论文部分内容阅读
浮点遗传算法是一种模拟生物进化的最化搜索法 ,由于其运算简单、稳定性好、不需要计算目标函数的导数、高精度和能处理多维数值问题 ,浮点遗传算法在科学研究和工程技术中得到了广泛应用 .通过对浮点遗传算法收敛性的分析 ,本文证明了“简单浮点遗传算法不收敛于全局最优解 ,而每代保留最优个体的浮点遗传算法才收敛于全局最优解” .在此基础上 ,本文设计了一种采用连续突变和每代保留最优个体的改进浮点遗传算法 ,它克服了精确度与计算量之间的矛盾 .本文利用该算法较好地解决了半导体器件模型参数提取问题 ,使计算量降低了约 2 7%
Floating-point genetic algorithm is a kind of simulation of biological evolution of the most search method, due to its simple operation, good stability, without the need to calculate the derivative of the objective function, high precision and can handle multi-dimensional numerical problems, floating-point genetic algorithm in scientific research and Engineering technology has been widely used.By analyzing the convergence of floating-point genetic algorithm, this paper proves that “the simple floating-point genetic algorithm does not converge to the global optimal solution, and each generation retains the best individual floating-point genetic algorithm converges In the global optimal solution. ”On this basis, we design an improved floating-point genetic algorithm that uses continuous mutation and preserves the optimal individuals for each generation, which overcomes the contradiction between accuracy and computational complexity.In this paper, The algorithm solves the problem of extracting the parameters of the semiconductor device model well and reduces the computation amount by about 27%