基于精英保留策略遗传算法的生物电阻抗特征参数提取-《中国医学物理学杂志》

文章信息/Info

Title:: Extraction of bio-impedance characteristic parameters through genetic algorithm based on elitist preservation strategy

Author(s):: XU Yifan; CAO Le; HUANG Jingwei; KAN Xiu; School of Electronic and Electrical Engineering, Shanghai University of Engineering Science, Shanghai 201620, China

Keywords:: Keywords: bio-impedance spectrum Cole-Cole model least absolute deviation elitist preservation strategy genetic algorithm

摘要:: 由于Cole-Cole模型可以对生物组织和生化物质特性进行有效表征，使得如何从实测生物阻抗频谱（BIS）数据中准确提取模型参数具有重要研究意义。针对传统遗传算法具有收敛速度慢、易陷入局部最优解的问题，分析种群个体特别是精英个体的特征，使用最小一乘法重新定义适应度函数，并制定精英保留策略。对标准BIS数据集和实际测量情况进行分析，通过添加奇异点和随机噪声，建立符合实际情况的3类仿真数据集。基于仿真数据集进行3组实验，将精英保留策略遗传算法与其他3种算法的拟合结果进行比较，结果表明本算法具有拟合精度高、鲁棒性强的优点，为全面准确分析生物组织生理病理状态提供了有效工具。

Abstract:: Abstract: Since the Cole-Cole model can effectively characterize the biological tissues and biochemical substances, it is of great significance to accurately extract the model parameters from the measured biological impedance spectrum (BIS) data. In order to solve the problem that traditional genetic algorithm has slow convergence speed and is easy to fall into local optimal solution, the population individuals characteristics, especially elite individuals, are analyzed, the fitness function is redefined by the least absolute deviation (LAD), and the elite preservation strategy is formulated. The standard BIS data set and the actual measurement situation are analyzed, and by adding singular points and random noise, three types of simulation data sets that meet the actual situation are established. Three sets of experiments were performed based on the simulation data set, and the fitting results of the proposed genetic algorithm (EGA-LAD) considering the elite retention strategy are compared with the other three algorithms. The results show that EGA-LAD has high fitting accuracy and robustness, which provides an effective tool for comprehensive and accurate analysis of the physiological and pathological state of biological tissues.