基于现场可编程门阵列实现二维光栅优化设计算法

Realization of FPGA-based optimum design of two-dimensional grating

提出了一种基于现场可编程门阵列实现的遗传算法对二元光栅直接进行二维优化设计的方法。采用二次多项式函数描述二元光栅曲面的面形,给出了基于现场可编程门阵列的遗传算法优化光栅的解决方案;以多项式系数为优化设计对象,选用了适合本设计的硬件实现编码、选择、交叉、变异算子、适应度计算算法,同时引进了精英保存策略来提高程序的健壮性和加快收敛速度。算法充分考虑硬件处理的并行性和流水线特点,利用Verilog HDL语言编程,在Altera的cycloneⅡEP2C50器件上实现。结果表明,该算法对二元光栅优化设计计算速度比软件实现的快四十倍以上,有效提高了二维二元光栅优化设计的速度。

A method of Field Programmable Gate Array（FPGA） - based Genetic Algorithm （GA） for directly optimizing two - dimen- sional （2 - D） binary gratings is proposed. The surface of binary grating is described as a quadratic polynomial function and the solution of grating optimization with FPGA - based GA is given. Taking polynomial function coefficient for the object of optimum, the a- daptive coding, selection, crossover, mutation and the fitness calculation are selected and the elite preservation strategy is introduced to improve the program＇s robustness and increase the convergence speed. Verilog HDL language is used to program the algorithm after fully considering the parallel and pipeling characteristics of the algorithm, and then the algorithm is implemented in cycloneⅡ EP2C50 of Altera. The results show that the speed of this binary grating optimum design is more than forty times faster than that of realization by software. The speed of 2 - D binary grating optimum design is raised effectively.