Information Technology Reference
In-Depth Information
- in terms of journeys (preferred means of transport, cost);
- in terms of their daily schedules (in relation to the preferences of students and
depending on the transport available).
The timetable of students therefore becomes “dynamic”. The functionalities
associated with the system can thus:
- provide transport information that is relevant to the user;
- plan itineraries;
- manage a personal schedule;
- spread information by ensuring the usability of the human-machine interface;
- manage the presentation of information using different aids.
Its originality lies in offering transport solutions corresponding to the planned
events. Indeed, the addition of an event linked to a journey in the schedule sets off
the calculation of a corresponding itinerary. Moreover, this itinerary is chosen
according to the profile of the user registered in the system. This project is based on
the field of multi-agent systems [MAN 02] and research on personalization and
learning [ANL 06], [PET 03]. Technical tests have been carried out in order to
detect possible defects, introduced in the methodological phase of definition and
implementation of the system. The proposed schedule functions by using Webpages
as well as a personal digital assistant (PDA).
The laboratory prototype was tested and evaluated by around 20 students with
criteria relating to utility and usability, all the while putting emphasis on
personalization [SOU 08a], [SOU 08b]. Let us recall here that usability mainly
accounts for the quality of human-machine interaction in terms of how easy it is to
learn and use. Utility determines whether the interface meets the needs of the user
[NIE 93]. In terms of results on utility, 55% of users were very satisfied, 40% were
satisfied and 5% were not very satisfied. With regards to usability of the system, the
rate of user satisfaction was 41% and 59% were very satisfied. These good results
can be explained by the fact that the service is quite simple and was fairly quick and
easy to learn. The tests show that the planned personalization functionalities have
indeed been carried out, but they also highlight which areas need to be improved.
These in particular have to do with the performance of the developed prototype,
which demonstrate a need to optimize the research and filter algorithms before
being applied on a more large-scale use.
