Distributed Problem Solving: The Contract Net Approach

Abstract

The ongoing revolution in LSI technology is drastically reducing the cost of computer components, making multiple processor architectures economically viable. These architectures have the potential to provide several computational advantages over uniprocessor architectures, including speed, reliability, and efficient matching of available processing power to problem complexity. This has led to a search for problem solving methods which can exploit the new technology. This paper presents one approach to problem solving in such architectures. The paper proposes a model of a distributed problem solver which consists of a collection of processors connected with communications and control mechanisms that enable them to operate concurrently, and enable them to cooperate in solving complex problems. The term 'distributed' rather than 'parallel' is used to emphasize that the individual processors are loosely-coupled; that is, the time a processor node spends in communication is small with respect to the time it spends in computation.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Sep 01, 1978
Accession Number
ADA060795

Entities

People

  • Randall Davis
  • Reid G. Smith

Organizations

  • Stanford University

Tags

Communities of Interest

  • Biomedical
  • Energy and Power Technologies
  • Materials and Manufacturing Processes
  • Sensors

DTIC Thesaurus Topics

  • Artificial Intelligence
  • Automated Speech Recognition
  • Communications Protocols
  • Computer Communications
  • Computer Programming
  • Computer Science
  • Computers
  • Computing System Architectures
  • Contracts
  • Detectors
  • Electrical Engineering
  • Information Processing
  • Information Transfer
  • Language
  • Message Processing
  • Negotiations
  • Networks

Fields of Study

  • Computer science

Readers

  • Artificial Intelligence
  • Parallel and Distributed Computing.
  • Systems Analysis and Design