Menu:

logo UPMC
logo LIP6


Latest news:

September 30, 2016: More information on social events

July 7, 2016:Invited speakers announced

June 20, 2016: Program online

June 16, 2016:
Registration open

April 12, 2016:
Accepted Special Session on Intelligent Computing for Smart Cities

March 30, 2016:
Accepted Special Issue on New trends and Innovations in Intelligent Distributed Computing

March 2016:
New extended deadline:
April 4

December 2015:
The CFP is online.

October 2015:
The website is online.

Invited Speakers


Carlos Cotta Universidad de Málaga, Spain
Memetic Algorithms, Self-* Properties, and Ephemeral Environments

Abstract

Evolutionary and memetic algorithms are very well suited to parallel environments thanks to their flexibility and decentralized nature. However, their deployment on emergent computational environments of a dynamic and unsteady nature -such as P2P or volunteer-computing networks- is not exempt of challenges. Although these techniques are inherently resilient to some extent, they need being endowed with self-control abilities in order to react and readjust their behavior in response to the fluctuations of the environment. Such abilities fall under the umbrella term of self-* properties. This talk provides an overview of recent research in this direction, outlining different mechanisms to boost the resilience and performance of memetic algorithms in unstable environments.


About Carlos Cotta

Carlos Cotta obtained his MSc and PhD in Computer Science from the University of Málaga (UMA), Spain in 1994 and 1998 respectively. He holds a tenured Professorship in the Department of Lenguajes y Ciencias de la Computación of this University since 2001. His main research areas involve metaheuristic optimization -in particular hybrid and memetic approaches- with a focus on both algorithmic and applied aspects (particularly combinatorial optimization) as well as complex systems. He has published +200 articles on these topics, and is involved in the programme committees and editorial boards of major conferences and journals in the area.



Serge Haddad ENS Cachan, France
Fault diagnosis in probabilistic models

Abstract

In this talk, I will consider partially observable probabilistic systems prone to faults and study the problem of (1) the specification of a diagnoser, (2) the existence of a diagnoser (called diagnosability), and (3) the synthesis of such diagnosers. There are two requirements about diagnoser, soundeness and reactivity, that can be expressed in several ways and we will establish an almost complete picture of the relations between the different specifications. Then I will adopt an algorithmic point of view and discuss the complexity of the diagnosability problem (and related notions) in finite state systems and the optimal size of diagnosers. At the end of the talk, I will briefly give an overview of diagnosability of infinite-state systems and active diagnosis where the diagnoser can in addition control the system.


About Serge Haddad

Serge Haddad is a former student of ENS Cachan in mathematics. After his PhD in computer science, he got in 1988 an assistant professor position at University Pierre et Marie Curie. Then he became full professor of University Paris-Dauphine in 1993 and moved to ENS Cachan in 2008. His main theoretical interests concern the specification and analysis of parallel and distributed systems featuring quantitative aspects like time and/or probability. He has also applied the techniques he developped to multi-agent systems, web services and more recently to diagnosis. He has numerous national collaborations and his main international collaborations are with Universities of Torino and Zaragossa. He received five best paper awards from international conferences between 2005 and 2014.