Project outline

The growing demand for connectivity prompts the wireless communications research community to continuously innovate. The novel techniques which are explored to increase this connectivity allow a better use of the available frequency bands. The number of simultaneous communications on these frequential bands is growing steadily.

As a result, the information volume to be processed is growing constantly. The communication chips’ processing power therefore needs to be increased. That is why the industry is orienting itself towards multi-processor systems-on-chip (MPSoC). These complex systems-on-chip allow integrating several programmable processors, specialized memory and various intellectual properties so as to optimize performance at a low cost.

The research efforts conducted by Professors Gagnon and Thibeault have shown the importance of partitioning the tasks within a signal processing chip. The partitioning stage consists in establishing some rules that will distribute the calculations in time among the various processors. This stage is one of the primary bottlenecks during the design of a wireless communications system. Because of the strong competition on this market, designers lie in wait for any tool that could shorten the design process while optimizing performance.

The purpose of this research project is to explore a task partitioning method based on evolutionary algorithms, which are capable of solving complex problems. These algorithms, which are based on the evolution theory, offer a great potential to study a wide range of unexplored solutions and to choose the best of them.

Download the scientific overview (pdf)

Expected results

• Explore and develop a partitioning optimization approach based on evolutionary algorithms.
• Exploration of many partitioning techniques for a reduced set of criteria.
• Optimal use of the available material resources.
  

Our partners

• NSERC

Team members

Professors: François Gagnon and Claude Thibeault
Specialist: Pascal Giard
Post-doc researcher: Marc Joliveau
Students:Mohamed El Euch, Olivier Alnet, Tarek Baallal, François Gourlay

CREDITS — © COPYRIGHT 2007 –INFO@COMUNITE.CA