Arrow Research search

Author name cluster

Dave McArthur

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

2 papers
1 author row

Possible papers

2

AAAI Conference 1982 Conference Paper

A Framework for Distributed Problem Solving

  • Dave McArthur

Situations in which several agents must interact to achieve goals present difficulties of coordination and cooperation not found in single-agent problem solving contexts. Techniques for coordination and cooperation required in group problem solving are not well understood because most AI models deal with cases in which problems are solved by a single agent. In this paper we present a framework for distributed problem solving that describes some of the expertise an agent working in a multi-agent environment must have. An application of the framework to the domain of air-traffic control is discussed. Here each aircraft is viewed as an agent that must cooperate with others to achieve a conflict-free plan.

PDF Details

IJCAI Conference 1981 Conference Paper

Tuning of Search of the Problem Space for Geometry Proofs

  • AUTOPILOT: A Distributed Planner for Air Fleet Control Perry W. Thorndyke
  • Dave McArthur

In planning a proof, a student searches through a space of inferences leading forward from the givens of the problem and backward from the to-be-proven statement. One dimension of growth of expertise is that students become more tuned in the search of this problem space. This can be shown to result from the application of various learning operators to production embodiments of the inference rules. Rules are evaluated after the solution of a problem according to whether they led to or led away from the solution. Rules that contributed to a solution are strengthened and an attempt is made to formulate general versions of these rules that will apply in other situations. Rules that led away from the solution are weakened and a discrimination process is evoked to try to add features to the rules that will try to restrict them to the correct circumstances of application. Composition is a learning process that collapses successful sequences of rule operations into single macro-rule productions. There is also a process that converts the backward reasoning rules formed by composition into forward reasoning rules. The effect of these learning processes is to put into production conditions tests for problem features that are heunstically predictive of the rule's success.

PDF