Information Technology Reference
In-Depth Information
Figure 3.27 can provide us with an insight into this matter. It takes the elements
of a demonstrator developed in the context of the PREDIT AGENPERSO project
(human-machine interfaces based on PERSOnnal software AGENts of information
to the collective transport users) [PET 03b], bringing together LAMIH, INRETS
and Archimed. The personalized Web page, visible in this figure, is exclusively
meant for an identified user. The system knows via the intermediary of his schedule
that he must go to Bordeaux. The agents put themselves at the service of the user by
recuperating information that is relevant to him: the need to reserve the hotel before
3pm, collection of plans likely to be useful in the context of this trip, etc. The agents
also anticipate future trips. For example, several days later the user must go to
Marseille, but the intelligent agents detect that there is advanced notice of a strike
that could disrupt the trip; the user perhaps needs to consider a re-planning the
journey and is warned via the intermediary of the personalized system.
By generalization, other elements of personalized information can be expressed
with mobility information, via knowledge of the preferences of the user in terms of
leisure, for example.
Intelligent agents can indeed go in search of information likely to complement
the user's trip by informing themselves as to the possibilities of shows and events at
the destination place, as well as about the
best
restaurants (best in the sense of the
adaptation to the user's criteria), the most
interesting
museums (for example, “the
museum of pans, dishes and other kitchen utensils” if the user enjoys cooking), all
the while preparing the
best
way (according to the user's criteria) to access it.
Agents can obtain information about television programs if the user is too tired to
leave the hotel in the evening (the study taking into account the hundreds of
channels that might be available). Numerous other generalization ideas can of
course be envisaged, which opens up new research avenues.
3.7. Conclusion
This chapter has described a contribution to the personalization of ISs, in view
of improving the HMI and moving towards adaptive, intelligent HMIs, within a
huge international research movement. We have put forward a method called
PerMet (
PERsonalization
METhodology
) for the development of PISs. This method
enables both the implementation of a new PIS as well as the personalization of an
already existing IS. PerMet proposes an iterative and incremental development
model and allows the specific phases linked to the development of services and the
specific phases linked to personalization to be carried out in parallel. We have also
put forward PerSyst (PERsonalization SYSTem), which is a PS that supports the
PerMet method, consisting of agents at the service of users. PerMet and PerSyst
were validated in different applications based on real or simulated data for the
