不等式约束优化一个基于滤子思想的广义梯度投影算法

讨论非线性不等式约束优化问题, 借鉴于滤子算法思想,提出了一个新型广义梯度投影算法.该方法既不使用罚函数又无真正意义下的滤子.每次迭代通过一个简单的显式广义投影法产生搜索方向,步长由目标函数值或者约束违反度函数值充分下降的Armijo型线搜索产生.算法的主要特点是: 不需要迭代序列的有界性假设;不需要传统滤子算法所必需的可行恢复阶段;使用了ε积极约束集减小计算量.在合适的假设条件下算法具有全局收敛性, 最后对算法进行了初步的数值实验.     

一个新的无约束优化超记忆梯度算法

本文提出一种新的无约束优化超记忆梯度算法,算法利用当前点的负梯度和前一点的负梯度的线性组合为搜索方向,以精确线性搜索和Armijo搜索确定步长．在很弱的条件下证明了算法具有全局收敛性和线性收敛速度．因算法中避免了存贮和计算与目标函数相关的矩阵,故适于求解大型无约束优化问题．数值实验表明算法比一般的共轭梯度算法有效．     

一类新的多步曲线搜索下的超记忆梯度法

研究一类新的求解无约束优化问题的超记忆梯度法,分析了算法的全局收敛性和线性收敛速率.算法利用一种多步曲线搜索准则产生新的迭代点,在每步迭代时同时确定下降方向和步长,并且不用计算和存储矩阵,适于求解大规模优化问题.数值试验表明算法是有效的.     

Parallel SSLE algorithm for large scale constrained optimization

In this paper, a parallel SSLE algorithm is proposed for solving large scale constrained optimization with block-separable structure. At each iteration, the PVD sub-problems are solved inexactly by the SSLE algorithm, which successfully overcomes the constraint inconsistency exited in most SQP-type algorithm, and decreases the computation amount as well. Without assuming the convexity of the constraints, the algorithm is proved to be globally convergent to a KKT point of the original problem.     

New spectral PRP conjugate gradient method for unconstrained optimization

In this paper, a new spectral PRP conjugate gradient algorithm has been developed for solving unconstrained optimization problems, where the search direction was a kind of combination of the gradient and the obtained direction, and the steplength was obtained by the Wolfe-type inexact line search. It was proved that the search direction at each iteration is a descent direction of objective function. Under mild conditions, we have established the global convergence theorem of the proposed method. Numerical results showed that the algorithm is promising, particularly, compared with the existing several main methods.     

A superlinearly convergent method of quasi-strongly sub-feasible directions with active set identifying for constrained optimization

Combining the norm-relaxed sequential quadratic programming (SQP) method and the idea of method of quasi-strongly sub-feasible directions (MQSSFD) with active set identification technique, a new SQP algorithm for solving nonlinear inequality constrained optimization is proposed. Unlike the previous work, at each iteration of the proposed algorithm, the norm-relaxed quadratic programming (QP) subproblem only consists of the constraints corresponding to an active identification set. Moreover, the high-order correction direction (used to avoid the Maratos effect) is yielded by solving a system of linear equations (SLE) which also includes only the constraints and their gradients corresponding to the active identification set, therefore, the scale and the computation cost of the high-order correction directions are further decreased. The arc search in our algorithm can effectively combine the initialization processes with the optimization processes, and the iteration points can get into the feasible set after a finite number of iterations. Furthermore, the arc search conditions are weaker than the previous work, and the computation cost is further reduced. The global convergence is proved under the Mangasarian–Fromovitz constraint qualification (MFCQ). If the strong second-order sufficient conditions are satisfied, then the active constraints are exactly identified by the identification set. Without the strict complementarity, superlinear convergence can be obtained. Finally, some elementary numerical results are reported.     

A new sequential systems of linear equations algorithm of feasible descent for inequality constrained optimization

Based on a new efficient identification technique of active constraints introduced in this paper, a new sequential systems of linear equations （SSLE） algorithm generating feasible iterates is proposed for solving nonlinear optimization problems with inequality constraints. In this paper, we introduce a new technique for constructing the system of linear equations, which recurs to a perturbation for the gradients of the constraint functions. At each iteration of the new algorithm, a feasible descent direction is obtained by solving only one system of linear equations without doing convex combination. To ensure the global convergence and avoid the Maratos effect, the algorithm needs to solve two additional reduced systems of linear equations with the same coefficient matrix after finite iterations. The proposed algorithm is proved to be globally and superlinearly convergent under some mild conditions. What distinguishes this algorithm from the previous feasible SSLE algorithms is that an improving direction is obtained easily and the computation cost of generating a new iterate is reduced. Finally, a preliminary implementation has been tested.     

A nonmonotone supermemory gradient algorithm for unconstrained optimization

This paper presents a nonmonotone supermemory gradient algorithm for unconstrained optimization problems. At each iteration, this proposed method sufficiently uses the previous multi-step iterative information and avoids the storage and computation of matrices associated with the Hessian of objective functions, thus it is suitable to solve large-scale optimization problems and can converge stably. Under some assumptions, the convergence properties of the proposed algorithm are analyzed. Numerical results are also reported to show the efficiency of this proposed method.     

一类新的曲线搜索下的记忆梯度法

提出一类新的求解无约束优化问题的记忆梯度法,在较弱条件下证明了算法具有全局收敛性和线性收敛速率.算法采用曲线搜索方法,在每一步同时确定搜索方向和步长,收敛稳定,并且不需计算和存储矩阵,适于求解大规模优化问题.数值试验表明算法是有效的.     

AN EFFECTIVE SEQUENTIAL QUADRATIC PROGRAMMING ALGORITHM FOR NONLINEAR OPTIMIZATION PROBLEMS

In this paper,a new globally convergent algorithm for nonlinear optimization prablems with equality and inequality constraints is presented. The new algorithm is of SQP type which determines a search direction by solving a quadratic programming subproblem per itera-tion. Some revisions on the quadratic programming subproblem have been made in such a way that the associated constraint region is nonempty for each point x generated by the algorithm, i. e. , the subproblems always have optimal solutions. The new algorithm has two important properties. The computation of revision parameter for guaranteeing the consistency of quadratic sub-problem and the computation of the second order correction step for superlinear convergence use the same inverse of a matrix per iteration, so the computation amount of the new algorithm will not be increased much more than other SQP type algorithms; Another is that the new algorithm can give automatically a feasible point as a starting point for the quadratic subproblems pe     

无约束优化的超记忆梯度算法及其收敛特征

研究一种新的无约束优化超记忆梯度算法，算法在每步迭代中充分利用前面迭代点的信息产生下降方向，利用Wolfe线性搜索产生步长，在较弱的条件下证明了算法的全局收敛性。新算法在每步迭代中不需计算和存储矩阵，适于求解大规模优化问题。     

一种新的无约束优化线搜索算法

在对各种有效的线搜索算法分析的基础上,给出了一种求解光滑无约束优化问题的新的线搜索算法.对于目标函数是二次连续可微且下有界的无约束优化问题,算法具有与Wolfe-Powell线搜索算法相同的理论性质.在每一步迭代中算法至多需要计算两次梯度,对于计算目标函数梯度花费较大的情形可以节省一定的计算量.数值试验表明本文算法是可行的和有效的.     

一类无需线性搜索的记忆梯度法

提出一类新的求解无约束优化问题的记忆梯度法,证明了算法的全局收敛性.当目标函数为一致凸函数时,对其线性收敛速率进行了分析.新算法在迭代过程中无需对步长进行线性搜索,仅需对算法中的一些参数进行预测估计,从而减少了目标函数及梯度的迭代次数,降低了算法的计算量和存储量.数值试验表明算法是有效的.     

一类新型DL共轭梯度法研究

提出了求解无约束优化问题的新型DL共轭梯度方法. 同已有方法不同之处在于,该方法构造了一种修正的Armijo线搜索规则,它不仅能给出当前迭代步步长, 而且还能同时确定计算下一步搜索方向时需要用到的共轭参数值. 在较弱的条件下, 建立了算法的全局收敛性理论. 数值试验表明,新型共轭梯度算法比同类方法具有更好的计算效率.     

不等式约束优化一个新型可行QP-free算法

本文对非线性不等式约束优化问题提出了一个新的可行 QP-free 算法. 新算法保存了现有算法的优点, 并具有以下特性: (1) 算法每次迭代只需求解三个具有相同系数矩阵的线性方程组, 计算量小; (2) 可行下降方向只需通过求解一个线性方程组即可获得, 克服了以往分别求解两个线性方程组获得下降方向和可行方向, 然后再做凸组合的困难;(3) 迭代点均为可行点, 并不要求是严格内点; (4) 算法中采用了试探性线搜索,可以进一步减少计算量; (5) 算法中参数很少,数值试验表明算法具有较好的数值效果和较强的稳定性.     

A superlinearly convergent strongly sub-feasible SSLE-type algorithm with working set for nonlinearly constrained optimization

In this paper, by means of a new efficient identification technique of active constraints and the method of strongly sub-feasible direction, we propose a new sequential system of linear equations (SSLE) algorithm for solving inequality constrained optimization problems, in which the initial point is arbitrary. At each iteration, we first yield the working set by a pivoting operation and a generalized projection; then, three or four reduced linear equations with a same coefficient are solved to obtain the search direction. After a finite number of iterations, the algorithm can produced a feasible iteration point, and it becomes the method of feasible directions. Moreover, after finitely many iterations, the working set becomes independent of the iterates and is essentially the same as the active set of the KKT point. Under some mild conditions, the proposed algorithm is proved to be globally, strongly and superlinearly convergent. Finally, some preliminary numerical experiments are reported to show that the algorithm is practicable and effective.     

求解阵列天线自适应滤波问题的一种调比随机逼近算法

提出了求解阵列天线自适应滤波问题的一种调比随机逼近算法.每一步迭代中,算法选取调比的带噪负梯度方向作为新的迭代方向.相比已有的其他随机逼近算法,这个算法不需要调整稳定性常数,在一定程度上解决了稳定性常数选取难的问题.数值仿真实验表明,算法优于已有的滤波算法,且比经典Robbins-Monro （RM）算法具有更好的稳定性.     

修正的SQP型并行变量分配算法

Ferris 和Mangasarian 提出求解最优化问题的PVD(并行变量分配)算法, 此算法是把变量分为主要变量和辅助变量, 分配到p个处理机上, 每个处理机除了负责更新本处理机的主要变量外, 同时还沿着给定的方向更新辅助变量, 使算法的鲁棒性和灵活性得到了很大的提高. 该文基于文献[6]提出一种修正的SQP型PVD算法, 构造其搜索方向是下降方向和可行方向的组合, 并对此方向给予一个高阶修正, 使此算法很好地防止 Maratos 效应发生, 而且能够克服在求解子问题时出现约束不相容的情况. 在合适的条件下, 推导出此算法具有全局收敛性.     

A superlinearly convergent norm-relaxed SQP method of strongly sub-feasible directions for constrained optimization without strict complementarity

In this paper, a kind of optimization problems with nonlinear inequality constraints is discussed. Combined the ideas of norm-relaxed SQP method and strongly sub-feasible direction method as well as a pivoting operation, a new fast algorithm with arbitrary initial point for the discussed problem is presented. At each iteration of the algorithm, an improved direction is obtained by solving only one direction finding subproblem which possesses small scale and always has an optimal solution, and to avoid the Maratos effect, another correction direction is yielded by a simple explicit formula. Since the line search technique can automatically combine the initialization and optimization processes, after finite iterations, the iteration points always get into the feasible set. The proposed algorithm is proved to be globally convergent and superlinearly convergent under mild conditions without the strict complementarity. Finally, some numerical tests are reported.     

On the eigenvalues of the saddle point matrices discretized from Navier–Stokes equations

In this paper, we proposed an implementation of stochastic perturbation of reduced gradient and bisection (SPRGB) method for optimizing a non-convex differentiable function subject to linear equality constraints and non-negativity bounds on the variables. In particular, at each iteration, we compute a search direction by reduced gradient, and optimal line search by bisection algorithm along this direction yields a decrease in the objective value. SPRGB method is desired to establish the global convergence of the algorithm. An implementation and tests of SPRGB algorithm are given, and some numerical results of large-scale problems are presented, which show the efficient of this approach.