single-rb.php

JRM Vol.8 No.5 pp. 454-458
doi: 10.20965/jrm.1996.p0454
(1996)

Paper:

Acquiring Objective Functions in Distributed Rule-Based Systems from Examples

Kenichi Matsuura and Yukinori Kakazu

Faculty of Engineering, Hokkaido University, Nishi-8, Kita-13, Kita-ku, Sapporo, 060 Japan

Received:
June 13, 1996
Accepted:
August 10, 1996
Published:
October 20, 1996
Keywords:
Distributed problem solving, Machine learning, Traveling salesman problem
Abstract

There are some great features in distributed problem solving systems, such as fault tolerance, robustness and so on. This system performs problem solving with search depending on an objective function. Distributed rulebased problem systems are considered to be of the same type. That is to say, the set of rules and the objective function exist separately within the system. However, in distributed rule-based systems, a set of rules should hold the objective function. The system should have a set of rules only, and the objective function should exist within that set of rules. In this paper, our objective is to acquire the objective function of a distributed rule-based system. A rule generation mechanism analyzes some given examples and acquires strategies for problem solving to a set of rules. In this way, the set of rules of the examples class in the domain represents the objective function of that class in the domain. Therefore, a solution using those rules keeps the same features as the examples if the problem belongs to the examples class that generates the set of rules. The system implemented by this theory has been applied to the domain of traveling salesman problem. This system has generated a set of rules that has held the objective function of its domain.

Cite this article as:
Kenichi Matsuura and Yukinori Kakazu, “Acquiring Objective Functions in Distributed Rule-Based Systems from Examples,” J. Robot. Mechatron., Vol.8, No.5, pp. 454-458, 1996.
Data files:

*This site is desgined based on HTML5 and CSS3 for modern browsers, e.g. Chrome, Firefox, Safari, Edge, Opera.

Last updated on Oct. 22, 2021