Abstract
Recommending a service to a travel group is often not a trivial task. In order to produce a satisfactory recommendation result, one not only has to compromise every user’s preferences and constraints but also satisfy some criterion to achieve the goals of travel group. In this paper, we demonstrate the design of a restaurant recommendation system in a multi-agent environment. First, the client agent could construct the user preference model for every user and evaluate all recommendation items in the database based on this model. Then all client agents would send the evaluation results as the users’ public preference to the recommendation system. And the system will integrate all information and put it together as the group preference. Finally the system will analyze the information stored in the group preference model, determine the criterion for the travel group, and search for the best solution in the database based the criterion to satisfy multi-users’ preferences in the travel group. We will illustrate the system with scenarios to recommend the restaurants satisfying different criteria for a travel group in Taipei city.