一种求解无约束优化问题的非单调信赖域算法

提出了非单调信赖域算法求解基于锥模型的无约束优化问题,该算法在求解信赖域子问题时充分利用了当前迭代点的一阶梯度信息。提出了一个新的信赖域半径的选取机制,并和经典的信赖域方法作比较分析。设定了一些条件,在这些假设条件下证明了算法是整体收敛的。数值实验结果表明,该算法对基于锥模型的无约束优化问题的求解是行之有效的,拓展了非单调信赖域算法的应用领域。     

前向神经网络信赖域学习算法的研究

本文介绍了引入信赖域优化理论解决神经网络中学习问题的新算法,提出了计算有效信赖域步方法,以保证信赖域算法的正确性,采用变系数方法避免了信赖域半径自适应调整过程中不稳定和低效的问题。实验表明,信赖域学习算法优于变尺度算法。     

一类新的带记忆模型的非单调信赖域算法

本文就无约束优化问题提出了一个带记忆模型的非单调信赖域算法。与传统的非单调信赖域算法不同,文中的信赖域子问题的逼近模型为记忆模型,该模型使我们可以从更全面的角度来求得信赖域试探步,从而避免了传统非单调信赖域方法中试探步的求取完全依赖于当前点的信息而过于局部化的困难。文中提出了一个带记忆模型的非单调信赖域
域算法,并证明了其全局收敛性。     

求解半定规划的新算法

为了提高求解半定规划问题的运算效率,提出了一种新的求解半定规划的非单调信赖域算法。将半定规划的最优性条件转化为无约束优化问题,并构造无约束优化问题的信赖域子问题,修正信赖域半径的校正条件,当初始搜索点处于峡谷附近时仍能搜索到全局最优解。实验结果表明,对于小规模和中等规模的半定规划问题,该算法的迭代次数都比经典的内点算法少,运行速度快。     

求解半定规划的新算法

为了提高求解半定规划问题的运算效率,提出了一种新的求解半定规划的非单调信赖域算法。将半定规划的最优性条件转化为无约束优化问题,并构造无约束优化问题的信赖域子问题,修正信赖域半径的校正条件,当初始搜索点处于峡谷附近时仍能搜索到全局最优解。实验结果表明,对于小规模和中等规模的半定规划问题,该算法的迭代次数都比经典的内点算法少,运行速度快。     

用改进的信赖域方法求解二次插值模型

一种改进的信赖域方法被用来解无约束最优化问题,当目标函数的导数信息不可利用或者求解目标函数的导数代价太大。通常,考虑用二次插值模型来逼近目标函数,并且用传统的信赖域方法求解这个二次模型。传统的信赖域方法将被改进,并且形成两个改进的信赖域子问题。改进的信赖域方法的创新点在于:求解二次模型在一个参数化的信赖域中,修改这个模型在另一个参数化的信赖域当中。在这两个新的信赖域中,可以分别很快地找到一个好的下降方向和一个具有均衡性的插值点。这个改进的方法不但节省了函数值计算次数而且提高了解的精度。实验结果表明,针对测试问题,提出的方法的确是优于传统的信赖域方法的。     

一种BFGS校正的改进信赖域方法

本文利用经典的信赖域方法,针对无约束优化问题,对信赖域进行改进,并在此基础上对算法进行BFGS校正。数值实验证明,相比传统的信赖域方法,改进的信赖域方法在计算效率上有了很大提高;而加入BFGS校正后,新算法相比改进的信赖域方法又有了进一步的提高。     

一种求解半定规划的非单调信赖域算法

提出一种求解半定规划的非单调信赖域算法。利用推广至矩阵域的光滑Fischer-Burmeister函数,转化半定规划的最优性条件,改写半定规划的中心路径,得到与其等价的无约束优化问题的非线性可微光滑方程组,在求解信赖域子问题时,利用当前迭代点的一阶梯度信息,给出信赖域半径的选取机制。仿真结果表明,与经典的内点算法相比,对于一般规模(n, m≤30)的半定规划问题,该算法的运行速度较快。对于大规模的半定规划问题(n, m＞30),该算法更适合处理Norm min、Lovasz这2类问题。     

非单调信赖域方法求解无约束非光滑优化问题

提出了非单调信赖域算法求解无约束非光滑优化问题,并和经典的信赖域方法作比较分析。同时,设定了一些条件,在这些假设条件下证明了该算法是整体收敛的。数值实验结果表明,非单调策略对无约束非光滑优化问题的求解是行之有效的,拓展了非单调信赖域算法的应用领域。     

基于回代信赖域的电阻层析成像算法

针对电阻层析成像(ERT)技术中反演问题的病态性,提出一种改进的回代信赖域算法BTR(Backtracking Trust Region),并将其应用于气/水两相流的可视化测量。该算法通过信赖域算法获得迭代方向,通过回代技术获得迭代步长,可在减小重建误差的同时,提高成像速度。利用Comsol软件进行仿真,并设计ERT系统对各种典型流型进行测量,验证了算法的可行性。通过与Landweber算法、共轭梯度算法和现存的信赖域算法的比较,证明本文方法明显改进了成像精度和实时性。     

An adaptive nonmonotone trust region algorithm

Based on an eigenvalue analysis conducted on the scaled memoryless quasi-Newton updating formulas BFGS and DFP, an adaptive choice for the trust region radius is proposed. Then, using a trust region ratio obtained from a nonmonotone line search strategy, an adaptive nonmonotone trust region algorithm is developed. Under proper conditions, it is briefly shown that the proposed algorithm is globally and locally superlinearly convergent. Numerical experiments are done on a set of unconstrained optimization test problems of the CUTEr collection, using the Dolan–Moré performance profile. They show efficiency of the proposed algorithm.     

A trust region method based on a new affine scaling technique for simple bounded optimization

In this paper, we propose a new trust region affine scaling method for nonlinear programming with simple bounds. Our new method is an interior-point trust region method with a new scaling technique. The scaling matrix depends on the distances of the current iterate to the boundaries, the gradient of the objective function and the trust region radius. This scaling technique is different from the existing ones. It is motivated by our analysis of the linear programming case. The trial step is obtained by minimizing the quadratic approximation to the objective function in the scaled trust region. It is proved that our algorithm guarantees that at least one accumulation point of the iterates is a stationary point. Preliminary numerical experience on problems with simple bounds from the CUTEr collection is also reported. The numerical performance reveals that our method is effective and competitive with the famous algorithm LANCELOT. It also indicates that the new scaling technique is very effective and might be a good alternative to that used in the subroutine fmincon from Matlab optimization toolbox.     

A BFGS trust-region method for nonlinear equations

In this paper, a new trust-region subproblem combining with the BFGS update is proposed for solving nonlinear equations, where the trust region radius is defined by a new way. The global convergence without the nondegeneracy assumption and the quadratic convergence are obtained under suitable conditions. Numerical results show that this method is more effective than the norm method.     

Adaptive augmented Lagrangian methods: algorithms and practical numerical experience

In this paper, we consider augmented Lagrangian (AL) algorithms for solving large-scale nonlinear optimization problems that execute adaptive strategies for updating the penalty parameter. Our work is motivated by the recently proposed adaptive AL trust region method by Curtis et al. [An adaptive augmented Lagrangian method for large-scale constrained optimization, Math. Program. 152 (2015), pp. 201–245.]. The first focal point of this paper is a new variant of the approach that employs a line search rather than a trust region strategy, where a critical algorithmic feature for the line search strategy is the use of convexified piecewise quadratic models of the AL function for computing the search directions. We prove global convergence guarantees for our line search algorithm that are on par with those for the previously proposed trust region method. A second focal point of this paper is the practical performance of the line search and trust region algorithm variants in Matlab software, as well as that of an adaptive penalty parameter updating strategy incorporated into the Lancelot software. We test these methods on problems from the CUTEst and COPS collections, as well as on challenging test problems related to optimal power flow. Our numerical experience suggests that the adaptive algorithms outperform traditional AL methods in terms of efficiency and reliability. As with traditional AL algorithms, the adaptive methods are matrix-free and thus represent a viable option for solving large-scale problems.     

An efficient strategy for reliability-based multidisciplinary design optimization using BLISS

This paper presents an efficient reliability-based multidisciplinary design optimization (RBMDO) strategy. The conventional RBMDO has tri-level loops: the first level is an optimization in the deterministic space, the second one is a reliability analysis in the probabilistic space, and the third one is the multidisciplinary analysis. Since it is computationally inefficient when high-fidelity simulation methods are involved, an efficient strategy is proposed. The strategy [named probabilistic bi-level integrated system synthesis (ProBLISS)] utilizes a single-level reliability-based design optimization (RBDO) approach, in which the reliability analysis and optimization are conducted in a sequential manner by approximating limit state functions. The single-level RBDO is associated with the BLISS formulation to solve RBMDO problems. Since both the single-level RBDO and BLISS are mainly driven by approximate models, the accuracy of models can be a critical issue for convergence. The convergence of the strategy is guaranteed by employing the trust region–sequential quadratic programming framework, which validates approximation models in the trust region radius. Two multidisciplinary problems are tested to verify the strategy. ProBLISS significantly reduces the computational cost and shows stable convergence while maintaining accuracy.     

Update strategies for kriging models used in variable fidelity optimization

Many optimization methods for simulation-based design rely on the sequential use of metamodels to reduce the associated computational burden. In particular, kriging models are frequently used in variable fidelity optimization. Nevertheless, such methods may become computationally inefficient when solving problems with large numbers of design variables and/or sampled data points due to the expensive process of optimizing the kriging model parameters in each iteration. One solution to this problem would be to replace the kriging models with traditional Taylor series response surface models. Kriging models, however, were shown to provide good approximations of computer simulations that incorporate larger amounts of data, resulting in better global accuracy. In this paper, a metamodel update management scheme (MUMS) is proposed to reduce the cost of using kriging models sequentially by updating the kriging model parameters only when they produce a poor approximation. The scheme uses the trust region ratio (TR-MUMS), which is a ratio that compares the approximation to the true model. Two demonstration problems are used to evaluate the proposed method: an internal combustion engine sizing problem and a control-augmented structural design problem. The results indicate that the TR-MUMS approach is very effective; on the demonstration problems, it reduced the number of likelihood evaluations by three orders of magnitude compared to using a global optimizer to find the kriging parameters in every iteration. It was also found that in trust region-based method, the kriging model parameters need not be updated using a global optimizer—local methods perform just as well in terms of providing a good approximation without affecting the overall convergence rate, which, in turn, results in a faster execution time.