Please use this identifier to cite or link to this item:
doi:10.22028/D291-24887
Title: | Using integrated knowledge acquisition to prepare sophisticated expert plans for their re-use in novel situations |
Author(s): | Schmalhofer, Franz Bergmann, Ralf Kühn, Otto Schmidt, Gabriele |
Language: | English |
Year of Publication: | 1992 |
OPUS Source: | Kaiserslautern ; Saarbrücken : DFKI, 1992 |
SWD key words: | Künstliche Intelligenz Automatische Handlungsplanung Wissenserwerb |
DDC notations: | 004 Computer science, internet |
Publikation type: | Report |
Abstract: | Plans which were constructed by human experts and have been repeatedly executed to the complete satisfaction of some customer in a complex real world domain contain very valuable planning knowledge. In order to make this compiled knowledge re-usable for novel situations, a specific integrated knowledge acquisition method has been developed: First, a domain theory is established from documentation materials or texts, which is then used as the foundation for explaining how the plan achieves the planning goal. Secondly, hierarchically structured problem class definitions are obtained from the practitioners' highlevel problem conceptualizations. The descriptions of these problem classes also provide operationality criteria for the various levels in the hierarchy. A skeletal plan is then constructed for each problem class with an explanation-based learning procedure. These skeletal plans consist of a sequence of general plan elements, so that each plan element can be independently refined. The skeletal plan thus accounts for the interactions between the various concrete operations of the plan at a general level. The complexity of the planning problem is thereby factored in a domain-specific way and the compiled knowledge of sophisticated expert plans can be re-used in novel situations. |
Link to this record: | urn:nbn:de:bsz:291-scidok-36446 hdl:20.500.11880/24943 http://dx.doi.org/10.22028/D291-24887 |
Series name: | Research report / Deutsches Forschungszentrum für Künstliche Intelligenz [ISSN 0946-008x] |
Series volume: | 92-25 |
Date of registration: | 25-Jun-2011 |
Faculty: | SE - Sonstige Einrichtungen |
Department: | SE - DFKI Deutsches Forschungszentrum für Künstliche Intelligenz |
Collections: | SciDok - Der Wissenschaftsserver der Universität des Saarlandes |
Files for this record:
File | Description | Size | Format | |
---|---|---|---|---|
RR_92_25.pdf | 98,95 kB | Adobe PDF | View/Open |
Items in SciDok are protected by copyright, with all rights reserved, unless otherwise indicated.