Finding and sharing common actions for distributed agents' collaboration

Abstract

This paper develops the description for agents to collaborate each other by sharing common actions and the way to find such common actions by distributed agents. Agents can collaborate each other and reduce useless actions by sharing common actions among their actions, but it is hard to find the common actions among their actions. Hence, we develop a description to find common actions among agents. We define the processes to find common actions based on /spl pi/-calculus. In the description, actions are described as the parallel processes through networks and resulted in the sets. The sets are the target sets including the intersection of some sets. Some sets have common actions and the others have uncommon actions by excluding common actions. Each agent will perform his own uncommon actions, which exclude common actions, and the part of the common actions. Moreover, we explain the descriptions to distribute actions among agents. They are constructed by the independent processes in /spl pi/-calculus. Thus, the leader of a group, who uses the descriptions, distributes actions among agents in parallel. Finally, we show the operational semantics of the descriptions.