Séminaires

Séminaires

A venir

  • Date 2016-08-29 à 10:00
  • Orateur : Ming Liu
  • Lieu : IBGBI, 23 Boulevard de France, 91000 Evry
  • Salle : Salle 334
  • Title : Bi-objective Optimization of a Single Machine Batch Scheduling Problem with Energy Cost Consideration
  • Résumé : With the increasing energy price, the rapid growth of electricity demand and severe challenges for sustainable development, energy-efficient scheduling is becoming more and more important for power-intensive manufacturing industry, especially for batch production companies. This work investigates a bi-objective single machine batch scheduling problem with non-identical job sizes, the time-of-use (TOU) electricity prices, and different energy consumption rates of the machine. The first objective is to minimize the makespan and the second is to minimize the total energy costs, by considering both the machine utilization and the economic cost. The problem is formulated as an integer programming model. Exact and heuristic methods are developed to obtain a set of nondominated solutions. Computational experiments on randomly generated instances show the effectiveness of the methods. A study case of a real-world glass manufacturing company is also conducted.

Short-bio of Dr. Liu

Dr. Ming Liu is an associated professor at Tongji University, China, with over 10 years of experience in Scheduling and Logistics. He received his PHD in Laboratorie Genie Industriale, Ecole Centrale Paris, (2009 France), and another PhD in Management Science and Engineering, Xi’an Jiaotong University (2010 China). Dr. Liu’s research focuses on the applications of Operations Research in container logistics at seaports, including: quay crane scheduling, berth allocation, yard storage allocation, Inland container transportation, etc. Dr. Liu’s area of expertise includes but not limited in mathematical programming, algorithms design, numerical analysis. He has published more than 40 journal papers, including <Transportation Science>, <Transportation Research Part B>, <Transportation Research Part E>, <European Journal of Operational Research>, <International Journal of Production Economics>, <International Journal of Production Research>, etc.

Passé

Optimal algorithm for quay crane double cycling problem
  • Date 2016-07-12 à 14:30
  • Orateur : Ming Liu
  • Lieu : IBGBI, 23 Boulevard de France, 91000 Evry
  • Salle : Salle 334
  • Résumé : Maritime transport is the backbone of international trade and a key driver of globalization. Around 80 percent of global trade by volume and over 70 per cent by value is transported by sea and handled by ports worldwide.

Ocean container transport now is a critical element of any supply chain. In particular, due to increased port congestion, heightened pressure on energy usage, and carbon emission, the immediate effect of container transportation on supply chain has recently attracted much attention. For container terminals, as quay cranes are the most expensive equipment, their efficiency impacts the throughput of seaports. In my talks, I will present some operational research methods to address key issues in quay crane scheduling problems.

Models and Logics for Strategic Reasoning with incomplete Information
  • Date 2014-01-16 à 14:00
  • Lieu : IBGBI, 23 Boulevard de France, 91000 Evry
  • Salle : Petit amphi
  • Résumé : Over the past decade Logic has become an increasingly popular and successful framework for modeling and analyzing strategic reasoning in games and multi-agent systems. Various formal logical systems have been proposed and studied for that purpose. Besides the purely technical and intrinsically logical problems that have arisen in these studies, a multitude of new conceptual questions related to the semantics of these logics have emerged. These questions refer to the fundamental notions of strategies and strategic abilities of agents/players and coalitions of agents to achieve objectives, particularly in the context of incomplete information. In this talk I will first briefly introduce (if necessary) multi-agent transition systems, aka concurrent game models (CGM), and the probably most popular modal logic for strategic reasoning in multi-agent systems ATL*, where one can formally express statements about the strategic ability of an agent or a coalition of agents, to achieve a goal, such as: “The agent (or, coalition of agents) A has a strategy such that, if A follows that strategy, then the goal G will be achieved, no matter what the other agents do”. Then I will focus on concurrent game models with (static) incomplete information and will give semantics of ATL* and its epistemic extensions in them, allowing strategic reasoning under incomplete information. Lastly, I will discuss some related open problems and some ideas for their solutions.
Risk Bounds and Learning Algorithms for the Regression Approach to Structured Output Prediction
  • Date 2014-01-23 à 14:00
  • Lieu : IBGBI, 23 Boulevard de France, 91000 Evry
  • Salle : Petit amphi
  • Résumé : In contrast with classification and regression where the learner tries to predict a scalar output, structured output prediction attempts at predicting outputs which can be very complex such as sequences, parse trees, and graphs. However, most learning algorithms, such as structured SVM and Max Margin Markov Networks require a prohibitive training time due to the fact that they have to solve a (often NP-hard) pre-image problem for each training example and for each update of the predictor produced by the learner. Here, we provide some conditions under which, a vector-valued regression approach, which avoids the pre-image problem during learning, is justified and has rigorous guarantees. More precisely, we show that the quadratic regression loss is a convex surrogate of the structured prediction loss when the output kernel satisfies some condition with respect to the structured prediction loss. We provide two upper bounds of the prediction risk that depend on the empirical quadratic risk of the predictor. The minimizer of the first bound is the regularized least-square predictor proposed by Cortes Mohri and Weston (2007) while the minimizer of the second bound is a predictor that has never been proposed so far. Both predictors are compared on practical tasks.
Rule-based modeling of protein-protein interaction networks: deriving pathways from facts
  • Date 2014-02-20 à 14:00
  • Orateur : Jean Krivine (CNRS researcher, PPS laboratory, Université Paris Diderot)
  • Lieu : IBGBI, 23 Boulevard de France, 91000 Evry
  • Salle : 334
  • Résumé : Signaling pathways of the cell are causal cascades of protein-protein interactions (that involve complex formation and post-transcriptional modifications) that are triggered in response to the reception of some (protein) signal at the surface of the cell. Such pathways are generally classified according to their phenotypic outcome, such as cell growth, death or motion. In this talk, which requires no deep knowledge about computer science nor biology, I will introduce Kappa, a rule-based formalism designed to model protein-protein interaction networks in a scalable fashion. Kappa is equipped with a stochastic simulator (KaSim) that may output time trajectories of particular protein complexes. We will see that by tracking the causal interplay between Kappa rule applications, it becomes possible to “explain” the appearance of particular observables (for instance a gene activation) during a simulation. We will argue that such “explanations” are good candidates for being “formal signaling pathways”. It is noteworthy that standard causal semantics that are considered in computer science fail to give satisfactory “formal pathways”. As a consequence the “formal pathways” that I will introduce in this talk require a notion of “non local causality”. This new concept, albeit non standard in computer science, seems to match the intuitive notion of causality in signaling pathways. We will conclude about the general interest of “non local causality” for the computer science community.
Strategic Reasoning with Epistemic Goals
  • Date 2014-06-26 à 15:00
  • Orateur : Tiago de Lima (PhD, CRIL, Université d’Artois and CNRS)
  • Lieu : IBGBI, 23 Boulevard de France, 91000 Evry
  • Salle : 334
  • Résumé : The objective of this talk is to show how logic can be used to model strategic reasoning. The board game Cluedo will be used as a running example to show how a formalism called dynamic epistemic logic can be used to model reasoning in scenarios where information is crucial for agents to attend their objectives. The talk is meant to be accessible to all computer scientists, thus, no previous knowledge about such formalism is required.
Simplicité en biologie: le cas du cycle cellulaire
  • Date 2015-10-14 à 14:30
  • Orateur : Damien Coudreuse (CR, CR1IGDR)
  • Lieu : IBGBI, 23 Boulevard de France, 91000 Evry
  • Salle : 334
  • Résumé : La prolifération des cellules eucaryotes repose sur un

mécanisme complexe et fortement conservé qui permet l’organisation d’une séquence précise d’événements interdépendants. Paradoxalement, le développement de nos connaissances des mécanismes moléculaires impliqués dans ce processus représente un obstacle à notre compréhension du circuit de base qui est nécessaire et suffisant pour assurer la division d’une cellule. Malgré l’émergence d’une vision de plus en plus intégrée du cycle cellulaire dans l’ensemble des mécanismes physiologiques de la cellule, nos travaux ont révélé la surprenante simplicité du noyau de contrôle de ce système. Ainsi, en utilisant des levures qui opèrent avec des circuits de régulation synthétiques, nous avons pu contourner la complexité inhérente aux réseaux de régulation endogènes pour déterminer les signaux essentiels qui assurent la prolifération cellulaire. Cette approche nous permet aujourd’hui de nous interroger sur la façon dont ce mécanisme, et plus généralement tout circuit de régulation simple, peut évoluer.

Last modified: 2016/08/25 09:44