中文摘要：

精确的短时交通预测是建立智能交通系统的一个重要前提,而具有明显周期性特点的交通流量的预测是其中的一个重要环节。为实现交通流量的准确预测,提出一种基于自适应惯性权重的粒子群优化（AωPSO）最小二乘支持向量机（LS-SVM）的短时交通流量预测方法,通过引入粒子种群多样性,设计自适应惯性权重调节方法,借助PSO算法的寻优能力实现LS-SVM参数的优化,减少人为因素对参数选择的影响,提高LS-SVM的泛化能力和预测精度。实验结果表明,与BP网络、LS-SVM等方法相比,该方法具有精度高、泛化能力强的特点。

英文摘要：

Precise short-term traffic prediction is one of the important prerequisites for establishing the intelligent transportation systems, and the prediction of traffic flow with obvious periodicity is one important part. In order to achieve accurate traffic flow prediction, a short-term traffic flow predicting method is proposed based on the least-square support vector machine （LS-SVM） optimized by particle swarm with an adaptive inertia weight. After designing the swarm diversity function of particles and adaptive inertia weight strategy, the parameters of LS-SVM are optimized without any more human participation. The generalization and accuracy of LS-SVM are improved based on the control of swarm diversity control and inertia weight adaptive adjustment. Experimental results demonstrate that, the proposed method can obtain a higher accuracy than other common neural network and LS-SVM methods.