Search

Accurate and timely traffic flow forecasting is the key technology to address the issue of urban traffic congestion, which is significant for intelligent transportation system. However, it is a quite challenging task to develop an efficient, robust and automatically generated forecasting model. In this paper, we propose a novel model, ensemble forecasting architecture based on non-negative constrained sparse autoencoder and extreme learning machine(NCAE-ELM-EA). This model is designed using NCAE to extract traffic characteristics layer-by-layer through a greedy unsupervised algorithm, and the prediction network formed by connecting NCAE and ELM serves as base learner of each lag pool. Then, employing adaptive enhanced integration algorithm to build an ensemble optimized forecasting model with scalable size suitable for each road segment. The model was applied to the actual data collected from I5 NB highways in Portland, USA, and J6-J7(N) freeways in the United Kingdom, with higher accuracy and lower labor costs compared with existing predictors.

In the TD-SCDMA system, there is a problem that non-emergency voice services probability of dropped call is not ideal when we use the dynamic allocation algorithm. To solve this problem, we propose a new algorithm, which is called protect channel reservation queue movable boundary dynamic channel allocation. This algorithm set aside channels for dropped calls of non-emergency voice services service, in order to keep blocking probability for service, the introduction of no patient customers, combined with the data packet to borrow algorithms. Simulation results show that, comparing to the movable boundary algorithm, the proposed algorithm reduce the drop probability for non-emergency voice services, and maintain a variety of service blocking probability, reducing the packet loss probability data services.

To solve the problem that the parameters in grey neural network (GNN) are difficult to determine, the improved Particle Swarm Optimization (IPSO) algorithm is employed to search the optimums by the introduction of a threshold of velocity. When the particle velocity is less than the threshold, an accelerated momentum is applied on the particle to reinitialize the particle velocity and position. The proposed approach is used to predict the telephone traffic of two regions. The forecasting results are compared with those of GNN, Grey Neural Network optimized by Particle Swarm Optimization (PSO-GNN) and Back-Propagation Neural Network (BPNN). The experimental results show high prediction accuracy.

In order to make accurate forecast of mobile Internet diffusion trend. This paper proposes a method which is based on Bass innovation diffusion model. To solve the problem that the parameters of the model are difficult to estimate, a modified particle swarm optimization algorithm (PSO) whose inertia weight changes dynamically is introduced to search the most precise parameters. The application of the index of population density helps determine the convergence status of population. When the algorithm is trapped into local optimum, the combination of Cauchy mutation and Gaussian mutation is applied on the global best particle. The results of the experiment show that optimized Bass model is suitable for predicting mobile Internet diffusion trend and the proposed algorithm can effectively improve the precision of the optimal solution as compared to existed estimation methods.

In order to improve the prediction accuracy of busy telephone traffic which is influenced by multiple factors, this paper proposes a combined forecasting model which takes the influence of multiple factors into consideration and combines three models ——wavelet transform, autoregressive integrated moving average (ARIMA) model and least squares support vector machines (LSSVM) model, LSSVM is optimized by particle swarm optimization (PSO). Correlation analysis is firstly applied to the busy telephone traffic data to obtain the key factors which influence the busy telephone traffic. Then wavelet transform is used to decompose and reconstruct the telephone traffic data to get the low-frequency and high-frequency components. The low-frequency component is loaded into ARIMA model to predict, while the high-frequency component and the obtained key factors are loaded into PSO-LSSVM model to predict. Finally the forecasting result is achieved by the superposition of predictive values. The simulation results show that the proposed model has higher prediction accuracy and strong generalization ability.

This study presents a busy telephone traffic prediction model that combines wavelet transformation and least squares support vector machine. Firstly, decompose preprocessed telephone traffic data with Mallat algorithm and get low frequency component and high frequency component. Secondly, reconfigure each component and use LS_SVM model to predict each reconfigure one. Then the traffic can be achieved. The results of experiments have testified higher prediction accuracy and stability of this combined traffic prediction model.

In this paper, an extreme learning machine (ELM)-based improved layered ensemble architecture (EILEA) for time series forecasting is proposed. Compared with multilayer perceptron (MLP)-based layered ensemble architecture (named LEA), our proposed structure has improved in two aspects. First of all, we proposed an inflection point estimation-based density peaks clustering algorithm to replace K-means algorithm used by LEA, which can automatically determine the number of clusters without the influence of the choice of initial cluster centers on the clustering results. Second, EILEA employs ELM as individual learners, taking advantage of its high learning speed to greatly improve the prediction speed. The proposed architecture has been tested on the time series data of neural network (NN)3 competition. In terms of forecasting accuracy and speed, the experimental results have revealed clearly that our proposed structure is better than LEA.

Vertical handoff is of great significance to achieve seamless connectivity in heterogeneous networks. It is necessary to enhance the location awareness of the mobile station (MS) as well as to reduce the adverse effect of fading so that it is able to predict future network conditions accurately. An optimal Savitzky-Golay filtering based vertical handoff algorithm is adopted in this essay, which equips the MS to decide network coverage boundaries accurately. Simulation results show that this method can reduce the effects of fading in the received signal, thereby reducing the ping-pong effect availably. It can also improve the performance of the network.