We consider the problem of minimizing a convex function over a subset of R that is not necessarily convex (minimization of a convex function over the integer points in a polytope is a special case). We define a family of duals for this problem and show that, under some natural conditions, strong duality holds for a dual problem in this family that is more restrictive than previously considered ...

Abstract. Coherent unit actions on a binary, quadratic operad were introduced by Loday and were shown by him to give Hopf algebra structures on the free algebras when the operad is also regular with a splitting of associativity. Working with such operads, we characterize coherent unit actions in terms of linear equations of the generators of the operads. We then use these equations to give all ...

We study the moduli space M(6; 3, 6, 4) of simple rank 6 vector bundles E on P with Chern polynomial 1 + 3t + 6t+4t and properties of these bundles, especially we prove some partial results concerning their stability. We first recall how these bundles are related to the construction of sextic nodal surfaces in P having an even set of 56 nodes (cf. [Ca-To]). We prove that there is an open set, c...

This paper studies the problem of continuous time expected utility maximization of consumption together with addictive habit formation in general incomplete semimartingale financial markets. Introducing an auxiliary state processes and a modified dual space, we embed our original problem into an auxiliary time-separable utility maximization problem with the shadow random endowment. We establish...

By J. Douglas, Jr., R. B. Kellogg, and R. S. Varga We consider the iterative solution of the system of linear equations (1 ) (Xi + X-, + • ■ ■ + Xn)z = /, n 2; 2, where each Xj, 1 | j | n, is a Hermitian and positive definite N X N matrix. If n = 2, the iterative methods of Peaceman-Rachford [1, Chapter 7], or D'yakonov [2] and Kellogg [3], may be used to solve (1). In this paper these methods ...

We investigate novel, dual algorithms for dynamic programming and reinforcement learning, based on maintaining explicit representations of stationary distributions instead of value functions. In particular, we investigate the convergence properties of standard dynamic programming and reinforcement learning algorithms when they are converted to their natural dual form. Here we uncover advantages...

We model agents’ preferences by cash-invariant concave functionals defined on L∞, and formulate the optimal risk allocation problem as their infimal-convolution. We study the case of agents whose choice functionals are law-invariant with respect to different probability measures and show how, in this case, the value function preserves a desirable dual representation (equivalent to the Fatou pro...

This paper studies fully convex problems of Bolza in the conjugate duality framework of Rockafellar. We parameterize the problem by a general Borel measure which has a direct interpretation in certain problems of financial economics. We derive a dual representation for the optimal value function in terms of continuous dual trajectories and we give conditions for the existence of solutions. Comb...

The topic of this paper is the discrete-time l1-norm minimisation problem with convolution constraints. We find primal initial conditions for which the dual optimal solution is periodic. Periodicity of the dual optimal solution implies satisfaction of a simple linear recurrence relation by the primal optimal solution. c © 2007 Elsevier B.V. All rights reserved.

Generalized Chow forms were introduced by Cayley for the case of 3-space; their zero set on the Grassmannian G(1, 3) is either the set Z of lines touching a given space curve (the case of an ‘honest’ Cayley form), or the set of lines tangent to a surface. Cayley gave some equations for F to be a generalized Cayley form, which should hold modulo the ideal generated by F and by the quadratic equa...