Distributed Caching Strategy on Protein Folding

From Distributed Information and Intelligence Analysis Group

This research focuses on improving time performance of evolutionary computing in solving complex optimization problems by reusing piecewise solutions intelligently. Our strategy to reduce computational load required for evaluating fitness function during evolutionary process has been developed by local cache as well as distributed caching migration. Although our practical setting is a form of genetic algorithms, it can be widely used for various types of evolutionary computing. Previously, we applied the idea into square and triangular HP protein folding problems through distributed systems as well as serial systems. Our future research plan is to apply this idea into other problems with higher time bounds and investigate the impact of the strategy on solving those problems.