Parallel Genetic Algorithm using Compute Unified Device Architecture (CUDA) for Designing Energy Saving Glass Coating Structure

摘要

Parallel Processing is proven to provide faster speed in generating the chromosome of the coating structure of Energy saving glass. The technique that has been used is parallel genetic algorithm, since parallel genetic algorithm require high speed of CPU in order to generate the result, yet high performance of CPU will require additional cost, with the use of GPU it is proven that the speed of the GPU is 100 times faster than the CPU which can shorten the duration in generating the code and result. The result will be in binary format which is the best chromosome that will be selected used as the design for the coating structure. This paper will cover the study of GPU and CUDA platform and analyze the result of the coating. The simulation result of the coating structure will be compare with different glass width or border and different frequency with same material and same coating structure design. This paper will discuss on the coating structure based on the simulation result whether the glass width can affect the signal frequency and which type of application that suitable with it. Speed test will also cover in this paper to prove that parallel processing does shorten the time of processing.