Project Tasks
Task 1
Task Denomination
Tabling Common Platform
Starting Date
January 2008
Duration in Months
36
Expected Results
Currently, YapTab is already recognized as one of the most important
implementations of a tabling execution model. YapTab's design relies
on several important contributions. The goal of this task is to take
advantage of this potential and improve YapTab in order to produce a
powerful implementation that can be of wider interest to application
developers. The design and implementation of these features will be
based in several important limitations and practical deficiencies that
we have encountered when applying tabling to a variety of real-world
problems. Relevant contributions of this task include the proposals of
improving our tabling engine with support for negation; support for
storing tables externally using a relational database system; support
for alternative data structures and algorithms for representing and
managing the table space; support for alternative tabling mechanisms;
and better support for deterministic applications.
Task 2
Task Denomination
Tabling and Answer Set Programming
Starting Date
July 2008
Duration in Months
30
Expected Results
The main goal of this task is to study how tabling technology can be
used to improve the overall performance of ASP solvers. In particular,
an important result of this research should be the design and
implementation of an ASP system incorporating tabling technology as a
means to significantly reduce the search space for these kind of
problems. With such a system combining the power of tabling with that
of ASP, we expect to be able to improve the efficiency of particular
complex combinatorial search problems, such as Product Configuration,
Model Checking, Planning and Diagnosis problems.
Task 3
Task Denomination
Tabling and the Extended Andorra Model
Starting Date
July 2008
Duration in Months
30
Expected Results
The main goal of this task is to integrate tabling within the BEAM
system. Tabling is an elegant solution that fits naturally with the
EAM. First, tabling contributes to one of the main premises of the
EAM, reducing search space through the reuse of goals. Second, tabling
avoids looping thus guaranteeing termination for programs with finite
solutions. Third, both paradigms are less sensible to goal
ordering. At the end of this task, we expect to have a system that
fully supports tabled logic programs running within the Extended
Andorra Model environment. With such a system combining the power of
tabling with that of EAM, we expect to be able to improve the
declarativeness and efficiency of particular search and database
applications, like ILP and DDB applications.
