Srdjan Vesic (CRIL, CNRS - Université d’Artois)
We study semantics that evaluate arguments in argumentation graphs, where each argument has a basic strength, and may be attacked by other arguments. We start by defining a set of principles, each of which is a property that a semantics could satisfy. We provide a formal analysis and comparison of existing semantics. Finally, we define three novel semantics that satisfy more principles than existing ones.
Anastasia Paparrizou (CRIL, CNRS - Université d’Artois)
It is well-known that the simultaneous satisfaction of a set of constraints is intractable in general and problems can become very difficult to solve as their size increases. Constraint Programming has developed various techniques to tackle this inherent problem. In this talk, I will speak about one of the most important such techniques which is the enforcement of a local consistency. I will mainly focus on less “traditional” local consistencies and recent promising results (presented in IJCAI’17).
Hubie Chen (Birkbeck, University of London)
This will be a swashbuckling adventure through the jungles of complexity. Animals such as SAT, CSP, and their quantified variants will be encountered and, sometimes, tamed. Seatbelts are advised; this full-throttle journey may veer into proof complexity territory before its finale.
Séminaire Solving Multiobjective Discrete Optimization Problems with Propositional Minimal Model Generation (CP 2017)
Takehide Soh (Kobe University, Japan)
This talk about SAT encoding consists of two parts.
The first part briefly explains recent our work around SAT
encoding—a SAT-based CP system “scarab” and a hybrid encoding
integrating the order and log encodings.
In the second part, as an application of the order encoding, we detail
a propositional logic based approach to solve MultiObjective Discrete
Optimization Problems (MODOPs).
In this approach, there exists a one-to-one correspondence between
a Pareto front point of MODOP and a P-minimal model of
the CNF formula obtained from MODOP.
This correspondence is achieved by adopting the order encoding
as encoding for multiobjective functions.
Finding the Pareto front is done by enumerating all P-minimal models.
The beauty of the approach is that each Pareto front point is
blocked by a single clause that contains at most one literal
for each objective function.
We evaluate the effectiveness of this approach by empirically
contrasting it to a state-of-the-art MODOP solving technique.