Summary:
Automated Planning is the area of Artificial Intelligence that deals with
choosing the
required sequence of actions of an agent or actor to achieve a desired state.
Specifically, in the relative bibliography, there are many tools that deal with
planning,
mostly in cases of preference-based planning or in cases of multi-agent
planning.
However, there seems to be a lack of research in the area combining the above,
that is,
preference-based planning for many actors. This area presents particular
interest,
because such applications are directly applicable to everyday situations, such
as
collective decision making about an action plan. The contribution of the
current thesis is
a way of formalizing such problems, so as to allow their solution using
existing tools,
creating plans in a centralized way, taking into account the preferences of
each actor
participating in the plan. As a proof of principle this approach was applied
successfully
in different instances of an example of decision making: specifically the
process to be
taken by a committee when all members are called to comment officially on a
letter
given by the chairman of the committee.
Keywords:
automated planning, multi-agent planning, preferences, goals, constraint programming
