Implementation of a Cooperative Agent Architecture based on the Language-Action Perspective
E. Verharen, Frank P. M. Dignum, Sander Bos · 1997
. In this paper the architecture and implementation of Cooperative Information Agents (CIA) is described. Taking a language-action perspective to the design of CIAs allows for the specification of obligations and authorizations, and results in the separation of tasks (things the agent must do) and contracts (mutually agreed commitments to the course of communication). The architecture describes the functional components of a CIA: task manager (responsible for managing the agenda), contract manager (managing and negotiating contracts), communication manager (responsible for all external communication), and service execution manager (managing the execution of actions). The prototype agents showhow a formal logical theory for communicating agents can be used as a sound basis for an actual implementation. 1 Introduction Traditionally an information system (IS) was considered as one central database and a set of users accessing the database through application programs or directly via an S...