CNRS Laboratory for the Analysis and Modeling of Decision Systems

Paris, France

CNRS Laboratory for the Analysis and Modeling of Decision Systems

Paris, France
SEARCH FILTERS
Time filter
Source Type

Bonnet E.,Hungarian Academy of Sciences | Paschos V.Th.,CNRS Laboratory for the Analysis and Modeling of Decision Systems
RAIRO - Operations Research | Year: 2017

We discuss approximability in FPT-time for the class of subset optimization graph problems where a feasible solution S is a subset of the vertex set of the input graph. This class encompasses many well-known problems, such as min dominating set, min vertex cover, max independent set, min feedback vertex set. We study approximability of such problems with respect to the dual parameter n - k where n is size of the vertex set and k the standard parameter. We show that under such parameterization, many of these problems, while W[·]-hard, admit parameterized approximation schemata. © EDP Sciences, ROADEF, SMAI 2017.


Cazenave T.,CNRS Laboratory for the Analysis and Modeling of Decision Systems
Communications in Computer and Information Science | Year: 2017

Monte Carlo Tree Search (MCTS) is a general search algo-rithm that has improved the state of the art for multiple games and optimization problems. Nested Rollout Policy Adaptation (NRPA) is an MCTS variant that has found record-breaking solutions for puzzles and optimization problems. It learns a playout policy online that dynami-cally adapts the playouts to the problem at hand. We propose to enhance NRPA using more selectivity in the playouts. The idea is applied to three different problems: Bus regulation, SameGame and Weak Schur numbers. We improve on standard NRPA for all three problems. © Springer International Publishing AG 2017.


Bouyssou D.,CNRS Laboratory for the Analysis and Modeling of Decision Systems | Marchant T.,Ghent University
Journal of Informetrics | Year: 2014

This paper analyzes several well-known bibliometric indices using an axiomatic approach. We concentrate on indices aiming at capturing the global impact of a scientific output and do not investigate indices aiming at capturing an average impact. Hence, the indices that we study are designed to evaluate authors or groups of authors but not journals. The bibliometric indices that are studied include classic ones such as the number of highly cited papers as well as more recent ones such as the h-index and the g-index. We give conditions that characterize these indices, up to the multiplication by a positive constant. We also study the bibliometric rankings that are induced by these indices. Hence, we provide a general framework for the comparison of bibliometric rankings and indices. © 2014 Elsevier Ltd.


Aloulou M.A.,CNRS Laboratory for the Analysis and Modeling of Decision Systems | Dolgui A.,CNRS Laboratory of Informatics, Modeling and Optimization of Systems | Kovalyov M.Y.,National Academy of Sciences of Belarus
International Journal of Production Research | Year: 2014

Non-deterministic lot-sizing models are considered which serve for an explicit determination of lot sizes in an uncertain environment. Taxonomy components for such models are suggested and a bibliography structured according to these components is presented. The taxonomy components are numeric characteristics of a lot-sizing problem, names of uncertain parameters and names of approaches to model the uncertainty. The bibliography covers more than 300 publications since the year 2000. © 2013 Taylor & Francis.


Cornaz D.,CNRS Laboratory for the Analysis and Modeling of Decision Systems | Galand L.,CNRS Laboratory for the Analysis and Modeling of Decision Systems | Spanjaard O.,CNRS Laboratory for Informatics
IJCAI International Joint Conference on Artificial Intelligence | Year: 2013

This paper is devoted to complexity results regarding specific measures of proximity to single-peakedness and single-crossingness, called "single-peaked width" [Cornaz et al., 2012] and "single-crossing width". Thanks to the use of the PQ-tree data structure [Booth and Lueker, 1976], we show that both problems are polynomial time solvable in the general case (while it was only known for single-peaked width and in the case of narcissistic preferences). Furthermore, we establish one of the first results (to our knowledge) concerning the effect of nearly single-peaked electorates on the complexity of an NP-hard voting system, namely we show the fixed-parameter tractability of Kemeny elections with respect to the parameters "single-peaked width" and "single-crossing width".


Cazenave T.,CNRS Laboratory for the Analysis and Modeling of Decision Systems
Theoretical Computer Science | Year: 2016

Monte Carlo Tree Search (MCTS) is the state of the art algorithm for General Game Playing (GGP). We propose to learn a playout policy online so as to improve MCTS for GGP. We also propose to learn a policy not only using the moves but also according to the features of the moves. We test the resulting algorithms named Playout Policy Adaptation (PPA) and Playout Policy Adaptation with move Features (PPAF) on Atarigo, Breakthrough, Misere Breakthrough, Domineering, Misere Domineering, Knightthrough, Misere Knightthrough and Nogo. The experiments compare PPA and PPAF to Upper Confidence for Trees (UCT) and to the closely related Move-Average Sampling Technique (MAST) algorithm. © 2016.


Mayag B.,CNRS Laboratory for the Analysis and Modeling of Decision Systems
Communications in Computer and Information Science | Year: 2016

In a context of Multiple Criteria Decision Aid, we present a decision model explaining some French hospitals rankings in weight loss surgery. To take into account interactions between medical indicators, we elaborated a model based on the 2-additive Choquet integral. The reference subset, defined during the elicitation process of this model, is composed by some specific alternatives called binary alternatives. To validate our approach, we showed that the proposed 2-additive Choquet integral model is able to approximate the hospitals ranking, in weight loss surgery, published by the French magazine “Le Point” in August 2013. © Springer International Publishing Switzerland 2016.


Gourves L.,CNRS Laboratory for the Analysis and Modeling of Decision Systems
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2015

This paper deals with states that are immune to group deviations. Group deviations help the players of a strategic game to escape from undesirable states but they compromise the stability of a system. We propose and analyse a solution concept, called profitable deviation strong equilibrium, which is between two well-known equilibria: the strong equilibrium and the super strong equilibrium. The former precludes joint deviations by groups of players who all benefit. The latter is more demanding in the sense that at least one member of a deviating coalition must be better off while the other members cannot be worst off. We study the existence, computation and convergence to a profitable deviation strong equilibrium in three important games in algorithmic game theory: job scheduling, max cut and singleton congestion game © Springer International Publishing Switzerland 2015.


Gourves L.,CNRS Laboratory for the Analysis and Modeling of Decision Systems | Monnot J.,CNRS Laboratory for the Analysis and Modeling of Decision Systems | Tlilane L.,CNRS Laboratory for the Analysis and Modeling of Decision Systems
IJCAI International Joint Conference on Artificial Intelligence | Year: 2013

We consider the problem of equitably allocating a set of indivisible goods to n agents so as to maximize the utility of the least happy agent. [Demko and Hill, 1988] showed the existence of an allocation where every agent values his share at least Vn(α), which is a family of nonincreasing functions in a parameter α, defined as the maximum value assigned by an agent to a single good. A deterministic algorithm returning such an allocation in polynomial time was proposed [Markakis and Psomas, 2011]. Interestingly, Vn(α) is tight for some values of α, i.e. it is the best lower bound on the valuation of the least happy agent. However, it is not true for all values of α. We propose a family of functions Wn such that Wn(x) ≥ Vn(x) for all x, and Wn(x) > Vn(x) for values of x where Vn(x) is not tight. The new functions Wn apply on a problem which generalizes the allocation of indivisible goods. It is to find a solution (base) in a matroid which is common to n agents. Our results are constructive, they are achieved by analyzing an extension of the algorithm of Markakis and Psomas.


Paschos V.T.,CNRS Laboratory for the Analysis and Modeling of Decision Systems
4OR | Year: 2015

We outline a relatively new research agenda aiming at building a new approximation paradigm by matching two distinct domains, the polynomial approximation and the exact solution ofNP-hard problems by algorithms with guaranteed and non-trivial upper complexity bounds. We show how one can design approximation algorithms achieving ratios that are “forbidden” in polynomial time (unless a very unlikely complexity conjecture is confirmed) with worst-case complexity much lower than that of an exact computation. © 2015, Springer-Verlag Berlin Heidelberg.

Loading CNRS Laboratory for the Analysis and Modeling of Decision Systems collaborators
Loading CNRS Laboratory for the Analysis and Modeling of Decision Systems collaborators