Abstract
Tour planning involves the detail scheduling of scenic spots, accommodation, and routes. In general users merely provide vague preferences about the scenic spots, hotels, and type of transportation, and thus the travel agent has to arrange feasible schedules and tourist routes with these constraints and preferences. However customer satisfaction often ascribes to the smooth execution of the tourist route and high cost-effectiveness. So, in this thesis, we focus on the route planning part of a tourist plan. But in contrast to central travel agent planning based on collected timetables, multiple transportation agents are introduced to offer dynamic latest timetables and flexible ticket fares. A transportation agent adjusts vehicle, flight, or ship dispatch based on customer demands and offers appealing fares based on profits. To find out the best route fitting user preferences, a heuristic most suitable path finding algorithm is used to generate user preferred routes with variations as candidate routes, and the transportation agents within the candidate routes compete in a user-preference based auction to decide the best route. In this thesis, we show how multiple transportation agents could negotiate and form coalition with each other to form competitive and appealing tourist routes. Besides higher customer satisfaction, the integration of the profit, privacy, and business strategies of transportation agents makes the multi-agent environment more complete and believable.