Sami Erol Gelenbe
Education: PhD. of Computer sciences
University: Imperial College London
Information and Communications Technology (ICT) was Awarded to Prof. Sami Erol Gelenbe for his pioneering research on “Modeling and Performance Evaluation of Computer Systems.”
He invented the random neural network and the eponymous G-networks or Gelenbe Networks as a model for queueing systems. The work shows that one can evaluate the performance of ICT systems that incorporate important control functions, such as removing overload. Now these methods are widely accepted standard approaches in the telecommunications and ICT industry.
Sami Erol Gelenbe is a Turkish computer scientist, electronic engineer and applied mathematician who is a member of Science Academy of Turkey (Bilim Academy), and holds the “Dennis Gabor” professorship in Electrical and Electronic Engineering at Imperial College, London.
After a childhood spent in Istanbul and Alexandria (Egypt), he graduated from Ankara College in 1962 and the Middle East Technical University in 1966, winning the K.K. Clarke Research Award for work on "partial flux switching magnetic memory systems". He continued at Polytechnic University (now part of New York University), earning a master's degree and a PhD thesis on "Stochastic automata with structural restrictions”.
At the INRIA institute and the University of Paris-Orsay, Erol created the team that developed the “Queueing Network Analysis Package”. He invented Diffusion Approximations for computer performance, derived transmission schemes to optimize the throughput of random access communications that are the basis of the well-known MAC protocols, and established schemes for maximizing the reliability of data bases.
Then, working at the University of Paris-Descartes and at Duke University, he developed new product form queueing networks with negative customers and triggers known as G-networks or Gelenbe Networks. The work shows that one can mathematically evaluate the performance of ICT systems that incorporate important control functions, such as removing overload and moving work dynamically among servers. He also introduced a new spiked stochastic neural network model known as the Random Neural Network, developed its mathematical solution and learning algorithms, and applied it to both engineering and biological problems.
Today, we view systems such as Skype as a common means of human communication over the packet based Internet. In 1978, before packet networks were available, motivated by the need to improve voice telephony, Erol and a colleague from the company LMT made a major invention (patent filed in 1979 and awarded in 1982) regarding Voice Transmission over a packet network router, whose detailed architecture was described in a second patent filed by Erol and two colleagues on behalf of Thomson CSF (now Thales) in 1979 (also awarded in 1982).
Dr. Gelenbe has founded the ISCIS (International Symposium on Computer and Information Sciences) series of conferences that since 1986 are held annually in Turkey, the USA and Europe to bring together Turkish computer scientists with their international counterparts. He currently works on the interaction between Energy Consumption and Quality of Service in ICT, and on improving the security of Networks and the Cloud. He has graduated over 75 PhD students, and mentored numerous post-doctoral fellows.
The heart of ICT system performance methodology is based on queueing networks and their computationally efficient “product form” solution. The single major breakthrough in this area after the 1990’s is Erol’s new result, now known as G-Networks or Gelenbe Networks, show that one can evaluate the performance of ICT systems that incorporate important control functions.
In queueing theory, a discipline within the mathematical theory of probability, a G-network (Generalized queueing network or Gelenbe network), is an open network of G-queues first introduced by Erol Gelenbe as a model for queueing systems with specific control functions, such as traffic re-routing or traffic destruction, as well as a model for neural networks.
Erol and his colleagues gave up financial reward, in exchange for distributing this software freely to universities worldwide, so as to spread knowledge and social benefit. Now these methods are widely accepted standard approaches in the telecommunications and ICT industry.