Person:
Isbell, Charles L.

Permanent Link
https://hdl.handle.net/1853/71360
Associated Organization(s)
Organizational Unit
ORCID
ArchiveSpace Name Record

Filters

2006 - 2007 4
4
4
4
4
Reset filters

Settings
search.filters.applied.f.isSeriesOfPublicationTitle: College of Computing Technical Report Series ×

Publication Search Results

Now showing 1 - 4 of 4
  • Thumbnail Image
    Item
    Autonomous Nondeterministic Tour Guides: Improving Quality of Experience with TTD-MDPs
    (Georgia Institute of Technology, 2007) Cantino, Andrew S. ; Roberts, David L. ; Isbell, Charles L.
    In this paper, we address the problem of building a system of autonomous tour guides for a complex environment, such as a museum with many visitors. Visitors may have varying preferences for types of art or may wish to visit different areas across multiple visits. Often, these goals conflict. For example, many visitors may wish to see the museum's most popular work, but that could cause congestion, ruining the experience. Thus, our task is to build a set of agents that can satisfy their visitors' goals while simultaneously providing quality experiences for all. We use targeted trajectory distribution MDPs (TTD-MDPs), a technology developed to guide players in an interactive entertainment setting. The solution to a TTD-MDP is a probabilistic policy that results in a specific distribution of trajectories through a state space. We motivate TTD-MDPs for the museum tour problem, then describe the development of a number of models of museum visitors. Additionally, we propose a museum model and simulate tours using personalized TTD-MDP tour guides for each kind of visitor. We explain how the use of probabilistic policies reduces the congestion experienced by visitors while preserving their ability to pursue and realize goals.
  • Thumbnail Image
    Item
    Horizon-based Value Iteration
    (Georgia Institute of Technology, 2007) Zang, Peng ; Irani, Arya ; Isbell, Charles L.
    We present a horizon-based value iteration algorithm called Reverse Value Iteration (RVI). Empirical results on a variety of domains, both synthetic and real, show RVI often yields speedups of several orders of magnitude. RVI does this by ordering backups by horizons, with preference given to closer horizons, thereby avoiding many unnecessary and incorrect backups. We also compare to related work, including prioritized and partitioned value iteration approaches, and show that our technique performs favorably. The techniques presented in RVI are complementary and can be used in conjunction with previous techniques. We prove that RVI converges and often has better (but never worse) complexity than standard value iteration. To the authors’ knowledge, this is the first comprehensive theoretical and empirical treatment of such an approach to value iteration.
  • Thumbnail Image
    Item
    Accord: Middleware Support for Contextual, Ubiquitous Data Management on User Devices
    (Georgia Institute of Technology, 2006) Cooper, Brian F. ; Isbell, Charles L. ; Pierce, Jeffrey S. ; Roberts, David L. ; Bhat, Sooraj
    People increasingly use a diverse array of computational devices, including desktop PCs, one or more laptops, a cell phone, a PDA, tablet PCs, digital music players, automobile computers, and so on. We present Accord, a middleware system we have implemented to manage user data across all of these devices. Accord emulates an ideal abstraction we call a user data-space: a virtual space in which user files exist independent of any particular physical device. Users put files into the space with whatever device is convenient, and later access those files using any of their devices. This abstraction is difficult to implement, and requires Accord to predict when a file will be needed and on which device. We describe two mechanisms the middleware uses to support such predictions: an object graph, which records contextual and statistical information about file objects, and a file transfer planner, which uses predictions to determine how to efficiently move files between devices despite connectivity, bandwidth and storage constraints. Predictions can be constructed using simple usage statistics, or from more complex machinelearned models of user activities. We also present experimental results demonstrating the effectiveness of our system.
  • Thumbnail Image
    Item
    A Decision-Theoretic Approach to File Consistency in Constrained Peer-to-Peer Device Networks
    (Georgia Institute of Technology, 2006) Roberts, David L. ; Bhat, Sooraj ; Isbell, Charles L. ; Cooper, Brian F. ; Pierce, Jeffrey S.
    As users interact with an increasing array of personal computing devices, maintaining consistency of data across those devices becomes significantly more difficult. Typical solutions assume either access to centralized servers, continual connectivity, or unbounded storage and CPU capacity. In practice, users own devices with widely varying processing and storage capabilities that use intermittent or sparsely-connected networks and incur (often asymmetric) transfer costs. We identify the conditions that enable the seamless management of a user's data across devices and present a multi-agent system built upon a decision-theoretic approach to constructing and executing multiple plans to achieve consistency in a peer-to-peer, partially observable, non-deterministic environment. We analyze the performance of these plans in comparison to a standard epidemic replication algorithm used in many database consistency applications.