改进量子多目标进化算法用于SOC软硬件划分

Division of hardware and software of SOC system based on improved quantum multi-objective evolutionary algorithm

采用量子多目标进化算法对从任务级进行抽象建模所得到的系统模型进行软硬件划分,并针对SOC系统设计中存在的特点,对量子多目标进化算法进行改进。采用量子个体编码方案,避免个体编/解码的冗余。并将Pareto最优概念与多目标优化相结合,从而实现了兼顾系统面积、功耗、时间等参数的软硬件划分方法。仿真对比实验结果表明,该算法一次运行可以获得多个Pareto最优解,为各个目标函数之间权衡分析提供了有效的工具,提高了设计效率。在满足系统性能要求下,可为复杂SOC系统提供多个设计目标的全局优化方案。

This paper used quantum multi-objective evolutionary algorithm to division of software and hardware of SOC system,which model obtained by abstract from task-level. Improved the quantum multi-objective evolutionary algorithm for the SOC system design features. The quantum individual coding schemes was to avoid the redundancy of individual codec. And the concept of Pareto optimal combination with the multi-objective optimization in order to achieve a balance between system size, power and time parameters of the hardware and software division method. Simulation comparison experiments show that the algorithm can get more than one run Pareto optimal solution,for each trade-off analysis between the objective function provides an effective tool to improve the design efficiency. To meet system performance requirements,it can be more complex SOC system design goals of the global optimization program.