Multiagent Systems in Travel Planning

Abstract—Currently, as a result of globalization and otherfactors, the tourism market has remained high. Assisting thecustomer in order to satisfy their desires and constraints is amajor challenge in this sector. This paper presents an agent-basedsystem for developing a travel plan. In the system, according tothe preferences informed by tourists, several agents cooperate inorder to construct the travel plan that best suits the client needs. Atourism domain ontology was developed to enable communicationbetween the agents. The system relies on the assistance of thegreedy algorithm to trace the routes. A heuristic function wasdesigned to ensure the fulfillment of the customer goals.Keywords—multiagent; ontology; Travel Planning I. I NTRODUCTION It’s known that building a travel packet that satisfies themajority of desires and constraints of the clients is a toughtask. There are several decisions to be taken and restrictions tobe satisfied such as: displacements, schedules, time and costconstraints, accommodation restrictions, tourist attractions tobe visited, etc. Therefore, a growing interest for automated sys-tems that supports travel packets construction can be noticednowadays. The agent-based systems approach can be helpfulin this domain due to its characteristics: many autonomousagents (representatives from airlines, hotels, bus companies,tourist attractions) with their own goals interact to build atravel plan. To implement such a system is necessary to firstestablish a common vocabulary, using a lightweight ontology,so that agents can communicate. It is then necessary to designagents, the roles to be played and the interaction betweenagents in order to produce a travel plan that meets the usersneeds. This paper presents a Multiagent system (MAS) [1] forthe travel planning problem. A tourism domain ontology wasdeveloped to enable communication between the agents. Thesystem relies on the assistance of the greedy algorithm to tracethe routes and the information required for the plan creationis located in a relational database. A heuristic function wasdesigned to ensure the fulfillment of the customer goals.II. R