Oil Layer Recognition by Support Vector Machine Based on Quantum-behaved Particle Swarm Optimization ⋆

In order to overcome the problems of the slow training speed and low recognition accuracy in the oil layer recognition, an improved method for oil layer recognition is put forward based on Support Vector Machine (SVM), which uses the Sequential Minimal Optimization (SMO) algorithm to train SVM model, and uses the Quantum-behaved Particle Swarm Optimization (QPSO) algorithm to optimize the parameters of SVM. Actual application in an oilfield in Xinjiang Region shows that the effect of oil layer classification is very good, and the QPSO-SVM is superior to the PSO-SVM and the improved PSO-SVM.