基于IPSO混沌支持向量机的网络流量预测研究

针对传统混沌支持向量机参数寻优算法的不足,提出了一种改进的粒子群(IPSO)算法。该算法通过延长迭代的开始阶段和最后阶段的搜索时间,实现了算法的全局搜索与局部搜索能力之间的平衡,进而优化模型参数,建立了基于IPSO优化的混沌支持向量机预测模型。应用实例结果表明,该模型对网络流量预测是有效可行的,并具有较高的寻优效率、预测精度和较好的稳态性能。

基于启发式算法的混沌支持向量机流量预测

针对现有混沌支持向量机回归模型存在流量预测效率低下的问题,利用差分进化(DE)算法、遗传算法和粒子群优化算法确定模型的径向基核函数系数、惩罚系数、不敏感系数等参数,在此基础上建立改进的混沌支持向量机回归模型进行流量预测。实例表明,相比其他启发式算法,DE算法能以较高的效率搜索到混沌支持向量机回归模型的最优参数,并且该模型具有较高的预测精度。

基于混沌SVM与BP神经网络预测材料价格指数

为了解决工程造价指数难以预测非线性结构、数据拟合难度大、预测模型参数求解过于固定化、预测模型可靠性不高等问题,文中在混沌时间序列理论的基础上,结合机器学习算法支持向量机(SVM)技术和BP神经网络算法,提出混沌SVM与BP神经网络组合预测模型。实例研究证明,该组合预测模型的精度比SVM预测模型、混沌SVM预测模型、BP神经网络预测模型和GM(1,1)预测模型的高,具有拟合非线性和预测线性波动的能力,可用于工程造价指数预测。

基于优化蚁群算法的钢轨轮廓识别

针对传统蚁群算法在钢轨图像识别中存在的问题,对蚁群算法进行4个方面的优化。初始化过程优化:采用一维Logistic混沌映射序列非线性迭代方程,使蚁群的初始化分布更加均匀,以避免大量的无关运算;搜索过程优化:在蚁群的搜索初期采用随机搜索策略,根据图像灰度梯度值自动设置阈值,初步确定图像中钢轨边缘的像素点,然后建立区域搜索模型,以进行钢轨边缘的精确搜索和描绘;搜索步长优化:在搜索初期,采用大步长随机搜索策略识别钢轨边缘的像素点,然后利用小步长区域搜索策略对钢轨边缘像素点做更精确地识别,从而实现钢轨轮廓的精确识别,并减少了搜索时间和算法的收敛时间;信息素更新策略优化:每完成1次搜索,根据自动设置的信息素最大、最小浓度值更新信息素,防止陷入局部最优。对实际采集到的直线和曲线线路上的钢轨图像分别用Canny边缘检测算子、传统算法和优化算法进行钢轨轮廓识别的对比试验,结果表明:优化算法具有更好的健壮性和识别效率。

奇异谱分析在中长期径流预测中的应用研究

针对目前中长期径流预测精度较低的问题,运用奇异谱分析法对隔河岩水库1951～2009年入库径流资料进行预处理得到重构序列,以达到浓缩主要信息和减小误差的目的。分别采用自回归模型和混沌支持向量机模型对原始序列和处理后的重构序列进行模拟预测。结果表明,应用奇异谱分析法进行资料的预处理可以大大提高中长期径流的预测精度。

一种新型煤灰分双能量γ射线检测方法

针对传统方法对煤灰分检测误差大的问题,提出基于混沌最小二乘支持向量机(chaos-LSSVM)的煤灰分双能量γ射线检测方法。其中,双能量γ射线透射法可减小煤炭形状、厚度、粒度、堆密度等因素引入的检测误差,而最小二乘支持向量机算法可减小标定误差,混沌算法可优化最小二乘支持向量机计算进程中惩罚系数g和核函数宽度参数δ。通过241Am和137Cs作为低能和中能γ射线源煤灰分的实验验证,Chao-LSSVM检测方法灰分平均相对误差可达到0.80%,与传统标定方法(直线逼近、最小二乘逼近方法)煤灰分检测的平均相对误差2.22%和3.19%相比,本文提出的方法具有优化的煤灰分检测准确度。

基于预测和动态阈值的流量异常检测机制研究

针对流量异常检测中的基线和阈值难以精确刻画的问题,提出了一种基于预测和动态阈值的异常检测机制。通过构造混沌支持向量机预测模型对流量基线值进行确定;采用假设检验的异常检验方法,利用一天中各时段对应训练集拟合残差符合正态分布的特点构造符合t分布的随机变量,进而计算各时段预测残差的置信区间来动态地确定网络流量的阈值。实验结果表明,预测模型具有很高的预测精度,该异常检测机制具有一定的可行性。

城市轨道交通自动售检票系统实时进站客流量异常检测

阈值范围的合理设定,对AFC(自动售检票)系统实时进站客流量数据的有效性检测至关重要。采用改进的小数据量法计算历史进站客流数据时间序列的Lyapunov指数,验证该序列的混沌特性;利用C_C方法确定混沌时间序列的时间延迟和最佳嵌入维数,对原时间序列进行相空间重构,确定模型训练测试样本集;采用大范围网格搜索方法优化混沌支持向量机回归模型参数,利用优化后的模型预测各时段的进站客流量;利用训练样本集中各时段进站客流量预测残差序列的分布特性,确定在某一置信度下各时段进站客流量预测残差的置信区间,从而确定未来时段的进站客流量异常检测的阈值上限和阈值下限。试验结果表明,该方法有效地增强了系统对于AFC系统实时进站客流量数据有效性检测的能力。

Deformation prediction of tunnel surrounding rock mass using CPSO-SVM model

A new method integrating support vector machine (SVM),particle swarm optimization (PSO) and chaotic mapping (CPSO-SVM) was proposed to predict the deformation of tunnel surrounding rock mass.Since chaotic mapping was featured by certainty,ergodicity and stochastic property,it was employed to improve the convergence rate and resulting precision of PSO.The chaotic PSO was adopted in the optimization of the appropriate SVM parameters,such as kernel function and training parameters,improving substantially the generalization ability of SVM.And finally,the integrating method was applied to predict the convergence deformation of the Xiakeng tunnel in China.The results indicate that the proposed method can describe the relationship of deformation time series well and is proved to be more efficient.

Intelligent prediction on air intake flow of spark ignition engine by a chaos radial basis function neural network

To ensure the control of the precision of air-fuel ratio(AFR)of port fuel injection(PFI)spark ignition(SI)engines,a chaos radial basis function(RBF)neural network is used to predict the air intake flow of the engine.The data of air intake flow is proved to be multidimensionally nonlinear and chaotic.The RBF neural network is used to train the reconstructed phase space of the data.The chaos algorithm is employed to optimize the weights of output layer connection and the radial basis center of Gaussian function in hidden layer.The simulation results obtained from Matlab/Simulink illustrate that the model has higher accuracy compared to the conventional RBF model.The mean absolute error and the mean relative error of the chaos RBF model can reach 0.0017 and 0.48,respectively.

Support vector machine forecasting method improved by chaotic particle swarm optimization and its application

By adopting the chaotic searching to improve the global searching performance of the particle swarm optimization (PSO), and using the improved PSO to optimize the key parameters of the support vector machine (SVM) forecasting model, an improved SVM model named CPSO-SVM model was proposed. The new model was applied to predicting the short term load, and the improved effect of the new model was proved. The simulation results of the South China Power Market’s actual data show that the new method can effectively improve the forecast accuracy by 2.23% and 3.87%, respectively, compared with the PSO-SVM and SVM methods. Compared with that of the PSO-SVM and SVM methods, the time cost of the new model is only increased by 3.15 and 4.61 s, respectively, which indicates that the CPSO-SVM model gains significant improved effects.

Optimization of support vector machine power load forecasting model based on data mining and Lyapunov exponents

According to the chaotic and non-linear characters of power load data,the time series matrix is established with the theory of phase-space reconstruction,and then Lyapunov exponents with chaotic time series are computed to determine the time delay and the embedding dimension.Due to different features of the data,data mining algorithm is conducted to classify the data into different groups.Redundant information is eliminated by the advantage of data mining technology,and the historical loads that have highly similar features with the forecasting day are searched by the system.As a result,the training data can be decreased and the computing speed can also be improved when constructing support vector machine(SVM) model.Then,SVM algorithm is used to predict power load with parameters that get in pretreatment.In order to prove the effectiveness of the new model,the calculation with data mining SVM algorithm is compared with that of single SVM and back propagation network.It can be seen that the new DSVM algorithm effectively improves the forecast accuracy by 0.75%,1.10% and 1.73% compared with SVM for two random dimensions of 11-dimension,14-dimension and BP network,respectively.This indicates that the DSVM gains perfect improvement effect in the short-term power load forecasting.

Fuzzy least squares support vector machine soft measurement model based on adaptive mutative scale chaos immune algorithm

In order to enhance measuring precision of the real complex electromechanical system,complex industrial system and complex ecological & management system with characteristics of multi-variable,non-liner,strong coupling and large time-delay,in terms of the fuzzy character of this real complex system,a fuzzy least squares support vector machine(FLS-SVM) soft measurement model was established and its parameters were optimized by using adaptive mutative scale chaos immune algorithm.The simulation results reveal that fuzzy least squares support vector machines soft measurement model is of better approximation accuracy and robustness.And application results show that the relative errors of the soft measurement model are less than 3.34%.

Corrosion depth prediction based on non-linearity method

Pipeline of oil and gas have an increased risk because of pipeline punctures and rupture caused by corrosion. Therefore it is very important to have a reliable way for pipeline corrosion prediction. The corrosion depth prediction models that based on the support vector machines and chaos were introduced in this paper. A real example was given in this paper. The predicted results showed that the prediction models have a more higher precision. The two corrosion depth prediction models are reasonable in corrosion research, which can supply a scientific basis for pipeline safety management, service life prediction and repair.

Discrete-Time Chaotic Systems Synchronization Based on Vector Norms Approach

Suitable stabilization conditions obtained for continuous chaotic systems are generalized to discrete-time chaotic systems. The proposed approach, leading to these conditions for complete synchronization is based on the use of state feedback and aggregation techniques for stability studies associated with the arrow form matrix for system description. The results are successfully applied for two identical discrete-time hyper chaotic Henon maps with different orders and also for non-identical discrete-time chaotic systems with same order namely the Lozi and the Ushio maps.