Modular composition of intelligent control policies for FMS models

The scheduling problem in flexible manufacturing systems (FMSs) is addressed. A rule-based scheduling system with the following main characteristics: it uses an opportunistic and/or hierarchical schema; it has four operative levels there exist some rules to choose the most suitable level; and the information needed by the scheduling system is obtained during FMS model stage. The authors' approach mainly consists of a library of Petri net (PN) models of FMS elements, called objects; and some primitives to relate these objects. Objects have associated information which is used to make decisions when it is necessary. This information is retrieved and grouped during the modeling stage to obtain more information about the problem. The decision making system is implemented on rules. An example is given to evaluate the performance of the decision making system.<
>