Existing traveler information systems provide mass information to a large group of users who plan to travel in a similar way. Such information, although provided at a real time basis, may not satisfy the real need of personal travel as the information is provided without much consideration on individual's preferences. The provision of personalized travel information through the development of a three-tier client server based traveler information system on Internet (TISONI) is discussed in the paper. In this system, personalized travel advices are provided through a combinatorial modeling approach, which combines models describing travel behavior and shortest path algorithm. TISONI is developed to provide personalized and multimodal transit travel information to users. The three-tier structure adopted in TISONI facilitates a convenient and efficient way to update and transmit data between client and server. The core of TISONI is a path choice model, which was developed by calibrating users' behavior in mode and route choices. The personalized information covers two main aspects: (1) recognition of the users and their preferences and (2) personalized information provided to users by learning. Based on the architecture presented, real time data transmission and process, user interfaces and connectivity were introduced integrating with path choice models. Java servlets, JSP, as new technologies, run on a TISONI server, features high performance, database connectivity, stability, and security and perform most dynamic content operations. An implementation and evaluation of TISONI prototype based on western part of Singapore was discussed. Such implementation aims to test and demonstrate the capability of dynamic modification, path search and decision models, computation speed, and users interface. The three main criteria including timeliness, intelligence and efficiency were used to evaluate the good performance of this prototype model. It is concluded that the proposed architecture is able to provide a stable and efficient platform for provision of personalized travel information to transit users.
Samenvatting