A Virtual Shop Navigation System Based on the Multi-Agent Technology

摘要

This paper presents an application of the multi-agent approach for an adaptive navigation in a virtual shop. The designed multi-agent system, called VSN~a, proposes to a geographycally distant visitor, a view of the shop adapted to its profile and preferences. The system architecture includes four types of intelligent agents: an interface agent which has for key task to reason about the visitor behavior and to generate a set of products of his interest. A planner agent which aims to generate a navigating plan adapted to the visitor profile and preferences. An information agent which has to retrieve information in response to queries, to select appropriate information sources and to manage multiple accesses to them. A navigator agent which performs the navigating plan to guide the visitor in the virtual shop and collects all the visitor's behavior information.