求解不可分离非凸非光滑问题的线性惯性ADMM算法

针对目标函数中包含耦合函数H(x,y)的非凸非光滑极小化问题,提出了一种线性惯性交替乘子方向法(Linear Inertial Alternating Direction Method of Multipliers,LIADMM)。为了方便子问题的求解,对目标函数中的耦合函数H(x,y)进行线性化处理,并在x-子问题中引入惯性效应。在适当的假设条件下,建立了算法的全局收敛性;同时引入满足Kurdyka-Lojasiewicz不等式的辅助函数,验证了算法的强收敛性。通过两个数值实验表明,引入惯性效应的算法比没有惯性效应的算法收敛性能更好。

求解一类非光滑凸优化问题的相对加速SGD算法

一阶优化算法由于其计算简单、代价小,被广泛应用于机器学习、大数据科学、计算机视觉等领域,然而,现有的一阶算法大多要求目标函数具有Lipschitz连续梯度,而实际中的很多应用问题不满足该要求。在经典的梯度下降算法基础上,引入随机和加速,提出一种相对加速随机梯度下降算法。该算法不要求目标函数具有Lipschitz连续梯度,而是通过将欧氏距离推广为Bregman距离,从而将Lipschitz连续梯度条件减弱为相对光滑性条件。相对加速随机梯度下降算法的收敛性与一致三角尺度指数有关,为避免调节最优一致三角尺度指数参数的工作量,给出一种自适应相对加速随机梯度下降算法。该算法可自适应地选取一致三角尺度指数参数。对算法收敛性的理论分析表明,算法迭代序列的目标函数值收敛于最优目标函数值。针对Possion反问题和目标函数的Hessian阵算子范数随变量范数多项式增长的极小化问题的数值实验表明,自适应相对加速随机梯度下降算法和相对加速随机梯度下降算法的收敛性能优于相对随机梯度下降算法。

车辆在非光滑路面的接触动力学分析方法研究

车辆接触动力学研究有助于更加全面地了解整车力学性能,对整车稳定性和座舱舒适性设计具有指导性作用.文章基于非光滑多体系统理论,对车辆系统展开接触动力学研究.考虑到车辆在行驶过程中,轮胎是与地面接触的唯一部件,轮胎动力学模型对整车动力学分析发挥着重要的作用.首先,借助于张拉整体结构的思想,建立轮胎的等效模型.然后,采用拉格朗日乘子方法,将约束方程引入到拉格朗日函数中,并基于第二类拉格朗日方程,推导得到整车系统的一般动力学方程.由于接触碰撞的存在,会导致车辆的状态变量呈现出非连续性.利用牛顿碰撞定律,建立了碰撞响应的非光滑数学列式,并利用光滑化的Fischer-Burmeister函数将非光滑数学列式进行等效替换,以提高计算效率.其次,将摩擦响应的求解转化为优化问题进行描述,以避免求解摩擦方向时出现奇异.另外,将车辆系统的一般动力学模型分解为光滑部分与非光滑部分进行求解,并利用广义α离散策略给出了详细的求解流程.最后,对整车系统展开了多工况的接触碰撞数值仿真,分析了不同工况下的接触响应对整车动力学性能的影响,验证了所提方法的有效性.

Accelerated Primal-Dual Projection Neurodynamic Approach With Time Scaling for Linear and Set Constrained Convex Optimization Problems

The Nesterov accelerated dynamical approach serves as an essential tool for addressing convex optimization problems with accelerated convergence rates.Most previous studies in this field have primarily concentrated on unconstrained smooth con-vex optimization problems.In this paper,on the basis of primal-dual dynamical approach,Nesterov accelerated dynamical approach,projection operator and directional gradient,we present two accelerated primal-dual projection neurodynamic approaches with time scaling to address convex optimization problems with smooth and nonsmooth objective functions subject to linear and set constraints,which consist of a second-order ODE(ordinary differential equation)or differential conclusion system for the primal variables and a first-order ODE for the dual vari-ables.By satisfying specific conditions for time scaling,we demonstrate that the proposed approaches have a faster conver-gence rate.This only requires assuming convexity of the objective function.We validate the effectiveness of our proposed two accel-erated primal-dual projection neurodynamic approaches through numerical experiments.

定向距离函数的光滑化方法及其应用

本文考虑定向距离函数的光滑化表示及其应用。首先在已有的两种光滑化方法的基础上,给出了这类特殊的非光滑函数的光滑化表示。作为特例,在二维空间中,给出该函数更具体的光滑化函数。最后利用定向距离函数的光滑化函数以及它在多目标优化问题标量化方法中的应用,建立非光滑多目标优化问题的光滑标量化模型,并给出了两者之间解集的关系。

基于BB步长的近端随机递归动量算法

研究了一个求解非凸非光滑复合优化问题的算法。首先,结合近端随机递归动量算法和改进的BB步长,提出了一种带BB步长的随机方差缩减算法(ProxSTORM-BB)求解非凸非光滑复合优化问题。该算法在迭代过程中通过动态调节步长来提高算法的计算效率,并且对初始步长的选取不敏感,解决了参数调优比较困难这一问题。然后,在合适的假设条件下证明了算法的收敛性。最后,通过数值实验验证了算法的有效性。

一种非精确非光滑信赖域算法

Aravkin等人提出了求解非光滑优化问题min_(x∈R^(d))f(x)+h(x)的非光滑信赖域算法(采用f的精确梯度),其中f是连续可微函数,h是邻近有界且下半连续的真函数。文章研究当该问题中f:=1/n ∑_(i=1)^(n)f_(i)(n很大且每个分量函数fi是连续可微)时,求解这类大规模可分离非光滑优化问题的有效算法。结合非精确算法和非光滑信赖域算法的思想,提出了用非精确梯度代替精确梯度的非精确非光滑信赖域算法。与非光滑信赖域算法(采用精确梯度)相比,该算法降低了每次迭代的计算量。在一定的假设条件下,证明了算法的迭代复杂度。

Nonsmooth finite-time control of uncertain second-order nonlinear systems

Nonsmooth finite-time stabilizing control laws have been developed for the double integrator system. The objective of this paper is to further explore the finite-time tracking control problem of a general form of uncertain secondorder affine nonlinear system with the new forms of terminal sliding mode （TSM）. Discontinuous and continuous finite-time controllers are also developed respectively without the singularity problem. Complete robustness can be acquired with the former, and enhanced robustness compared with the conventional boundary layer method can be expressed as explicit bounded function with the latter. Simulation results on the stabilizing and tracking problems are presented to demonstrate the effectiveness of the control algorithms.

Nonsmooth Equations of K-T Systems for a Constrained Minimax Problem

Using K-T optimality condition of nonsmooth optimization, we establish two equivalent systems of the nonsmooth equations for the constrained minimax problem directly. Then generalized Newton methods are applied to solve these systems of the nonsmooth equations. Thus a new approach to solving the constrained minimax problem is developed.

NONSMOOTH CRITICAL POINT THEOREMS AND ITS APPLICATIONS TO QUASILINEAR SCHRDINGER EQUATIONS

In this paper, the existence and nonexistence of solutions to a class of quasilinear elliptic equations with nonsmooth functionals are discussed, and the results obtained are applied to quasilinear SchrSdinger equations with negative parameter which arose from the study of self-channeling of high-power ultrashort laser in matter.

OPTIMALITY CONDITIONS AND DUALITY RESULTS FOR NONSMOOTH VECTOR OPTIMIZATION PROBLEMS WITH THE MULTIPLE INTERVAL-VALUED OBJECTIVE FUNCTION

In this paper, both Fritz John and Karush-Kuhn-Tucker necessary optimality conditions are established for a （weakly） LU-efficient solution in the considered nonsmooth multiobjective programming problem with the multiple interval-objective function. Further, the sufficient optimality conditions for a （weakly） LU-efficient solution and several duality results in Mond-Weir sense are proved under assumptions that the functions constituting the considered nondifferentiable multiobjective programming problem with the multiple interval- objective function are convex.

A New Definition for Generalized First Derivative of Nonsmooth Functions

In this paper, we define a functional optimization problem corresponding to smooth functions which its optimal solution is first derivative of these functions in a domain. These functional optimization problems are applied for non-smooth functions which by solving these problems we obtain a kind of generalized first derivatives. For this purpose, a linear programming problem corresponding functional optimization problem is obtained which their optimal solutions give the approximate generalized first derivative. We show the efficiency of our approach by obtaining derivative and generalized derivative of some smooth and nonsmooth functions respectively in some illustrative examples.

Newton type methods for solving nonsmooth equations

Numerical methods for the solution of nonsmooth equations are studied. A new subdifferential for a locally Lipschitzian function is proposed. Based on this subdifferential, Newton methods for solving nonsmooth equations are developed and their convergence is shown. Since this subdifferential is easy to be computed, the present Newton methods can be executed easily in some applications.

Merit functions for nonsmooth complementarity problems and related descent algorithm

Under some assumptions, the solution set of a nonlinear complementarity problem coincides with the set of local minima of the corresponding minimization problem. This paper uses a family of new merit functions to deal with nonlinear complementarity problem where the underlying function is assumed to be a continuous but not necessarily locally Lipschitzian map and gives a descent algorithm for solving the nonsmooth continuous complementarity problems. In addition, the global convergence of the derivative free descent algorithm is also proved.

A NONMOTOTONE ALGORITHM FOR MINIMIZING NONSMOOTH COMPOSITE FUNCTIONS

In this paper, we present a nonmonotone algorithm for solving nonsmooth composite optimization problems. The objective function of these problems is composited by a nonsmooth convex function and a differentiable function. The method generates the search directions by solving quadratic programming successively, and makes use of the nonmonotone line search instead of the usual Armijo-type line search. Global convergence is proved under standard assumptions. Numerical results are given.

A SQP METHOD FOR MINIMIZING A CLASS OF NONSMOOTH FUNCTIONS

In this paper,we present a successive quadratic programming(SQP)method for minimizing a class of nonsmooth functions,which are the sum of a convex function and a nonsmooth composite function.The method generates new iterations by using the Armijo-type line search technique after having found the search directions.Global convergence property is established under mild assumptions.Numerical results are also offered.

NONDESCENT SUBGRADIENT METHOD FOR NONSMOOTH CONSTRAINED MINIMIZATION

A kind of nondecreasing subgradient algorithm with appropriate stopping rule has been proposed for nonsmooth constrained minimization problem. The dual theory is invoked in dealing with the stopping rule and general global minimiizing algorithm is employed as a subroutine of the algorithm. The method is expected to tackle a large class of nonsmooth constrained minimization problem.

NONSMOOTH MODEL FOR PLASTIC LIMIT ANALYSIS AND ITS SMOOTHING ALGORITHM

By means of Lagrange duality of Hill＇s maximum plastic work principle theory of the convex program, a dual problem under Mises＇ yield condition has been derived and whereby a non-differentiable convex optimization model for the limit analysis is developed. With this model, it is not necessary to linearize the yield condition and its discrete form becomes a minimization problem of the sum of Euclidean norms subject to linear constraints. Aimed at resolving the non-differentiability of Euclidean norms, a smoothing algorithm for the limit analysis of perfect-plastic continuum media is proposed. Its efficiency is demonstrated by computing the limit load factor and the collapse state for some plane stress and plain strain problems.

NONLINEAR KRYLOV SUBSPACE METHODS FOR SOLVING NONSMOOTH EQUATIONS

Newton-FOM （Full Orthogonalization Method ） algorithm and NewtonGMRES （Generalized Minimum Residual Method） algorithm for solving nonsmooth equations are presented. It is proved that these Krylov subspace algorithms have the locally quadratic convergence. Numerical experiments demonstrate the effectiveness of the algorithms.

Higher-Order Minimizers and Generalized (F,<i>ρ</i>)-Convexity in Nonsmooth Vector Optimization over Cones

In this paper, we introduce the concept of a (weak) minimizer of order k for a nonsmooth vector optimization problem over cones. Generalized classes of higher-order cone-nonsmooth (F, ρ)-convex functions are introduced and sufficient optimality results are proved involving these classes. Also, a unified dual is associated with the considered primal problem, and weak and strong duality results are established.