In today’s fast-developing information technology landscape, neural network technology is applied across various industries, playing a crucial role in various predictions in public transport, mainly in vehicle travel time prediction, passenger flow prediction, and comprehensive evaluation prediction of public transport. This enables accurate predictions in public transport and enhances the management level of public transport services.
1. Overview
Artificial Neural Networks (ANNs) are a mathematical model of algorithms that mimic the behavioral characteristics of animal neural networks for distributed and parallel information processing. This network relies on the complexity of the system, adjusting the relationships between numerous interconnected internal nodes to achieve the goal of information processing, with the ability to self-learn and adapt. Public transport operations include information on vehicles, routes, and stations, making it a typical neural network structure. It integrates with intelligent scheduling systems to achieve reasonable adjustments in operational status, ensuring balanced and orderly vehicle operations, maximally satisfying passenger travel demands.
Neural networks are non-linear, adaptive information processing systems composed of numerous interconnected processing units. It is proposed based on the achievements of modern neuroscience research, attempting to process information by simulating how the brain’s neural networks handle and memorize information. Artificial neural networks have four fundamental characteristics: non-linearity, non-convexity, very qualitative, and non-limited.
2. Principles of Neural Network Algorithms
A neural network consists of an input layer, multiple hidden layers, and an output layer. The critical work of the neural network is to design the hidden layers and the weights between the neurons; adding a few hidden layers results in a shallow neural network (SNN); many hidden layers form a deep neural network (DNN). Due to the simplicity of operation and significant effects of neural network technology, it has been successfully applied in many fields. The structure of the neural network is as follows:
Neural networks are non-linear, adaptive information processing systems composed of numerous interconnected processing units. It is proposed based on the achievements of modern neuroscience research, attempting to process information by simulating how the brain’s neural networks handle and memorize information. Artificial neural networks have four fundamental characteristics: non-linearity, non-convexity, very qualitative, and non-limited.
3. Applications of Neural Network Algorithms
(1) Prediction of Bus Travel Time
The prediction of bus vehicle travel time is an essential part of the intelligent scheduling system. Applying neural network algorithms to predict bus arrival times can provide passengers with more accurate waiting times, thereby enhancing the quality of public transport services.
Using intelligent scheduling location data to model the neural network enables accurate predictions of bus vehicle travel times. First, initialize the selected sample data, then use the backpropagation algorithm to adjust the weights and biases of the network, making the output vector as close as possible to the expected vector. The neural network is composed of multiple individual neural networks, selecting prediction methods based on the prediction problem, mainly including regression problem predictions and classification problem predictions. The flowchart of the neural network algorithm is as follows:
(2) Passenger Flow Prediction
Implement the passenger flow prediction function based on the neural network modeling approach. First, identify the factors affecting passenger flow and set parameters for the initial data, including quantifiable factors such as the number of vehicles, population, and resident income, as well as non-quantifiable factors like the environment and urban characteristics. Then, input the network, preprocess both quantifiable and non-quantifiable data, normalizing it, such as linearization, calculating the predicted values, and then transforming according to the reverse rules. After preprocessing the data, proceed with genetic neural network training.
(3) Comprehensive Evaluation Prediction of Public Transport
Predict the results of the comprehensive evaluation model of public transport through neural networks, thereby enhancing the public transport planning evaluation system and urban management level. Neural networks are a multi-layer feedback grid that can better predict public transport evaluation results. First, input data is transmitted to the hidden layer nodes, analyzed and computed, and then the results are sent to the output nodes, ultimately outputting the results.
First, select evaluation index parameters, construct the evaluation model, and then apply neural network algorithms to process the model, planning data parameters, and implementing the network algorithms to better simulate the comprehensive evaluation process of public transport, ultimately completing a comparative analysis between the predicted values and evaluation values.
4. Conclusion
Public transport companies achieve intelligent management through neural network algorithm technology, processing and integrating collected data to achieve precise vehicle predictions and dynamic vehicle management, reducing the blind spots in public transport planning and management, and providing better technical support for management decision-making.
References
[1] Peng Xinjian, Weng Xiaoxiong. Prediction of Bus Travel Time Based on Firefly Algorithm Optimized BP Neural Network.
[2] Xia Guirong. Research and Application of Neural Network Integration in Bus Travel Time Prediction.
[3] Li Shuqing, Jiang Haiyan, Wang Fang. Research on the Application of Genetic Neural Networks in Bus Passenger Flow Prediction.
[4] Chen Guifu, Chen Danni. Research on the Comprehensive Evaluation Model of Nanping Public Transport Based on BP Neural Network.
[5] Baidu Encyclopedia.
Reference
END
About Us
Our center mainly engages in the development and application of advanced concepts, new products, and new technologies for urban public transport at home and abroad; investigation, collection, analysis, and prediction of basic data for urban transportation; planning and design of urban transportation systems and special studies; research on transportation economics and policies; construction of public transport standards system; research on the management theory system of public transport enterprises; consulting and evaluation of dual prevention system construction for transportation enterprises; research and development of computer software and hardware; design and development of intelligent transportation systems; training, meetings, and exhibitions.
Scan to follow us!
