論文

We consider a class of difference-of-convex (DC) optimization problems where the objective function is the sum of a smooth function and a possibly nonsmooth DC function. The application of proximal DC algorithms to address this problem class is well-known. In this paper, we combine a proximal DC algorithm with an inexact proximal Newton-type method to propose an inexact proximal DC Newton-type method. We demonstrate global convergence properties of the proposed method. In addition, we give a memoryless quasi-Newton matrix for scaled proximal mappings and consider a two-dimensional system of semi-smooth equations that arise in calculating scaled proximal mappings. To efficiently obtain the scaled proximal mappings, we adopt a semi-smooth Newton method to inexactly solve the system. Finally, we present some numerical experiments to investigate the efficiency of the proposed method, which show that the proposed method outperforms existing methods.

In this paper, we consider an active-set algorithm for solving large-scale bound constrained optimization problems. First, by incorporating a restart technique, we modify the active-set strategy by Yuan and Lu (2011) and combine it with the memoryless quasi-Newton method based on a modified spectral-scaling Broyden family. Then, we propose an algorithm of our method with the framework of the Armijo line search, and show its global convergence. Finally, we illustrate some numerical experiments to investigate how the parameter choice in our method affects numerical performance.

An important and often researched area of management science is mathematical modeling of a supply chain. Competitive situations can occur in supply chains owing to the involvement of multiple decision makers (players) that independently decide their behaviors. To investigate competitive supply chain networks, a supply chain network equilibrium (SCNE) model was proposed. Recently, particular attention has been paid to risk management of a supply chain. In equilibrium models, it is vital to consider players' decisions and interdependence relations. Thus, we consider competitive supply chain networks with uncertainties in the other players' strategies. In the proposed model, each player cannot know exactly the other players' strategies, and they decide their strategy using the minimax principle (that is, assuming the worst case). We call it the robust SCNE model. We formulate the robust SCNE model as a variational inequality problem (VIP) in which the set associated with the VIP is constructed by second-order cone constraints. We show the existence and uniqueness of the equilibrium under mild assumptions. In addition, we give, in a special case, some relations between players' strategies in the equilibrium and magnitudes of uncertainties. Finally, some numerical results are provided.

Memoryless quasi-Newton methods are studied for solving largescale unconstrained optimization problems. Recently, memoryless quasi-Newton methods based on several kinds of updating formulas were proposed. Since the methods closely related to the conjugate gradient method, the methods are promising. In this paper, we propose a memoryless quasi-Newton method based on the Broyden family with the spectral-scaling secant condition. We focus on the convex and preconvex classes of the Broyden family, and we show that the proposed method satisfies the sufficient descent condition and converges globally. Finally, some numerical experiments are given.

This study considers a proximal Newton-type method to solve the minimization of a composite function that is the sum of a smooth nonconvex function and a nonsmooth convex function. In general, the method uses the Hessian matrix of the smooth portion of the objective function or its approximation. The uniformly positive definiteness of the matrix plays an important role in establishing the global convergence of the method. In this study, an inexact proximal memoryless quasi-Newton method is proposed based on the memoryless Broyden family with the modified spectral scaling secant condition. The proposed method inexactly solves the subproblems to calculate scaled proximal mappings. The approximation matrix is shown to retain the uniformly positive definiteness and the search direction is a descent direction. Using these properties, the proposed method is shown to have global convergence for nonconvex objective functions. Furthermore, the R-linear convergence for strongly convex objective functions is proved. Finally, some numerical results are provided.

We consider proximal gradient methods for minimizing a composite function of a differentiable function and a convex function. To accelerate the general proximal gradient methods, we focus on proximal quasi-Newton type methods based on proximal mappings scaled by quasi-Newton matrices. Although it is usually dificult to compute the scaled proximal mappings, applying the memoryless symmetric rank-one (SR1) formula makes this easier. Since the scaled (quasi-Newton) matrices must be positive definite, we develop an algorithm using the memoryless SR1 formula based on a modified spectral scaling secant condition. We give the subsequential convergence property of the proposed method for general objective functions. In addition, we show the R-linear convergence property of the method under a strong convexity assumption. Finally, some numerical results are reported.

We consider iterative methods for unconstrained optimization on Riemannian manifolds. Though memoryless quasi-Newton methods are effective for large-scale unconstrained optimization in the Euclidean space, they have not been studied over Riemannian manifolds. Therefore, in this paper, we propose a memoryless quasi-Newton method in Riemannian manifolds. The proposed method is based on the spectral-scaling Broyden family with additional modifications to ensure the sufficient descent condition. We present an algorithm for the proposed method that uses the Wolfe line search conditions and show that this algorithm guarantees global convergence. We emphasize that global convergence is guaranteed without any assumptions regarding the convexity of the objective function or the isometric property of the vector transport. In addition, we derive appropriate selections for the parameter vector contained in the proposed method. Numerical experiments are conducted to compare the proposed method with conventional conjugate gradient methods using typical test problems. The results show that the proposed method is superior to the tested conjugate gradient methods.

in press