Behavioural state machines: agent programming and engineering.
Peter Novák · 2009
ion: the functional view on the semantics of mental state transformer provides a powerful abstraction for BSM agent program. Mental state transformer facilitates encapsulation of a set of particular interdependent objects stored in agent’s knowledge bases together with the dynamics associated with them, a part of agent’s overall behaviour. The construct of mst thus provides a similar abstract concept for programming cognitive agents as does an object in the object oriented programming paradigm. In particular, the functional semantics allows to view agent programs as functions modifying its mental states, decomposed into (or composed of) compounds of other encapsulated subfunctions, possibly with a similar nested structure. The remainder of this dissertation provides a substance to the above claims. In particular the Chapters 4 and 6 subsequently discuss formal support for the process of specification refinement and design of BSM agent program. Both heavily rely on the functional view on subprograms, i.e., mental state transformers.