[CCoE Notice] Friday seminar talk announcement: GPU-accelerated Meta-heuristics for Optimization

[Lewis, Lindsay R]

GPU-accelerated Meta-heuristics for Optimization

Dr. Weihang Zhu

Department of Industrial Engineering

Lamar University

When: 10 -11am, Friday, October 8, 2010

Where: Room 102D, Engr. Bldg 1

Abstract:

This seminar will present the interdisciplinary research work on GPU-accelerated meta-heuristics for optimization. A framework of massively parallel evolutionary algorithms with graphics hardware acceleration was established for global optimization and multi-objective optimization problems. GPU, the common graphics hardware found in modern personal computers (PC), can be used for data-parallel computing in a desktop setting. In this research, multiple population-based evolutionary algorithms were adapted to the data-parallel GPU computing platform featuring 'Single Instruction - Multiple Thread' (SIMT) to achieve orders of magnitude speedup with affordable hardware. Applications of the framework on classical problems such as Quadratic Assignment Problems (QAP) and new fields such as protein structure modeling (Computational Biology) will be discussed. The research result demonstrated a promising direction for high speed optimization with desktop parallel computing in a PC.

Short Bio:

Weihang Zhu (Ph.D., North Carolina State University, 2003) is an assistant professor of Industrial Engineering, directing Computational Optimization Lab, and Manufacturing Lab at Lamar University, Beaumont, Texas. He has more than 15 years of engineering research experience and has participated in the projects supported by NSF, NIST and DoD. He has about 50 technical publications on GPU computing, metaheuristics, design optimization, computational biology, CAD/CAM, medical simulation and haptics.

Gino Lim

Hari and Anjali Agrawal Faculty Fellow and Associate Prof.

Director of Science & Technology, SWTC Center

Department of Industrial Engineering

University of Houston

http://www.egr.uh.edu/ie/faculty/?e=lim
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://Bug.EGR.UH.EDU/pipermail/engi-dist/attachments/20101006/cfea5f04/attachment.html