Results 1  10
of
15
Temporal Abstraction in Reinforcement Learning
, 2000
"... Decision making usually involves choosing among different courses of action over a broad range of time scales. For instance, a person planning a trip to a distant location makes highlevel decisions regarding what means of transportation to use, but also chooses lowlevel actions, such as the moveme ..."
Abstract

Cited by 59 (2 self)
 Add to MetaCart
Decision making usually involves choosing among different courses of action over a broad range of time scales. For instance, a person planning a trip to a distant location makes highlevel decisions regarding what means of transportation to use, but also chooses lowlevel actions, such as the movements for getting into a car. The problem of picking an appropriate time scale for reasoning and learning has been explored in artificial intelligence, control theory and robotics. In this dissertation we develop a framework that allows novel solutions to this problem, in the context of Markov Decision Processes (MDPs) and reinforcement learning. In this dissertation, we present a general framework for prediction, control and learning at multipl...
Automated Modeling of Complex Systems to Answer Prediction Questions
 ARTIFICIAL INTELLIGENCE
, 1995
"... ..."
Studies of the Accuracy of Time Integration Methods for ReactionDiffusion Equations
 J. COMPUT. PHYS
, 2005
"... In this study we present numerical experiments of time integration methods applied to systems of reactiondi#usion equations. Our main interest is in evaluating the relative accuracy and asymptotic order of accuracy of the methods on problems which exhibit an approximate balance between the compe ..."
Abstract

Cited by 17 (3 self)
 Add to MetaCart
In this study we present numerical experiments of time integration methods applied to systems of reactiondi#usion equations. Our main interest is in evaluating the relative accuracy and asymptotic order of accuracy of the methods on problems which exhibit an approximate balance between the competing component time scales. Nearly balanced
A parallel block multilevel preconditioner for the 3d incompressible navierstokes equations
 J. Comput. Phys
, 2003
"... Abstract. The development of robust and efficient algorithms for both steadystate simulations and fullyimplicit time integration of the Navier–Stokes equations is an active research topic. To be effective, the linear subproblems generated by these methods require solution techniques that exhibit r ..."
Abstract

Cited by 13 (2 self)
 Add to MetaCart
Abstract. The development of robust and efficient algorithms for both steadystate simulations and fullyimplicit time integration of the Navier–Stokes equations is an active research topic. To be effective, the linear subproblems generated by these methods require solution techniques that exhibit robust and rapid convergence. In particular, they should be insensitive to parameters in the problem such as mesh size, time step, and Reynolds number. In this context, we explore a parallel preconditioner based on a block factorization of the coefficient matrix generated in an Oseen nonlinear iteration for the primitive variable formulation of the system. The key to this preconditioner is the approximation of a certain Schur complement operator by a technique first proposed by Kay, Loghin, and Wathen [26] and Silvester, Elman, Kay, and Wathen [46]. The resulting operator entails subsidiary computations (solutions of pressure Poisson and convection–diffusion subproblems) that are similar to those required for decoupled solution methods; however, in this case these solutions are applied as preconditioners to the coupled Oseen system. One important aspect of this approach is that the convection–diffusion and Poisson subproblems are significantly easier to solve than the entire coupled system, and a solver can be built using tools developed for the subproblems. In this paper, we apply smoothed aggregation algebraic multigrid to both subproblems. Previous work has focused on demonstrating the optimality of these preconditioners with respect to mesh size on serial, twodimensional, steadystate computations employing geometric multigrid methods; we focus on extending these methods to largescale, parallel, threedimensional, transient and steadystate simulations employing algebraic multigrid (AMG) methods. Our results display nearly optimal convergence rates for steadystate solutions as well as for transient solutions over a wide range of CFL numbers on the twodimensional and threedimensional liddriven cavity problem. 1. Introduction. Recently
Edinburgh EH9 3JZ,
, 1993
"... We compare the results for the dynamical evolution of star clusters derived from anisotropic gaseous models with the data from Nbody simulations of isolated and onecomponent systems, each having modest number of stars. The statistical quality of Nbody data was improved by averaging results from man ..."
Abstract
 Add to MetaCart
We compare the results for the dynamical evolution of star clusters derived from anisotropic gaseous models with the data from Nbody simulations of isolated and onecomponent systems, each having modest number of stars. The statistical quality of Nbody data was improved by averaging results from many Nbody runs, each with the same initial parameters but with different sequences of random numbers used to initialize positions and velocities of the particles. We study the development of anisotropy, the spatial evolution and energy generation by threebody binaries and its Ndependence. We estimate the following free parameters of anisotropic gaseous models: the time scale for collisional anisotropy decay and the coefficient in the formulae for energy generation by threebody binaries. To achieve a fair agreement between Nbody and gaseous models for the core in pre as well as in postcollapse only the energy generation by binaries had to be varied by N. We find that anisotropy has considerable influence on the spatial structure of the cluster particularly for the intermediate and outer regions.
Solitons and wavelets: Scale analysis and bases
, 2000
"... We use a onescale similarity analysis which gives specific relations between the velocity, amplitude and width of localized solutions of nonlinear differential equations, whose exact solutions are generally difficult to obtain. We also introduce kinkantikink compact solutions for the nonlinearnon ..."
Abstract
 Add to MetaCart
We use a onescale similarity analysis which gives specific relations between the velocity, amplitude and width of localized solutions of nonlinear differential equations, whose exact solutions are generally difficult to obtain. We also introduce kinkantikink compact solutions for the nonlinearnonlinear dispersion K(2,2) equation, and we construct a basis of scaling functions similar with those used in the multiresolution analysis. These approaches are useful in describing nonlinear structures and patterns, as well as in the derivation of the time evolution of initial data for nonlinear equations with finite wavelength soliton solutions. PACS numbers: 42.30.Sy, 02.30.Px, 43.35.+d, 47.20.Ky, 05.45.a, 1
Dynamical Evolution of Rotating Stellar Systems: III. The Effect of Mass Spectrum
, 2004
"... Since the pioneering studies of Lupton & Gunn (1987) and its subsequent applications to fit actual globular cluster observations it is evident that galactic globular clusters exhibit some degree of rotation, and that there is also a consistent ..."
Abstract
 Add to MetaCart
Since the pioneering studies of Lupton & Gunn (1987) and its subsequent applications to fit actual globular cluster observations it is evident that galactic globular clusters exhibit some degree of rotation, and that there is also a consistent
Tidally limited star clusters – comparing gaseous models with other techniques
, 2008
"... Dynamical modelling of globular clusters and other collisional stellar systems (like galactic nuclei, rich open clusters, and rich galaxy clusters) still poses a considerable challenge for both theory and computational requirements (in hardware and software). On the ..."
Abstract
 Add to MetaCart
Dynamical modelling of globular clusters and other collisional stellar systems (like galactic nuclei, rich open clusters, and rich galaxy clusters) still poses a considerable challenge for both theory and computational requirements (in hardware and software). On the