Fast convergent and computationally inexpensive policy evaluation is an essential part of reinforcement learning algorithms based on policy iteration. Algorithms such as LSTD, LSPE, FPKF and NTD, have faster convergence rates but they are computationally slow. On the other hand, there are algorithms that are computationally fast but with slower convergence rate, among them are TD, RG, GTD2 and ...

We report a new and efficient factorized algorithm for the determination of the adaptive compound mobility matrix B in a stochastic quasi-Newton method (S-QN) that does not require additional potential evaluations. For one-dimensional and two-dimensional test systems, we previously showed that S-QN gives rise to efficient configurational space sampling with good thermodynamic consistency [C. D....

Warped products provide a rich class of physically significant geometric objects. Warped product construction is an important method to produce a new metric with a base manifold and a fibre. We construct compact base manifolds with a positive scalar curvature which do not admit any non-trivial quasi-Einstein warped product, and non compact complete base manifolds which do not admit any non-triv...

Data Flow is a natural paradigm representing distributed programs able to express data and control dependencies between composing nodes. Our work concentrates on periodic Data-Flow programs having Hard Real-Time constraints. We propose a model and a string-based representation for a relevant class of Data-Flow programs and present a two-phase algorithm able to compute a schedule for a given set...

We provide criteria for a set of observed choices to be generated by a quasi-linear preference relation. These criteria are valid if choice sets are compact and downward closed and do not require preferences to be convex. We show that under the added assumption that choice sets can be represented by linear budget sets, our condition also implies the existence of a quasi-linear utility represent...

By a quasi-permutation matrix we mean a square matrix over the complex field C with non-negative integral trace. Thus every permutation matrix over C is a quasipermutation matrix. For a given finite group G, let p(G) denote the minimal degree of a faithful permutation representation of G (or of a faithful representation of G by permutation matrices), let q(G) denote the minimal degree of a fai...

By a quasi-permutation matrix we mean a square matrix over the complex field C with non-negative integral trace. Thus, every permutation matrix over C is a quasipermutation matrix. For a given finite group G, let p(G) denote the minimal degree of a faithful permutation representation of G (or of a faithful representation of G by permutation matrices), let q(G) denote the minimal degree of a fa...

For a scheme X, we construct a sheaf C of complexes on X such that for every quasi-compact open U ⊂ X, C(U) is quasi-isomorphic to the Hochschild complex of U [11]. Since C is moreover acyclic for taking sections on quasi-compact opens, we obtain a local to global spectral sequence for Hochschild cohomology if X is quasi-compact.

In this paper some Newton and quasi-Newton algorithms for the solution of inequality constrained minimization problems are considered. All the algorithms described produce sequences fx k g converging q-superlinearly to the solution. Furthermore , under mild assumptions, a q-quadratic convergence rate in x is also attained. Other features of these algorithms are that the solution of linear syste...