Integer programming based crowd behavior analysis for world expo

With the large-scale activities increasing gradually, crowd behavior analysis becomes more and more popular and important. Trajectory identification and estimation are key techniques in crowd behavior analysis. This paper develops an integer programming based trajectory identification algorithm. Shanghai World Expo 2010 is a typical example of large-scale activities and the proposed algorithm is implemented on World Expo data. The experimental results show that the tourists´ visiting patterns are different in different days due to various attractive strategies of the Expo site; and the tourists´ visiting patterns of different time periods in the same day are different as well.