Inter-agent cooperative communication method using TupleSpace for guiding users to alternative actions

Abstract

We propose an agent construction method that can identify action models from user action histories, propose alternative actions if users encounter situations different from their normal action models, and guide the users to those alternative actions. In addition, to realize an alternative action guidance service, we present a selection of terminal agents and network agents as well as an allocation of their functions and roles. Furthermore, we propose a steady-action model definition for realizing the service, a tuple classification and element constructions for TupleSpace-based inter-agent cooperative communication, and an alternative-action extraction algorithm. Next, we present a prototype system-based operation check of inter-agent cooperative communication and an evaluation of tuple-matching processing time. © Springer-Verlag Berlin Heidelberg 2011.