This paper studies issues concerning parallel execution of regular Fortran DO loops on an asynchronous shared-memory multiprocessor, where each iteration is the basic unit to be executed by a single processing element. An iteration is a dependent predecessor of another iteration if execution of the latter iteration has to wait until execution of the former iteration has completed. During the ex...

We consider implicit integration methods for the numerical solution of sti initial-value problems. In applying such methods, the implicit relations are usually solved by Newton iteration. However, it often happens that in subintervals of the integration interval the problem is nonsti or mildly sti with respect to the stepsize. In these nonsti subintervals, we do not need the (expensive) Newton ...

Finite Element problems are often solved using multigrid techniques. The most time consuming part of multigrid is the iterative smoother, such as Gauss-Seidel. To improve performance, iterative smoothers can exploit parallelism, intra-iteration data reuse, and inter-iteration data reuse. Current methods for parallelizing Gauss-Seidel on irregular grids, such as multi-coloring and ownercomputes ...

In our papers [Inverse Problems, 15, 309-327,1999] and [Numer. Math., 88, 347-365, 2001] we proposed algorithm REGINN being an inexact Newton iteration for the stable solution of nonlinear ill-posed problems. REGINN consists of two components: the outer iteration, which is a Newton iteration stopped by the discrepancy principle, and an inner iteration, which computes the Newton correction by so...

We describe a point-based policy iteration (PBPI) algorithm for infinite-horizon POMDPs. PBPI replaces the exact policy improvement step of Hansen’s policy iteration with point-based value iteration (PBVI). Despite being an approximate algorithm, PBPI is monotonic: At each iteration before convergence, PBPI produces a policy for which the values increase for at least one of a finite set of init...

That our mental lives are not transparent to us is one of the few Freudian doctrines that has been robustly confirmed by subsequent psychological research. Nevertheless, much work in the philosophy of language, as well as formal work in economics and computer science, assumes various “iteration principles” that can seem to conflict with this psychological truism. By “iteration principles” I hav...

This paper provides a comparison of three looping operations: Kleene’s repetition (‘star’), Elgot’s iteration (‘dagger’) and feedback (‘up arrow’). The comparison is based on an algebraic study of an algebra for flowgraphs (this is a generic name for digraph models like: automata, nets, flowchart schemes, etc.), called biflow. Equivalent presentations of biflows and biflows over algebraic or ma...

In reinforcement learning, least-squares temporal difference methods (e.g., LSTD and LSPI) are effective, data-efficient techniques for policy evaluation and control with linear value function approximation. These algorithms rely on policy-dependent expectations of the transition and reward functions, which require all experience to be remembered and iterated over for each new policy evaluated....

In this paper we consider approximate policy-iteration-based reinforcement learning algorithms. In order to implement a flexible function approximation scheme we propose the use of non-parametric methods with regularization, providing a convenient way to control the complexity of the function approximator. We propose two novel regularized policy iteration algorithms by addingL-regularization to...