Hybrid Hardware/Software Partitioning Methodology Based On Genetic And Negative Selection Algorithms

The traditional design methodology is unable to satisfy the needs of So C due to the smaller component size, higher integration level and more complex functions. So the hardware and software co-design methodology becomes a necessity. Hardware/Software partitioning is a key step during the hardware/software co-design. So the research on the methodology of hardware/software partitioning, describing the system and constructing the model of the system, improving the algorithm of partitioning has a significant meaning in both theory and practical application.This thesis introduces the area of hardware/software co-design and generalizes the domestic and abroad developing situation in the research field of hardware/software co-design. Then the methodology of constructing the mathematical model in the process of the design of embedded systems and the problems commonly existed in current partitioning technology are discussed. Based on the comparison and analysis of the features and advantages of the genetic algorithm and the negative selection algorithm in many terms, a hybrid algorithm is proposed based on genetic algorithm and negative selection, which obtains a better efficiency and ability of searching solutions by combing the advantages of genetic algorithm and negative selection algorithm. A few improvements have been made on the traditional cross operator and mutation operator to improve the quality and distribution range of the solution set. This thesis proposes a strategy that the self set eliminate parameter changes as the evolution process of the algorithm, which provides a better control and adjustment on the degree of eliminating individuals during the different stage of the evolution progress of the algorithm. At the end of this thesis,genetic algorithm, negative selection algorithm and the hybrid algorithm are respectively programmed and verified using the real data generated by TGFF tool. The verification results indicate that the hybrid algorithm can overcome the genetic algorithm ‘s disadvantage of the low speed of eliminating individuals and the negative selection algorithms ‘s disadvantage of weak searching ability in the initial stage of the evolution process, which makes the hybrid algorithm can rapidly obtain a solution set with better quality and distribution range.