The standard assumption for proving linear convergence of first order methods for smooth convex optimization is the strong convexity of the objective function, an assumption which does not hold for many practical applications. In this paper, we derive linear convergence rates of several first order methods for solving smooth non-strongly convex constrained optimization problems, i.e. involving an objective function with a Lipschitz continuous gradient that satisfies some relaxed strong convexity condition. In particular, in the case of smooth constrained convex optimization, we provide several relaxations of the strong convexity conditions and prove that they are sufficient for getting linear convergence for several first order methods such as projected gradient, fast gradient and feasible descent methods. We also provide examples of functional classes that satisfy our proposed relaxations of strong convexity conditions. Finally, we show that the proposed relaxed strong convexity conditions cover important applications ranging from solving linear systems, Linear Programming, and dual formulations of linearly constrained convex problems. IntroductionRecently, there emerges a surge of interests in accelerating first order methods for difficult optimization problems, for example the ones without strong convex objective function, arising in different applications such as data analysis [6] or machine learning [9]. Algorithms based on gradient information have proved to be effective in these settings, such as projected gradient and its fast variants [11], stochastic gradient descent [12] or coordinate gradient descent [18].

Dual decomposition is a powerful technique for deriving decomposition schemes for convex optimization problems with separable structure. Although the Augmented Lagrangian is computationally more stable than the ordinary Lagrangian, the prox-term destroys the separability of the given problem. In this paper we use another approach to obtain a smooth Lagrangian, based on a smoothing technique developed by Nesterov, which preserves separability of the problem. With this approach we derive a new decomposition method, called proximal center algorithm, which from the viewpoint of efficiency estimates improves the bounds on the number of iterations of the classical dual gradient scheme by an order of magnitude. This can be achieved with the new decomposition algorithm since the resulting dual function has good smoothness properties and since we make use of the particular structure of the given problem.

No abstract

No abstract

No abstract

No abstract

System performance for networks composed of interconnected subsystems can be increased if the traditionally separated subsystems are jointly optimized. Recently, parallel and distributed optimization methods have emerged as a powerful tool for solving estimation and control problems in large-scale networked systems. In this paper we review and analyze the optimization-theoretic concepts of parallel and distributed methods for solving coupled optimization problems and demonstrate how several estimation and control problems related to complex networked systems can be formulated in these settings. The paper presents a systematic framework for exploiting the potential of the decomposition structures as a way to obtain different parallel algorithms, each with a different tradeoff among convergence speed, message passing amount and distributed computation architecture. Several specific applications from estimation and process control are included to demonstrate the power of the approach.

In this paper, we propose a distributed algorithm for solving large-scale separable convex problems using Lagrangian dual decomposition and the interior-point framework. By adding self-concordant barrier terms to the ordinary Lagrangian, we prove under mild assumptions that the corresponding family of augmented dual functions is self-concordant. This makes it possible to efficiently use the Newton method for tracing the central path. We show that the new algorithm is globally convergent and highly parallelizable and thus it is suitable for solving large-scale separable convex problems.