Improvement of Parallal-Distributed Genetic-Algorithm VLSI55

点击翻译按钮获取中文摘要

Genetic Algorithms (GA) is a probabilistic technique for search and optimization which was inspired from life evolution. It is general purpose, and applicable to wide variety of problems, so it has been collecting attentions. To accelerate GA process, two techniques are the most promising : parallelization and VLSI implementation. We have designed GA-VLSI in a domestic pioneering research, and first developed parallel and distributed design of GA-VLSI.This research is to improve this parallel and distributed GA-VLSI further by introducing a dynamic adaptation technique for inter-cluster migration. This technique was proved to effective in our past research in software implementation on top of workstation network. This is the first attempt in the world.The research has a diverse background of hardware design, parallel and distributed computation, dynamic optimization and problem solving, so we needed the newest technique in each area. We investigated each area, such as hardware design methodologies and dynamic distributed computation, so as to integrate them into our research. Then, we achieved system design, circuit design, hardware simulation and evaluation, and logic synthesis. We obtained a good results out of them.