General primal-dual penalty/barrier path-following Newton methods for nonlinear programming

In the present article rather general penalty/barrier-methods are considered, that define a local continuously differentiable primal-dual path. The class of penalty/barrier terms includes most of the usual techniques like logarithmic barriers, SUMT, quadratic loss functions as well as exponential penalties, and the optimization problem which may contain inequality as well as equality constraints. The convergence of the corresponding general primal-dual path-following method is shown for local minima that satisfy strong second-order sufficiency conditions with linear independence constraint qualification (LICQ) and strict complementarity. A basic tool in the analysis of these methods is to estimate the radius of convergence of Newton's method depending on the penalty/barrier-parameter. Without using self-concordance properties convergence bounds are derived by direct estimations of the solutions of the Newton equations. Parameter selection rules are proposed which guarantee the local convergence of the considered penalty/barrier-techniques with only a finite number of Newton steps at each parameter level. Numerical examples illustrate the practical behavior of the proposed class of methods.     

Computational experience with known variable metric updates

Variable metric methods from the Broyden family are well known and commonly used for unconstrained minimization. These methods have good theoretical and practical convergence properties which depend on a selection of free parameters. We demonstrate, using extensive computational experiments, the influence of both the Biggs stabilization parameter and Oren scaling parameter on 12 individual variable metric updates, two of which are new. This paper focuses on a class of variable metric updates belonging to the so-called preconvex part of the Broyden family. These methods outperform the more familiar BFGS method. We also experimentally demonstrate the efficiency of the controlled scaling strategy for problems of sufficient size and sparsity.     

A Path Algorithm for Constrained Estimation

Many least-square problems involve affine equality and inequality constraints. Although there are a variety of methods for solving such problems, most statisticians find constrained estimation challenging. The current article proposes a new path-following algorithm for quadratic programming that replaces hard constraints by what are called exact penalties. Similar penalties arise in l ₁ regularization in model selection. In the regularization setting, penalties encapsulate prior knowledge, and penalized parameter estimates represent a trade-off between the observed data and the prior knowledge. Classical penalty methods of optimization, such as the quadratic penalty method, solve a sequence of unconstrained problems that put greater and greater stress on meeting the constraints. In the limit as the penalty constant tends to ∞, one recovers the constrained solution. In the exact penalty method, squared penalties are replaced by absolute value penalties, and the solution is recovered for a finite value of the penalty constant. The exact path-following method starts at the unconstrained solution and follows the solution path as the penalty constant increases. In the process, the solution path hits, slides along, and exits from the various constraints. Path following in Lasso penalized regression, in contrast, starts with a large value of the penalty constant and works its way downward. In both settings, inspection of the entire solution path is revealing. Just as with the Lasso and generalized Lasso, it is possible to plot the effective degrees of freedom along the solution path. For a strictly convex quadratic program, the exact penalty algorithm can be framed entirely in terms of the sweep operator of regression analysis. A few well-chosen examples illustrate the mechanics and potential of path following. This article has supplementary materials available online.     

An adaptive updating full-Newton step interior-point algorithm with modified Newton direction

We adopt the self-adaptive strategy to update the barrier parameter of a feasible primal-dual interior-point algorithm. We obtain two adaptive updating methods, namely, cheap updates and sharp updates. We compare the effectiveness of the short updates with the adaptive update methods on some benchmark problems. The numerical results show that the sharp updates method is superior to short updates and cheap updates methods.     

Modified Cross-Validation for Penalized High-Dimensional Linear Regression Models

In this article, for Lasso penalized linear regression models in high-dimensional settings, we propose a modified cross-validation (CV) method for selecting the penalty parameter. The methodology is extended to other penalties, such as Elastic Net. We conduct extensive simulation studies and real data analysis to compare the performance of the modified CV method with other methods. It is shown that the popular K-fold CV method includes many noise variables in the selected model, while the modified CV works well in a wide range of coefficient and correlation settings. Supplementary materials containing the computer code are available online.     

Computational Discretization Algorithms for Functional Inequality Constrained Optimization

In this paper, a functional inequality constrained optimization problem is studied using a discretization method and an adaptive scheme. The problem is discretized by partitioning the interval of the independent parameter. Two methods are investigated as to how to treat the discretized optimization problem. The discretization problem is firstly converted into an optimization problem with a single nonsmooth equality constraint. Since the obtained equality constraint is nonsmooth and does not satisfy the usual constraint qualification condition, relaxation and smoothing techniques are used to approximate the equality constraint via a smooth inequality constraint. This leads to a sequence of approximate smooth optimization problems with one constraint. An adaptive scheme is incorporated into the method to facilitate the computation of the sum in the inequality constraint. The second method is to apply an adaptive scheme directly to the discretization problem. Thus a sequence of optimization problems with a small number of inequality constraints are obtained. Convergence analysis for both methods is established. Numerical examples show that each of the two proposed methods has its own advantages and disadvantages over the other.     

Optimization via simulation: A review

We review techniques for optimizing stochastic discrete-event systems via simulation. We discuss both the discrete parameter case and the continuous parameter case, but concentrate on the latter which has dominated most of the recent research in the area. For the discrete parameter case, we focus on the techniques for optimization from a finite set: multiple-comparison procedures and ranking-and-selection procedures. For the continuous parameter case, we focus on gradient-based methods, including perturbation analysis, the likelihood ratio method, and frequency domain experimentation. For illustrative purposes, we compare and contrast the implementation of the techniques for some simple discrete-event systems such as the (s, S) inventory system and theGI/G/1 queue. Finally, we speculate on future directions for the field, particularly in the context of the rapid advances being made in parallel computing.     

Stochastic programming problems with generalized integrated chance constraints

If the constraints in an optimization problem are dependent on a random parameter, we would like to ensure that they are fulfilled with a high level of reliability. The most natural way is to employ chance constraints. However, the resulting problem is very hard to solve. We propose an alternative formulation of stochastic programs using penalty functions. The expectations of penalties can be left as constraints leading to generalized integrated chance constraints, or incorporated into the objective as a penalty term. We show that the penalty problems are asymptotically equivalent under quite mild conditions. We discuss applications of sample-approximation techniques to the problems with generalized integrated chance constraints and propose rates of convergence for the set of feasible solutions. We will direct our attention to the case when the set of feasible solutions is finite, which can appear in integer programming. The results are then extended to the bounded sets with continuous variables. Additional binary variables are necessary to solve sample-approximated chance-constrained problems, leading to a large mixed-integer non-linear program. On the other hand, the problems with penalties can be solved without adding binary variables; just continuous variables are necessary to model the penalties. The introduced approaches are applied to the blending problem leading to comparably reliable solutions.     

On choosing “optimal” shape parameters for RBF approximation

Many radial basis function (RBF) methods contain a free shape parameter that plays an important role for the accuracy of the method. In most papers the authors end up choosing this shape parameter by trial and error or some other ad hoc means. The method of cross validation has long been used in the statistics literature, and the special case of leave-one-out cross validation forms the basis of the algorithm for choosing an optimal value of the shape parameter proposed by Rippa in the setting of scattered data interpolation with RBFs. We discuss extensions of this approach that can be applied in the setting of iterated approximate moving least squares approximation of function value data and for RBF pseudo-spectral methods for the solution of partial differential equations. The former method can be viewed as an efficient alternative to ridge regression or smoothing spline approximation, while the latter forms an extension of the classical polynomial pseudo-spectral approach. Numerical experiments illustrating the use of our algorithms are included.     

Lavrentiev's regularization method for nonlinear ill-posed equations in Banach spaces

In this paper, we deal with nonlinear ill-posed problems involving m-accretive mappings in Banach spaces. We consider a derivative and inverse free method for the implementation of Lavrentiev regularization method. Using general H¨older type source condition we obtain an optimal order error estimate. Also we consider the adaptive parameter choice strategy proposed by Pereverzev and Schock(2005) for choosing the regularization parameter.     

Dynamic updates of the barrier parameter in primal-dual methods for nonlinear programming

We introduce a framework in which updating rules for the barrier parameter in primal-dual interior-point methods become dynamic. The original primal-dual system is augmented to incorporate explicitly an updating function. A Newton step for the augmented system gives a primal-dual Newton step and also a step in the barrier parameter. Based on local information and a line search, the decrease of the barrier parameter is automatically adjusted. We analyze local convergence properties, report numerical experiments on a standard collection of nonlinear problems and compare our results to a state-of-the-art interior-point implementation. In many instances, the adaptive algorithm reduces the number of iterations and of function evaluations. Its design guarantees a better fit between the magnitudes of the primal-dual residual and of the barrier parameter along the iterations.     

The penalty method for grid matching in mixed finite element methods

In this paper we prove the possibility of the use of the penalty method for grid matching in mixed finite element methods. We consider the Hermann-Johnson scheme for biharmonic equation. The main idea is to construct a perturbed problem with two parameters which play roles of penalties. The perturbed problem is built by the replacement of essential conditions on the interface in the mixed variational statement with natural conditions that contain parameters. The perturbed problem is discretized by the finite element method. We estimate the norm of the difference between a solution of the discrete perturbed problem and a solution of the initial problem; the obtained estimates depend on the step and the penalties. We give recommendations for the choice of penalties depending on the step.     

多元Fredholm积分方程自适应直接方法的优化

考虑了以各向异性光滑函数为核及自由项的多元Fredholm积分方程类的自适应直接方法的优化,得到误差阶的精确估计及相应的最优算法。     

Asymptotic theory of the adaptive Sparse Group Lasso

We study the asymptotic properties of a new version of the Sparse Group Lasso estimator (SGL), called adaptive SGL. This new version includes two distinct regularization parameters, one for the Lasso penalty and one for the Group Lasso penalty, and we consider the adaptive version of this regularization, where both penalties are weighted by preliminary random coefficients. The asymptotic properties are established in a general framework, where the data are dependent and the loss function is convex. We prove that this estimator satisfies the oracle property: the sparsity-based estimator recovers the true underlying sparse model and is asymptotically normally distributed. We also study its asymptotic properties in a double-asymptotic framework, where the number of parameters diverges with the sample size. We show by simulations and on real data that the adaptive SGL outperforms other oracle-like methods in terms of estimation precision and variable selection.

     

A uniform framework for the combination of penalties in generalized structured models

Penalized estimation has become an established tool for regularization and model selection in regression models. A variety of penalties with specific features are available and effective algorithms for specific penalties have been proposed. But not much is available to fit models with a combination of different penalties. When modeling the rent data of Munich as in our application, various types of predictors call for a combination of a Ridge, a group Lasso and a Lasso-type penalty within one model. We propose to approximate penalties that are (semi-)norms of scalar linear transformations of the coefficient vector in generalized structured models—such that penalties of various kinds can be combined in one model. The approach is very general such that the Lasso, the fused Lasso, the Ridge, the smoothly clipped absolute deviation penalty, the elastic net and many more penalties are embedded. The computation is based on conventional penalized iteratively re-weighted least squares algorithms and hence, easy to implement. New penalties can be incorporated quickly. The approach is extended to penalties with vector based arguments. There are several possibilities to choose the penalty parameter(s). A software implementation is available. Some illustrative examples show promising results.     

Bayesian adaptive Lasso

We propose the Bayesian adaptive Lasso (BaLasso) for variable selection and coefficient estimation in linear regression. The BaLasso is adaptive to the signal level by adopting different shrinkage for different coefficients. Furthermore, we provide a model selection machinery for the BaLasso by assessing the posterior conditional mode estimates, motivated by the hierarchical Bayesian interpretation of the Lasso. Our formulation also permits prediction using a model averaging strategy. We discuss other variants of this new approach and provide a unified framework for variable selection using flexible penalties. Empirical evidence of the attractiveness of the method is demonstrated via extensive simulation studies and data analysis.     

Approximative solutions to the bicriterion Vehicle Routing Problem with Time Windows

The Vehicle Routing Problem with Time Windows (VRPTW) is a combinatorial optimization problem. It deals with route planning and the distribution of goods from a depot to geographically dispersed customers by a fleet of vehicles with constrained capacities. The customers’ demands are known and each customer has a time window in which it has to be supplied. The time windows are assumed to be soft, that means, violations of the time windows are allowed, but associated with penalties. The problem is to organize the vehicle routes optimally, i.e. to minimize the total costs, consisting of the number of used vehicles and the total distance, and the penalties simultaneously. Thus, the problem is formulated as a bicriterion minimization problem and heuristic methods are used to calculate approximations of the Pareto optimal solutions. Experimental results show that in certain cases the allowance of penalties leads to significant savings of the total costs.     

Job scheduling with forbidden setups and two objectives using genetic algorithms and penalties

One of the most important tasks in service and manufacturing systems is how to schedule arriving jobs such that some criteria will be satisfied. Up to now there have been defined a great variety of scheduling problems as well as corresponding models and solution approaches. Most models suffer from such more or less restrictive assumptions like single machine, unique processing times, zero set-up times or a single criterion. On the other hand some classical approaches like linear or dynamic programming are practicable for small-size problems only. Therefore over the past years we can observe an increasing application of heuristic search methods. But scheduling problems with multiple machines, forbidden setups and multiple objectives are scarcely considered. In our paper we apply a Genetic Algorithm to such a problem which was found at a continuous casting plant. Because of the forbidden setups the probability for a random generated schedule to be feasible is nearly zero. To resolve this problem we use three kinds of penalties, a global, a local and a combined approach. For performance investigations of these penalty types we applied our approaches to a real world test instance with 96 jobs, three machines and two objectives. We tested five different penalty levels with 51 independent runs to evaluate the impact of the penalties.     

求解带线性约束的凸优化的一类自适应不定线性化增广拉格朗日方法

增广拉格朗日方法是求解带线性约束的凸优化问题的有效算法.线性化增广拉格朗日方法通过线性化增广拉格朗日函数的二次罚项并加上一个临近正则项,使得子问题容易求解,其中正则项系数的恰当选取对算法的收敛性和收敛速度至关重要.较大的系数可保证算法收敛性,但容易导致小步长.较小的系数允许迭代步长增大,但容易导致算法不收敛.本文考虑求解带线性等式或不等式约束的凸优化问题.我们利用自适应技术设计了一类不定线性化增广拉格朗日方法,即利用当前迭代点的信息自适应选取合适的正则项系数,在保证收敛性的前提下尽量使得子问题步长选择范围更大,从而提高算法收敛速度.我们从理论上证明了算法的全局收敛性,并利用数值实验说明了算法的有效性.     

Two optimal Dai–Liao conjugate gradient methods

Two adaptive choices for the parameter of Dai–Liao conjugate gradient (CG) method are suggested. One of which is obtained by minimizing the distance between search directions of Dai–Liao method and a three-term CG method proposed by Zhang et al. and the other one is obtained by minimizing Frobenius condition number of the search direction matrix. Global convergence analyses are made briefly. Numerical results are reported; they demonstrate effectiveness of the suggested adaptive choices.