Propose and Revise

  • Frank Puppe


The propose-and-revise strategy is suitable for deriving values for a set of parameters. It makes possible an efficient, sequential determination of the parameter values, even when the search space is locally or globally under- or over-determined or when cyclic relationships exist. Its flexibility results from the fact that the parameter values first proposed can be corrected at any time with “revision knowledge” if this turns out to be necessary. It is suitable for cyclic parameter relationships because loops can be broken by taking a cyclic parameter as a starting parameter and estimating its value. If a contradiction occurs in the subsequent calculations on account of the cyclic relations, the parameter value is raised or lowered (Fig. 23.1).


Knowledge Acquisition User Requirement Rule Interpreter Plausibility Check Revision Rule 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. Hartinger, M. and Wedel, T.: Stand des Expertensystems PAREX-CO zur wissensbasierten dynamischen Konfiguration von Parametern des PPS-Modularprogramms COPICS, Memo 1/1990, University of Erlangen-Nuremberg, Department of Business Computer Science, 1990.Google Scholar
  2. Marcus, S.: SALT: A Knowledge Acquisition Tool for Propose-and-Revise Systems, in Marcus, S. (ed.): Automating Knowledge Acquisition for Expert Systems, Chap. 4, Kluwer Academic Publishers, 1988 (a).Google Scholar
  3. Marcus, S., Stout, J. and McDermott, J.: VT: An Expert Elevator Designer That Uses Knowledge-Based Backtracking, AI Magazine 9, 95–111, 1988 (b).Google Scholar
  4. Marcus, S. and McDermott, J.: SALT: A Knowledge Acquisition Language for Propose-and-Revise Systems, AI Journal 39, 1–38, 1989.MATHGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1993

Authors and Affiliations

  • Frank Puppe
    • 1
  1. 1.Institut für Informatik Lehrstuhl für Künstliche Intelligenz Am HublandUniversität WürzburgWürzburgGermany

Personalised recommendations