Abstract
Up to now, the design and construction of SCS (Sequential Control Systems) have been performed by human experts. The Petri net modeling, which would be used as the formal specification in SCS is effective. As the result of the research and development of Petri net modeling, applications to FA, distributed discrete control systems, hardware and software of controllers, and simulation of SCS, have been reported. As the acquisition of knowledge is a main bottleneck now, computer aided design and construction of SCS is difficult.
In this paper, using a Petri-net-based representation to specify SCS, we distinguish the abstract level of knowledge from the concrete level. We explain then, the concept acquisition in SCS by learning. Utilizing a concept's concrete knowledge, we make up a set of relations having partial order. From this set the minimal element, maximal element, and middle element obtained by a set partition may be acquired. In order to show the efficiency of the method developed in this paper, we discuss an example.
