Time filter

Source Type

Toulouse, France

Liberti L.,Ecole Polytechnique - Palaiseau | Cafieri S.,ENAC | Savourey D.,Ecole Polytechnique - Palaiseau
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

Most optimization software performs numerical computation, in the sense that the main interest is to find numerical values to assign to the decision variables, e.g. a solution to an optimization problem. In mathematical programming, however, a considerable amount of symbolic transformation is essential to solving difficult optimization problems, e.g. relaxation or decomposition techniques. This step is usually carried out by hand, involves human ingenuity, and often constitutes the "theoretical contribution" of some research papers. We describe a Reformulation- Optimization Software Engine (ROSE) for performing (automatic) symbolic computation on mathematical programming formulations. © 2010 Springer-Verlag. Source

Marzuoli A.,Georgia Institute of Technology | Hurter C.,ENAC | Feron E.,Georgia Institute of Technology
Proceedings - 2012 Conference on Intelligent Data Understanding, CIDU 2012

With the predicted growth of air traffic, traffic flow managers need new tools to access information to support their decision making processes. Recent progress with information visualization tools enables users to explore large data sets and extract decisive knowledge. Their advantages for air traffic applications are presented in this paper. They can provide high level information to aggregate trajectories. With constant feedback due to human perception, a flow model of the airspace, reflecting its intrinsic structure, is elaborated and can be used for further research. © 2012 IEEE. Source

Cafieri S.,ENAC | Cafieri S.,Toulouse 1 University Capitole | Costa A.,Singapore University of Technology and Design | Hansen P.,HEC Montreal
Journal of Complex Networks

Finding communities in complex networks is a topic of much current research and has applications in many domains. On the one hand, criteria for doing so have been proposed, the most studied of which is modularity. On the other hand, properties to be satisfied by each community have been suggested. It has recently been observed that one of the best known such properties, i.e. the weak condition, proposed by Radicchi et al. (2004, Proc. Natl. Acad. Sci. USA, 101, 2658) was not satisfied by one or more communities in a partition which maximizes (approximately) some of the best known criteria. It was therefore proposed byWang et al. (2009, Lect. Notes in Oper. Res., 11, 142) to merge both approaches by maximizing a criterion subject to the weak condition. We consider five community-defining conditions, which we call cohesion conditions (strong, semi-strong, almost-strong, weak and extra-weak conditions). We add cohesion conditions, one at a time, as constraints to a modularity maximization problem, thus obtaining new mathematical optimization models, which we solve exactly. We study the impact of cohesion conditions on modularity maximization. Strong, semi-strong and almost-strong cohesion conditions appear to be generally too restrictive and the extra-weak condition too lax. The weak condition is verified by some but not all modularity maximizing partitions of the considered real-world networks. Imposition of this condition on those partitions for which some communities do not verify it reduces modularity moderately but sometimes changes the optimal number of communities and their composition.We also show, on a known example, that the strong, semi-strong and almost-strong conditions allow us to overcome the resolution limit of modularity. The behaviour of modularity maximization subject to cohesion constraints appears to be coherent with the detectability of the modular structure of the considered networks. © The authors 2014. Source

Crawled News Article
Site: http://www.materialstoday.com/news/

Spanish plastics technology center AIMPLAS reports that in 2015 it took part in 123 R&D projects in collaboration with 227 companies, generating a return of €12.6 million for the companies involved. These projects involve sectors such as packaging, automotive and transport, construction, environment and recycling, electrical-electronic and aeronautics. A total of 120 professionals work in AIMPLAS in areas related to R&D and also provide other services, such as technical assessments, analysis and testing and training to other plastics companies. The organization’s analysis and testing laboratory has the highest number of tests on plastics accredited by ENAC (the Spanish national accreditation body). During the last financial year, more than 6,500 technology services were developed and 158 training activities were organised, where 2 518 attendees from 889 companies received 3,963 training hours, AIMPLAS says. Its center has currently more than 2 000m2 of facilities and, in 2015, a new filaments extrusion line has been incorporated to this plant’s equipment to print parts by means of 3D printing. This story uses material from AIMPLAS, with editorial changes made by Materials Today. The views expressed in this article do not necessarily represent those of Elsevier.

Crawled News Article
Site: http://phys.org/biology-news/

So far, the genetic methods for establishing proper DNA identification used to take several days to produce some conclusive results. This innovative methodology is of great interest for the canning industry and is a tool that can help to authenticate what it says on the label. Changes in production processes have led to a revolution in the canning industry, since canned products are produced in many cases using imported frozen tuna fillets. These skinned fillets offer tremendous advantages in terms of productivity and yield of the processes, yet on occasions, in view of the difficulty in distinguishing between species visually, errors may occur in the labelling of canned products. The DNA system developed by the Molecular Biology Laboratory at AZTI-Tecnalia for detecting various species of canned and processed tuna, has received accreditation from the National Accreditation Body (ENAC). This methodology not only offers speed in identifying species but also reliable, conclusive results. So this new method denotes a great opportunity to guarantee that products are properly labelled and to ensure quality for consumers. Authentication methodologies for canned products are generally based on DNA fragment detection in any kind of processed fish sample, including canned fish. These methods, known as genetic methods, are always very reliable but the downside is that it takes several days to obtain a conclusive result. However, AZTI-Tecnalia's innovative system based on fluorescent probe detection enables an accurate result within 24 hours. The Molecular Biology Laboratory at AZTI-Tecnalia continues developing new methods to authenticate several marine species in order to provide effective tools for the fishing and canning industry to ensure their traceability systems. DNA technology enables specialists at the R&D centre to genetically identify Bay of Biscay anchovy and albacore tuna, among other fish. They also provide systems for authentication of Arabica coffee, cheeses under protected designation of origin, juices, and blends of meat, among other foodstuffs.

Discover hidden collaborations