfies the Krush-Kuhn-Tucker (KKT) conditions and at least one limit point satisfies second order necessary optimality conditions. In addition, if one limit point is a strong local minimizer and the Hessian is Lipschitz continuous in a neighborhood of that point, then the generated sequence converges globally to that point in the rate of at least 2-step quadratic. We are mainly concerned with the theoretical properties of the algorithm in this paper. Implementation issues and adaptation to large-scale problems will be addressed in
基金Supported by CERG: CityU 101005 of the Government of Hong Kong SAR, Chinathe National Natural ScienceFoundation of China, the Specialized Research Fund of Doctoral Program of Higher Education of China (Grant No.20040319003)the Natural Science Fund of Jiangsu Province of China (Grant No. BK2006214)
文摘In this paper we present a filter-trust-region algorithm for solving LC1 uncon- strained optimization problems which uses the second Dini upper directional derivative. We establish the global convergence of the algorithm under reasonable assumptions.
基金Supported by the National High Technology Research and Development Programme of China( No. 2006AA04Z245 ) and China Postdoctoral Science Foundation ( No. 200904500988 ).