Article ID: | iaor1992359 |
Country: | United States |
Volume: | 7 |
Issue: | 4 |
Start Page Number: | 291 |
End Page Number: | 319 |
Publication Date: | Sep 1990 |
Journal: | Simulation Transactions |
Authors: | Lee C., Zeigler B.P., Kim T.G. |
Keywords: | artificial intelligence |
A family of models is defined as a set of variants generated by applying admissible structure transformations to a nominal model. In variable structure modelling methodology, a model is able to select successive variants for its own structure depending on prevailing behavioral circumstances. Such a methodology has application to a variety of important problem areas. Knowledge representation and decision schemes are discussed for mediating the selection of the next variant and its realization in a structure state transition. Since such structure transformation must be locally implemented, the schemes involve distributed and cooperative interaction among model components. This approach is illustrated in the context of an adaptive computer architecture simulation model.