基于改进PSO算法的LS-SVM油层识别模型

为解决常规油层识别方法因其本身缺陷而无法取得理想效果的缺点，提出一种基于改进PSO算法的LS-SVM油层识别模型，即综合已有改进的PSO模型提出一种新的改进形式，并用此算法迭代求解LS—SVM中出现的矩阵方程，从而避免矩阵求逆，加快LS-SVM算法的训练速度，节省内存，而且求得最优解．实际应用表明，所提出的识别模型优于BP模型和经典SVM模型，识别精度高、收敛速度快、效果显著．

Oil layer recognition model of LS-SVM based on improved PSO algorithm

To solve the intrinsic shortcomings in general oil layer recognition methods,which is hard to obtain an ideal effect in application, the oil layer recognition model of LS-SVM based on the improved PSO algorithm is presented to iteratively solve the linear system of equations in LS-SVM algorithm, which is a new improved form by synthesized the existing model of PSO. By using the improved LS-SVM algorithm, the problem of solving inverse matrix is resolved and the training velocity of LS-SVM algorithm is quickened. Memory is saved and the least square solution is always got. The actual apptication shows that the improved LS-SVM model is superior to BP and SVM modet in oit layer recognition, which not only has much greater accuracy, but also improves the velocity of convergence.