Chunfu Shao

Study on the Application of A* Shortest Path Search Algorithm in Dynamic Urban Traffic

This paper studies the application of A* shortest path search algorithm in dynamic urban traffic in the search of optimal path in real-time and dynamic traffic environment. At first, this paper introduces A* search algorithm and the characteristics of urban traffic. Then, the A* search algorithm application processes in optimal path searching of urban traffic is analyzed. The analysis focuses on the heuristic function of A* search algorithm. In the course of algorithm design, not only are the accuracy of the prediction of travel time and the characteristics of urban traffic considered, but also the factor of time similarity in urban traffic is taken into account.

Simulation of pedestrian evacuation with affected visual field and absence of evacuation signs

A simulation of pedestrian evacuation in room with affected visual field and absence of the direction of evacuation signs will be presented based on cellular automata (CA) in this paper. Evacuation room is divided into exit-visible area and exit-invisible area by pedestrian visual field radius. Two basic dynamic parameters in Dynamic Parameters Model (DPM) are used to simulate the varied movement characteristics of pedestrian in different evacuation areas including normal evacuation movement and blind random movement. The effect of pedestrian visual field radius on evacuation time is studied without evacuation signs on wall. It is observed that evacuation time is dependent not only on the pedestrian visual field radius but also on the initial density. It is also found that the evacuation time will tend to the level and remain unchanged with visual field radius rising.

Study on Distributed and Parallel Search Strategy of Shortest Path in Urban Road Network

This paper mainly focuses on distributed and parallel search strategy of shortest path in urban road network, which is motivated by the fact that in the shortest path search in complicated urban road network, the search time consumed will increase sharply with the increase of the number of road network nodes. In the algorithm presented in the paper, the whole search road network is first cut into local sub-search road networks, and connections among each local sub-search road network are established through virtual auxiliary nodes; then searches of the shortest paths in local sub-search road networks are conducted in a parallel manner. Finally, the shortest path search of the whole network is completed on the basis of local and parallel search by the virtual auxiliary nodes. The shortest path problem from single-origin to single-destination is studied and analyzed. The correctness of the algorithm is proved.

Short-Term Traffic Flow Forecasting Based on Periodicity Similarity Characteristics

The methodology has been putted forward that the periodicity similarity should be consideration when using Elman neural network (ENN) to forecast short-term traffic flow, which is not only helpful to save training time, reduce training sample size, but also enhance forecasting efficiency. Firstly, training sample of ENN has been designed based on the periodicity similarity of traffic flow and network structure has been established aiming at improving ENN global stability. Secondly, short-term traffic flow forecasting method based on ENN have been established by taking daily, weekly and monthly periodicity similarity into account respectively. Finally, forecasting results have been evaluated by four error statistics from two aspects: forecasting effect and efficiency. The conclusion has been summarized to three aspects.

Travel Mode Choice Modeling Based on Improved Probabilistic Neural Network

Travel mode choice can be regarded essentially as a problem of pattern recognition. In the past, numerous researchers applied artificial neural network (ANN) to travel mode choice modelling. Feed-forward back propagation neural network (BPNN) is widely used, and most of the studies show that BPNN presents better prediction accuracy and generalization capacity. Probabilistic neural network (PNN) is also known as a general solution to pattern classification problems by following an approach developed in statistics. Comparing to BPNN, its shorter training time and better stability increase the reliability of simulation results, even so PNN is barely used in travel behaviour analysis. Therefore the paper tries to apply PNN for travel mode choice modelling. First, network structure is established based on the data obtained from resident trip survey, and then the K-means cluster algorithm is applied to optimize the hidden node number so that they can be modified dynamically. Moreover, the data centers and extended constants of gaussian radial basis function are changed adaptively during the learning progress, and the improved PNN is called KPNN. To reveal the superiority of KPNN, BPNN is also introduced to make a comparison. For BPNN, the optimal number of hidden nodes is recognized, and then the well trained network is applied. By proposing two performance measures, the classification and predictive capability of KPNN and BPNN is evaluated, it is found that KPNN outperforms BPNN on both the simplicity and prediction accuracy. Anyway, the simulation results show that KPNN can deal with the problem of travel choice modelling as well as, if not better than BPNN.

Application of wavelet neural networks for trip chaining recognition

The article develops a wavelet neural network for trip chaining pattern recognition. Based on the data obtained from Beijing Resident Trip Survey, a set of socioeconomic and demographic factors related to the of traveller situation which potentially influence trip-chaining patterns are selected as input variables of neural network, and a categorical trip chaining pattern (simple and complex trip chaining) are used as output variables. In order to quantify prediction accuracy, two performance measures are applied to evaluate it. Besides, BP neural network and a logistic regression model are also introduced to make a comparison, and the conclusions indicate wavelet neural network performs much better in convergence rate and prediction accuracy; actually its generalization capability is much better too.

Simulation of pedestrian evacuation from room with internal obstacles

A simulation of pedestrian evacuation from room with internal obstacles is presented based on an improved Dynamic Parameter Model (DPM) in the paper. In order to reduce evacuation imbalance caused by the asymmetry of obstacles layout, a special technology is introduced to compute the shortest estimated distance from anyone cell site to exits in room, which determines two basic dynamic parameter: direction-parameter and empty-parameter. Actual and imaginary distances are merged into the shortest estimated distance through considering the effects of pedestrian jam around obstacles on evacuation path selection. The pedestrian evacuation flows with single exit and multi-exit are simulated under different obstacles layouts from two aspects: with fixed and unfixed initialization site. It is observed that improved model can effectively reduce evacuation imbalance.

Traffic Modal Splitting Model Based on Game Theory with Incomplete Information

Traffic modal splitting plays an important role in the traditional four-stages-forecast model. Usually in a stable urban traffic network, traffic supply can meet the demand basically, and a variety of traffic modal will bear a fixed proportion of passengers. Compared with other trip mode, public transportation is often influenced by some uncertain factors, e.g. transferring and waiting, which will cause different marginal cost and finally come to a new balance or imbalance between demand and supply. In this paper, traffic modal was supposed as a rational gamer, and generalized trip cost was considered as the cost of gamer. Then, traffic mode share can finally be got via comparing supply of all kinds of traffic means. A case study is conducted and compared with disaggregate modal, which shows the game theory-based model is more effective and the outcome is closer to the real traffic modal share.