From process transition networks to behavior automata

A process transition network (PTN) provides a method for specifying the behavior of multiple agents in terms of changes and interdependence among the changes of the states of individual agents. A way to compile a PTN into a more easily implemented structure is presented. A structure as such is called a behavior automaton (BA). The transformation from PTNs to BAs addresses, in a broad sense, the issues of relating the design to the implementation of an intelligent control system. With regard to the modeling of intelligent behavior, such a transformation suggests a closed loop in which the behavior specification of the concerned agents can be examined and incrementally improved by examining the validation of the PTNs. Also explored are ways in which PTNs can be used to specify behaviors at different abstraction levels