Heuristic control system employing expert system, neural network and training pattern generating and controlling system

• US 5,465,204 A
• Filed: 11/05/1992
• Issued: 11/07/1995
• Est. Priority Date: 11/08/1991
• Status: Expired due to Term

• Current Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Original Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Inventors: Chang, Eric, Sekine, Masatoshi
• Primary Examiner(s): Gordon, Paul P.
• Application Number:
  US07/972,304
• Time in Patent Office: 1,097 Days
• Field of Search: 364/152, 364/490, 364/489, 390/20, 390/21, 390/23
• US Class Current: 700/32
• CPC Class Codes: G05B 13/029 using neural networks and expert systems
• CPC Class Codes: G06N 3/042 Knowledge-based neural networks; Logical representations of neural networks

Abstract

A heuristic control system and method for use with a computer-aided design, capable of learning complicated control and achieving an optimizing task with a fewer number of iteration. The heuristic control system includes: rule-based system for choosing and evaluating a rule among a plurality of given rules; training system for choosing strongly a rule whose evaluation result is favorable based on a predetermined value which evaluates an evaluation result of the rule-based system, and for generating a learning pattern; and neural network for designing an optimized circuit based on a signal fed from the training system and for sending a resultant signal to the rule-based system for another iteration of heuristic control. The learning method includes the steps of: choosing and evaluating a rule among a plurality of given rules; choosing strongly a rule whose evaluation result is favorable based on a predetermined value; generating a learning pattern which brings a desirable result based on the evaluating step; designing an optimized circuit based on the learning pattern generated; and choosing and evaluating iteratively a rule among a plurality of given rules.

Claims

1. A heuristic control system for use with a computer-aided design, comprising:

processing means including a predetermined learning pattern and a rule-invoking system;

evaluation means for evaluating an operational result of the processing means;

means for generating a training pattern based on a result obtained from the evaluation means; and

neural network means for learning to control the processing means by means of the training pattern supplied by the training pattern generating means.

2. The heuristic control system as recited in claim 1, wherein the training pattern generating means compares a plurality of evaluation results obtained by the evaluation means so as to choose a pattern with a favorable result.

3. A heuristic control system for use with a computer-aided design, comprising:

rule-based means for choosing and evaluating a rule among a plurality of given rules;

training means for choosing strongly a rule whose evaluation result is favorable based on a predetermined value which evaluates an evaluation result of the rule-based means, and for generating a learning pattern; and

neural network means for designing an optimized circuit based on a signal fed from the training means and for sending a resultant signal to the rule-based means for another iteration of heuristic control.

4. The heuristic control system as recited in claim 3, wherein the training means invokes and provides to the neural means an initial circuit when the evaluation results do not improve over a predetermined number of iteration.

5. The heuristic control system as recited in claim 3, wherein the neural network means includes a back-propagation system with multiple layers.

6. A learning method in computer-aided design, comprising the steps of

choosing and evaluating a rule among a plurality of given rules;

choosing strongly a rule whose evaluation result is favorable based on a predetermined value;

generating a learning pattern which brings a desirable result based on the evaluating step;

designing an optimized circuit based on the learning pattern generated; and

choosing and evaluating iteratively a rule among a plurality of given rules.

7. The method of claim 6, wherein in the learning pattern generating step includes providing an initial circuit when the evaluation results do not improve over a predetermined number of iterations.