CONTROL AND SCALING OF DISTRIBUTED AGENT SYSTEMS
msra(2007)
摘要
EXECUTIVE SUMMARY The Oregon Graduate Institute is pleased to submit this proposal for a 3-year effort to research "Control and Scaling of Distributed Agent Systems." The proposed research addresses two major challenges in developing a Tactical Picture Agent -- a system that can provide the user with an understanding of what the TPA will actually do, and techniques for addressing scalability in terms of number of users, information requests and information sources. The issue of comprehensibility is addressed by our providing an easy to use interface for creating and simulating a collection of "standing information requests" -- conditional notification and actions that are triggered by a combination of events in the data space. Regarding issue of scale, we will extend techniques from distributed databases and physical query optimization to reason about the physical properties (bandwidths, latencies, sheer size) of information sources in order to formulate efficient retrievals and monitoring of data. The proposed 3-year effort is estimated to cost $750,000.
更多查看译文
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络