‎In this paper‎, ‎some KKT type sufficient global optimality conditions‎ ‎for general mixed integer nonlinear programming problems with‎ ‎equality and inequality constraints (MINPP) are established‎. ‎We achieve‎ ‎this by employing a Lagrange function for MINPP‎. ‎In addition‎, ‎verifiable sufficient global optimality conditions for general mixed‎ ‎integer quadratic programming problems are der...

By simply transforming the quadratic matrix equation into an equivalent fixed-point equation, we construct a successive approximation method and a Newton’s method based on this fixed-point equation. Under suitable conditions, we prove the local convergence of these two methods, as well as the linear convergence speed of the successive approximation method and the quadratic convergence speed of ...

We survey a number of algorithms for the simple stochastic game problem, which is to determine the winning probability of a type of stochastic process, where the transitions are partially controlled by two players. We show that four natural approaches to solving the problem are incorrect, and present two new algorithms for the problem. The rst reduces the problem to that of nding a locally opti...

Let F be a subset of the n-dimensional Euclidean space R n represented in terms of a compact convex subset C 0 and a set P F of nitely or innnitely many quadratic functions on R n such that F = fx 2 C 0 : p(x) 0 (8p() 2 P F)g. In this paper, we investigate some fundamental properties related to the nite convergence of the successive SDP (semideenite programming) relaxation method proposed by th...

Successive overrelaxation (SOR) for symmetric linear complementarity problems and quadratic programs is used to train a support vector machine (SVM) for discriminating between the elements of two massive datasets, each with millions of points. Because SOR handles one point at a time, similar to Platt's sequential minimal optimization (SMO) algorithm which handles two constraints at a time and J...

In Dynamic Programming, convergence of algorithms such as Value Iteration or Policy Iteration results -in discounted problemsfrom a contraction property of the back-up operator, guaranteeing convergence to its fixedpoint. When approximation is considered, known results in Approximate Policy Iteration provide bounds on the closeness to optimality of the approximate value function obtained by suc...

A function on R with multiple local minima is approximated from below, via linear programming, by a linear combination of convex kernel functions using sample points from the given function. The resulting convex kernel underestimator is then minimized, using either a linear equation solver for a linear-quadratic kernel or by a Newton method for a Gaussian kernel, to obtain an approximation to a...

MAP inference for general energy functions remains a challenging problem. While most efforts are channeled towards improving the linear programming (LP) based relaxation, this work is motivated by the quadratic programming (QP) relaxation. We propose a novel MAP relaxation that penalizes the Kullback-Leibler divergence between the LP pairwise auxiliary variables, and QP equivalent terms given b...

We present a branch and cut algorithm that yields in nite time, a globally -optimal solution (with respect to feasibility and optimality) of the nonconvex quadratically constrained quadratic programming problem. The idea is to estimate all quadratic terms by successive linearizations within a branching tree using Reformulation-Linearization Techniques (RLT). To do so, four classes of linearizat...