Ph.D.: Swarm Intelligence model based strategy for multiple-fault analysis, diagnosis and repair in telecommunication wireline networks (ref.0005897)

With 231 million customers on five continents, Orange Group is one of the world's biggest telecommunications operators. Research and innovation in Orange Group is held by the worldwide network of Orange Labs and Techno centres. 5 000 people (researchers, engineers, designers and developers) are spread across 4 continents. “Orange Lab Networks” is the division in charge of innovation in the field of network infrastructure. The team offering you a Ph.D. is comprised of experts in the fields of copper and fibre network operation and maintenance. They are in charge of continually upgrading the capacity for France Telecom to analyze, diagnose and repair its wire line telecommunication networks for the mass market.
The proposed Ph.D. subject follows the strategic guidelines of the company. This strategic plan, named “Conquest 2015”, particularly targets our networks and customers. Indeed, the aim is to make Orange number 1 in terms of quality of services and customer experience. The explicit goals of this thesis are to improve the quality of services, to have more efficient networks and to improve the experience of customers facing a telecommunication services blackout or reduction.

your role
Our issues are telecommunication wireline networks and fault management. They comprise detection, isolation, recovery of faults and diagnosis reliability. The first initial off site diagnosis is crucial for:
• the relevance of the following actions be taken to recover from the failure (call out or not)
• the speed of repair
• the cost of the entire aftersales process
• customer satisfaction.
Our actual automatic diagnostics tools are based on Expert systems. However, with a large number of configurations, this approach turns out to be burdensome for designers and developers who maintain these tools. At the same time, the probability of occurrence of faults in large communication networks grows as they become widespread, complex and heterogeneous.
Researchers discuss the best way to diagnose Multiple-Faults problems: Network graph model, Swarm Intelligence. Swarm Intelligence consists of the use of multi-agent systems, inspired by the collective behaviour of social insects, such as ants or bees. Even though single agents have simple actions, a cooperative strategy gives rise to an overall capability of solving complex tasks – much like wasps building nests. We briefly guide the readers to the following papers:
• “Ant-search strategy based on likelihood trail intensity modification for multiple-fault diagnosis in sensor networks”, Pasquale Arpaia, Université de Sannio (Italie), janvier 2013
• “A unified theory of fault diagnosis and distributed fault management in communication networks”, Guy Berthet, Ecole Polytechnique de Lausanne, 1996
The main goal of our research is to use the available data store in a network graph and Swarm Intelligence based diagnosis tree.
From a concrete point of view, the work proposed by the team to the student is as following:
• code Swarm intelligence strategy and Network graph model in a new diagnosis tree (MySQL, Java, Spring)
• measure the pros and cons of that diagnosis tree compared to the current one - the expert method based one.
• propose evolution of Orange diagnosis strategies in the following areas :
o network deployment rules
o data store availability
o local treatment of data

your profile
You have achieved an Engineer diploma or Master of Science in computer sciences. You have competencies in algorithm, coding and applied math. Here you will also need to mobilizeknowledge in information theory and telecommunication networks.
You are autonomous and curious. You are keen to advance your Ph.D. in the international environment - to prove your ability to manage problems. The hosting team is very open to a native English speaker. French is not a fundamental necessity in order to work with your colleagues.

the main benefit of this offer
With this doctoral job, you will greatly enhance your many opportunities for further post-doctoral career development. You will be immersed in a multidisciplinary team, at the frontier of telecommunication and application development and coding.
You will be in contact with the current telecom industry standards, have the chance to deal with extremely sensitive data about telecommunication and after sales process reliabilities.
You will integrate with an attentive team that will pay real attention to you and your intellectual ability.
You will have access to international academic papers online, and to relevant computer and software applications.