The high level of automation, the high level of interrelationship betw
een components, and the productivity demands to which automated manufa
cturing systems are subjected are causing considerable increases in bo
th the frequency and significance of faults and, as a consequence of t
his, the real availability of machines is low, the product quality une
ven, and the reliability of systems insufficient. The simplicity, prec
ision, and generality of the graphs mean that they are very useful too
ls in the representation of manufacturing systems. At the same time, a
ny system modelled by means of graphs is an ideal candidate for coding
with AI tools. The physical and functional models of manufacturing sy
stems can be represented by means of petri nets, status diagrams, deci
sion trees, logic graphs, grafcets, fault diagnosis graphs, and so on.
In this work, the results of two projects initiated by the European C
ommunity are presented: ESPRIT 504 (1985-1988) and ESPRIT 2349 (1989-1
992).