TOP  >  Thesis/Dissertation  >  Making Route Schedule for Enhancing Tourists Satisfaction Based on Real-Time Information in Sightseeing Area

Making Route Schedule for Enhancing Tourists Satisfaction Based on Real-Time Information in Sightseeing Area

In this paper, we intend to make route schedules for tourists to visit as many spots which they wish to visit as possible. This technology can be applied to develop a system that recommends an appropriate schedule for each tourist. Using the schedule recommendation system, tourists can go around many spots efficiently and have a good time in a sightseeing area.

In order to make appropriate schedules for tourists, we need to estimate congestion of roads and spots in the future. For that purpose, we developed a simulator which places roads and spots and simulates the tourists' transition between spots. We move the tourist agents in the simulator and observe the congestion of roads and spots in the future.

This simulation needs the transition trend of tourists, but it is impossible to get the correct transition trend of tourists in advance. Suppose that we use a schedule recommendation system, we can get the transition trend of system user tourists who use this recommendation system and follow the recommended schedule, but cannot get the transition trend of general tourists who don't use this recommendation system. In a previous research, all tourists are supposed to use this recommendation system and %simulate congestion of routes and spots, the simulator knows the transition trend of all tourists. However, it is not a realistic assumption and the simulation results may be different from congestion in a real sightseeing area. As a result, we cannot make appropriate schedules for tourists in a real situation.%the results of the simulation is dubious.

In this paper, we estimate the transition trend of general tourists in simulation based on real-time information and update schedules so that we can cope with this problem. Real-time information that we use is the list of spots that each tourist has visited before that moment. We rerun the simulation reflecting the real-time information at that moment %to real sightseeing area and make appropriate schedules for tourists.

In the proposed method, we implement the transition trend of general tourists as the congestion trend in each spot. The reliable estimation results need to be got with a small amount of data so that simulation reflects the influence of general tourists based on real-time information. If we calculate the transition trend of general tourists directly with a small amount of data, we cannot get the reliable estimation results because the number of transitions(parameters) is the number of combinations between the spots and it is very large. In our method, we calculate the congestion trend in each spot based on real-time information. In this case, the number of parameters is the number of the spots and it is much smaller than the number of transitions. As the result, we can estimate them reliably with a small amount of data.

Simulation reflects the influence of general tourists estimated by real-time information so that we predict the future in the sightseeing area and update schedules. We reduce the capacity of each spot in simulation based on estimated results of the congestion trend, so simulation reflects the influence of general tourists. By using simulation results, our method updates the schedule of each system user tourist by reducing the number of visiting spots and changing visiting order of the spots.

To examine the effectiveness of the proposed method, we conducted simulation experiments, in which we evaluated the satisfaction of system user tourists by our method. In the first experiment, we compared our method with a previous method when the transition trend of tourists given in advance is incorrect. This experiment confirmed that our method can improve satisfaction of system user tourists by updating schedules based on real-time information. In the second experiment, we compared a method calculating the congestion trend in each spot with a method calculating the transition trend of general tourists directly. This experiment showed that our method can get reliable estimation results with a small amount of data. The last experiment investigated the validity of our method, in the cases which we modeled the transition strategy of general tourists in various way in evaluation simulation. These results show that our method can effectively reflect the real-time information and can update schedules for system user tourists adequately.