Method and device for optimizing traffic network based on large model

Abstract

The invention particularly relates to a method and a device for optimizing a traffic network based on a large model. According to the method and the device for optimizing the traffic network based on the large model, traffic data are collected and preprocessed, and a Transform model framework is adopted to train the traffic data in a mode of combining unsupervised learning and reinforcement learning so as to predict a future traffic state; according to a traffic flow prediction model based on LSTM and a knowledge graph, a traffic signal lamp timing intelligent optimization algorithm based on a Grey Model model is adopted to intelligently adjust a traffic signal lamp control strategy according to a traffic flow prediction value, traffic flow distribution is optimized, and dynamic optimization of a traffic network is realized. According to the method and the device for optimizing the traffic network based on the large model, dynamic optimization of the traffic network is realized by analyzing traffic data in real time, intelligently adjusting a traffic signal lamp control strategy and optimizing traffic flow distribution, and the efficiency and the safety of an urban traffic network, an expressway and a public traffic system can be remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of large models, and in particular to a method and device for optimizing a traffic network based on a large model. Background Art

[0002] With the rapid development of urbanization, traffic congestion has become a common problem faced by major cities around the world. Traditional traffic management methods, such as timed signal light control and manual dispatching, can no longer meet the growing traffic demand.

[0003] In recent years, the rapid development of artificial intelligence technology, especially the application of large model technology, has provided new solutions for traffic network optimization. With its powerful data processing and deep learning capabilities, large models have achieved remarkable results in image recognition, natural language processing and other fields, and their application in traffic network optimization has also shown great potential.

[0004] In order to optimize the traffic network, improve traffic efficiency, reduce congestion, and enhance travel safety, the present invention proposes a method and device for optimizing the traffic network based on a large model. Summary of the invention

[0005] In order to make up for the defects of the prior art, the present invention provides a simple and efficient method and device for optimizing the traffic network based on a large model.

[0006] The present invention is achieved through the following technical solutions:

[0007] A method for optimizing a transportation network based on a large model comprises the following steps:

[0008] Step S1: Data collection and feature extraction

[0009] Collect traffic flow monitoring data, weather information, and customized special event information related to traffic, and pre-process the collected data, including missing value processing, normalization processing, and time window division of time series data;

[0010] The self-attention mechanism of the Transformer model framework is used to process sequence data. The pre-processed data is trained through a combination of unsupervised learning and reinforcement learning to predict future traffic conditions.

[0011] Step S2: Traffic flow prediction model based on LSTM and knowledge graph

[0012] Step S2.1, collect traffic flow data, including historical traffic flow, road checkpoints, points of interest (POIs), and meteorological data, and design static and dynamic urban traffic knowledge graphs based on road traffic information and external factors to capture the semantic relationship between road traffic information and external factors;

[0013] Step S2.2, input the urban traffic knowledge graph into the relational evolution graph convolutional neural network RE-GCN, and use the relational evolution graph convolutional neural network RE-GCN to embed knowledge into the knowledge graph;

[0014] Step S2.3, using the long short-term memory network LSTM to predict traffic flow and capture the spatiotemporal relationship of traffic flow data;

[0015] Step S2.4, the traffic flow data is integrated with the information extracted from the urban traffic knowledge graph, and then input into the spatiotemporal multi-graph convolution module for training to extract spatiotemporal features;

[0016] Step S2.5, output the traffic flow prediction value through the fully connected layer of the traffic flow prediction model, and evaluate the performance of the traffic flow prediction model on a real traffic flow dataset;

[0017] Step S3: According to the traffic flow prediction value, the traffic light timing intelligent optimization algorithm based on the Grey Model model is used to intelligently adjust the traffic light control strategy, optimize the traffic flow distribution, and realize the dynamic optimization of the traffic network.

[0018] In step S1, traffic data is collected from three data sources, namely surveillance cameras, sensors and GPS, to obtain a comprehensive view of the traffic network;

[0019] The missing values ​​in the data are processed by the mean filling method; the input data is normalized so that the feature values ​​of different dimensions have the same scale; the time series data is divided into appropriate time windows to provide contextual information;

[0020] Extract key features from the preprocessed data, including traffic flow, speed, and accident information.

[0021] In step S3, the intelligent optimization algorithm for traffic light timing based on the Grey Model is used to intelligently adjust the traffic light control strategy, and the implementation steps are as follows:

[0022] Step S3.1, collect historical traffic flow data and record them in units of 15 minutes; use the accumulation generation method to accumulate the original data once to form a new data sequence;

[0023] Step S3.2, establish the Grey Model (1,1) model, and fit the model parameters by the least squares method;

[0024] Step S3.3, comprehensively considering actual delay, effective green time and saturation factors, and establishing a fitness function;

[0025] Step S3.4, using a genetic algorithm to optimize the timing of traffic lights, randomly generating a number of traffic light timing schemes as an initial population, and evaluating the fitness of each scheme according to a set objective function;

[0026] Step S3.5: Generate a new generation of population through roulette wheel selection, crossover and mutation operations, and iterate until the custom-set termination condition is reached.

[0027] The Grey Model-based intelligent optimization algorithm for traffic light timing has a built-in adaptive learning mechanism that can dynamically adjust the learning rate and model parameters according to real-time traffic conditions to adapt to changes in traffic flow.

[0028] A device for optimizing a traffic network based on a large model, comprising a data collection and feature extraction module, a traffic flow prediction model and an intelligent optimization algorithm module;

[0029] The data collection and feature extraction module is responsible for collecting traffic flow monitoring data, weather information, and customized special event information related to traffic, and preprocessing the collected data, including missing value processing, normalization processing, and time window division of time series data;

[0030] The self-attention mechanism of the Transformer model framework is used to process sequence data. The pre-processed data is trained through a combination of unsupervised learning and reinforcement learning to predict future traffic conditions.

[0031] Traffic flow prediction model, responsible for traffic flow prediction based on LSTM and knowledge graph;

[0032] The intelligent optimization algorithm module is responsible for intelligently adjusting the traffic light control strategy according to the traffic flow prediction value, using the traffic light timing intelligent optimization algorithm based on the Grey Model model, optimizing traffic flow distribution, and realizing dynamic optimization of the transportation network.

[0033] The data collection and feature extraction module collects traffic data from three data sources: surveillance cameras, sensors, and GPS to obtain a comprehensive view of the traffic network;

[0034] The missing values ​​in the data are processed by the mean filling method; the input data is normalized so that the feature values ​​of different dimensions have the same scale; the time series data is divided into appropriate time windows to provide contextual information;

[0035] Extract key features from the preprocessed data, including traffic flow, speed, and accident information.

[0036] The traffic flow prediction model performs traffic flow prediction, and the implementation steps are as follows:

[0037] Step S2.1, collect traffic flow data, including historical traffic flow, road checkpoints, points of interest (POIs), and meteorological data, and design static and dynamic urban traffic knowledge graphs based on road traffic information and external factors to capture the semantic relationship between road traffic information and external factors;

[0038] Step S2.2, using the relational evolution graph convolutional neural network RE-GCN to embed knowledge into the knowledge graph, inputting the urban traffic knowledge graph into the relational evolution graph convolutional neural network RE-GCN to achieve knowledge embedding;

[0039] Step S2.3, using the long short-term memory network LSTM to predict traffic flow and capture the spatiotemporal relationship of traffic flow data;

[0040] Step S2.4, the traffic flow data is integrated with the information extracted from the urban traffic knowledge graph, and then input into the spatiotemporal multi-graph convolution module for training to extract spatiotemporal features;

[0041] Step S2.5: Output the traffic flow prediction value through the fully connected layer of the traffic flow prediction model, and evaluate the performance of the traffic flow prediction model on a real traffic flow dataset.

[0042] The intelligent optimization algorithm module uses the traffic light timing intelligent optimization algorithm based on the Grey Model model to intelligently adjust the traffic light control strategy. The implementation steps are as follows:

[0043] Step S3.1, collect historical traffic flow data and record them in units of 15 minutes; use the accumulation generation method to accumulate the original data once to form a new data sequence;

[0044] Step S3.2, establish the Grey Model (1,1) model, and fit the model parameters by the least squares method;

[0045] Step S3.3, comprehensively considering actual delay, effective green time and saturation factors, and establishing a fitness function;

[0046] Step S3.4, using a genetic algorithm to optimize the timing of traffic lights, randomly generating a number of traffic light timing schemes as an initial population, and evaluating the fitness of each scheme according to a set objective function;

[0047] Step S3.5: Generate a new generation of population through roulette wheel selection, crossover and mutation operations, and iterate until the custom-set termination condition is reached.

[0048] The Grey Model-based intelligent optimization algorithm for traffic light timing has a built-in adaptive learning mechanism that can dynamically adjust the learning rate and model parameters according to real-time traffic conditions to adapt to changes in traffic flow.

[0049] A device for optimizing a traffic network based on a large model, characterized in that it includes a memory and a processor; the memory is used to store a computer program, and the processor is used to implement the above-mentioned method steps when executing the computer program.

[0050] A readable storage medium, characterized in that: a computer program is stored on the readable storage medium, and the computer program implements the above method steps when executed by a processor.

[0051] The beneficial effects of the present invention are as follows: the method and device for optimizing the traffic network based on a large model realizes dynamic optimization of the traffic network by analyzing traffic data in real time, intelligently adjusting the traffic signal light control strategy, optimizing traffic flow distribution, and can significantly improve the efficiency and safety of urban traffic networks, highways and public transportation systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0053] Attached Figure 1 Schematic diagram of the traffic flow prediction method of the present invention.

[0054] Attached Figure 2 Schematic diagram of the intelligent traffic light control method of the present invention. DETAILED DESCRIPTION

[0055] In order to enable those skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0056] The method for optimizing a transportation network based on a large model comprises the following steps:

[0057] Step S1: Data collection and feature extraction

[0058] Collect traffic flow monitoring data, weather information, and customized special event information related to traffic, and pre-process the collected data, including missing value processing, normalization processing, and time window division of time series data;

[0059] The self-attention mechanism of the Transformer model framework is used to process sequence data. The pre-processed data is trained through a combination of unsupervised learning and reinforcement learning to predict future traffic conditions.

[0060] Step S2: Traffic flow prediction model based on LSTM and knowledge graph

[0061] Step S2.1, collect traffic flow data, including historical traffic flow, road checkpoints, points of interest (POIs), and meteorological data, and design static and dynamic urban traffic knowledge graphs based on road traffic information and external factors to capture the semantic relationship between road traffic information and external factors;

[0062] Step S2.2, input the urban traffic knowledge graph into the relational evolution graph convolutional neural network RE-GCN, and use the relational evolution graph convolutional neural network RE-GCN to embed knowledge into the knowledge graph;

[0063] Step S2.3, using the long short-term memory network LSTM to predict traffic flow and capture the spatiotemporal relationship of traffic flow data;

[0064] Step S2.4, the traffic flow data is integrated with the information extracted from the urban traffic knowledge graph, and then input into the spatiotemporal multi-graph convolution module for training to extract spatiotemporal features;

[0065] Step S2.5, output the traffic flow prediction value through the fully connected layer of the traffic flow prediction model, and evaluate the performance of the traffic flow prediction model on a real traffic flow dataset;

[0066] Step S3: According to the traffic flow prediction value, the traffic light timing intelligent optimization algorithm based on the Grey Model model is used to intelligently adjust the traffic light control strategy, optimize the traffic flow distribution, and realize the dynamic optimization of the traffic network.

[0067] In step S1, traffic data is collected from three data sources, namely surveillance cameras, sensors and GPS, to obtain a comprehensive view of the traffic network;

[0068] The missing values ​​in the data are processed by the mean filling method; the input data is normalized so that the feature values ​​of different dimensions have the same scale; the time series data is divided into appropriate time windows to provide contextual information;

[0069] Extract key features from the preprocessed data, including traffic flow, speed, and accident information.

[0070] In step S3, the intelligent optimization algorithm for traffic light timing based on the Grey Model is used to intelligently adjust the traffic light control strategy, and the implementation steps are as follows:

[0071] Step S3.1, collect historical traffic flow data and record them in units of 15 minutes; use the accumulation generation method to accumulate the original data once to form a new data sequence;

[0072] Step S3.2, establish the Grey Model (1,1) model, and fit the model parameters by the least squares method;

[0073] Step S3.3, comprehensively considering actual delay, effective green time and saturation factors, and establishing a fitness function;

[0074] Step S3.4, using a genetic algorithm to optimize the timing of traffic lights, randomly generating a number of traffic light timing schemes as an initial population, and evaluating the fitness of each scheme according to a set objective function;

[0075] Step S3.5: Generate a new generation of population through roulette wheel selection, crossover and mutation operations, and iterate until the custom-set termination condition is reached.

[0076] The Grey Model-based intelligent optimization algorithm for traffic light timing has a built-in adaptive learning mechanism that can dynamically adjust the learning rate and model parameters according to real-time traffic conditions to adapt to changes in traffic flow.

[0077] The traffic light timing intelligent optimization algorithm based on the Grey Model model can predict traffic flow at the millisecond level and adjust traffic control strategies in real time.

[0078] The device for optimizing the traffic network based on a large model includes a data collection and feature extraction module, a traffic flow prediction model and an intelligent optimization algorithm module;

[0079] The data collection and feature extraction module is responsible for collecting traffic flow monitoring data, weather information, and customized special event information related to traffic, and preprocessing the collected data, including missing value processing, normalization processing, and time window division of time series data;

[0080] The self-attention mechanism of the Transformer model framework is used to process sequence data. The pre-processed data is trained through a combination of unsupervised learning and reinforcement learning to predict future traffic conditions.

[0081] Traffic flow prediction model, responsible for traffic flow prediction based on LSTM and knowledge graph;

[0082] The intelligent optimization algorithm module is responsible for intelligently adjusting the traffic light control strategy according to the traffic flow prediction value, using the traffic light timing intelligent optimization algorithm based on the Grey Model model, optimizing traffic flow distribution, and realizing dynamic optimization of the transportation network.

[0083] The data collection and feature extraction module collects traffic data from three data sources: surveillance cameras, sensors, and GPS to obtain a comprehensive view of the traffic network;

[0084] The missing values ​​in the data are processed by the mean filling method; the input data is normalized so that the feature values ​​of different dimensions have the same scale; the time series data is divided into appropriate time windows to provide contextual information;

[0085] Extract key features from the preprocessed data, including traffic flow, speed, and accident information.

[0086] The traffic flow prediction model performs traffic flow prediction, and the implementation steps are as follows:

[0087] Step S2.1, collect traffic flow data, including historical traffic flow, road checkpoints, points of interest (POIs), and meteorological data, and design static and dynamic urban traffic knowledge graphs based on road traffic information and external factors to capture the semantic relationship between road traffic information and external factors;

[0088] Step S2.2, using the relational evolution graph convolutional neural network RE-GCN to embed knowledge into the knowledge graph, inputting the urban traffic knowledge graph into the relational evolution graph convolutional neural network RE-GCN to achieve knowledge embedding;

[0089] Step S2.3, using the long short-term memory network LSTM to predict traffic flow and capture the spatiotemporal relationship of traffic flow data;

[0090] Step S2.4, the traffic flow data is integrated with the information extracted from the urban traffic knowledge graph, and then input into the spatiotemporal multi-graph convolution module for training to extract spatiotemporal features;

[0091] Step S2.5: Output the traffic flow prediction value through the fully connected layer of the traffic flow prediction model, and evaluate the performance of the traffic flow prediction model on a real traffic flow dataset.

[0092] The intelligent optimization algorithm module uses the traffic light timing intelligent optimization algorithm based on the Grey Model model to intelligently adjust the traffic light control strategy. The implementation steps are as follows:

[0093] Step S3.1, collect historical traffic flow data and record them in units of 15 minutes; use the accumulation generation method to accumulate the original data once to form a new data sequence;

[0094] Step S3.2, establish the Grey Model (1,1) model, and fit the model parameters by the least squares method;

[0095] Step S3.3, comprehensively considering actual delay, effective green time and saturation factors, and establishing a fitness function;

[0096] Step S3.4, using a genetic algorithm to optimize the timing of traffic lights, randomly generating a number of traffic light timing schemes as an initial population, and evaluating the fitness of each scheme according to a set objective function;

[0097] Step S3.5: Generate a new generation of population through roulette wheel selection, crossover and mutation operations, and iterate until the custom-set termination condition is reached.

[0098] The Grey Model-based intelligent optimization algorithm for traffic light timing has a built-in adaptive learning mechanism that can dynamically adjust the learning rate and model parameters according to real-time traffic conditions to adapt to changes in traffic flow.

[0099] The device for optimizing a traffic network based on a large model comprises a memory and a processor; the memory is used to store a computer program, and the processor is used to implement the above method steps when executing the computer program.

[0100] The readable storage medium stores a computer program, and when the computer program is executed by a processor, the above method steps are implemented.

[0101] Compared with the prior art, the method and device for optimizing the traffic network based on a large model have the following characteristics:

[0102] (1) Ability to process and fuse data from different sources, such as video surveillance, sensors, GPS, etc., to obtain a comprehensive view of the transportation network and achieve multimodal data fusion.

[0103] (2) An integrated adaptive learning mechanism can dynamically adjust the learning rate and model parameters according to real-time traffic conditions to adapt to changes in traffic flow.

[0104] (3) Able to predict traffic flow at the millisecond level and adjust and optimize traffic control strategies in real time.

[0105] In terms of urban traffic networks, by analyzing traffic flow and predicting congestion points, and intelligently adjusting traffic light control strategies, it is possible to optimize traffic flow distribution and reduce congestion.

[0106] In terms of highway traffic, it is possible to improve driving safety and efficiency by predicting traffic flow and accident risks, adjusting speed limits and lane assignments in real time.

[0107] For public transportation systems, by analyzing passenger flow and demand, it is possible to optimize bus and subway scheduling plans and improve public transportation efficiency and passenger satisfaction.

[0108] The embodiment described above is only one specific implementation of the present invention. Common changes and substitutions made by those skilled in the art within the scope of the technical solution of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for optimizing a transportation network based on a large model, characterized in that: The following steps are involved: Step S1: Data collection and feature extraction Collect traffic flow monitoring data, weather information, and customized special event information related to traffic, and pre-process the collected data, including missing value processing, normalization processing, and time window division of time series data; The self-attention mechanism of the Transformer model framework is used to process sequence data. The pre-processed data is trained through a combination of unsupervised learning and reinforcement learning to predict future traffic conditions. Step S2: Traffic flow prediction model based on LSTM and knowledge graph Step S2.1, collect traffic flow data, including historical traffic flow, road checkpoints, points of interest (POIs), and meteorological data, and design static and dynamic urban traffic knowledge graphs based on road traffic information and external factors to capture the semantic relationship between road traffic information and external factors; Step S2.2, input the urban traffic knowledge graph into the relational evolution graph convolutional neural network RE-GCN, and use the relational evolution graph convolutional neural network RE-GCN to embed knowledge into the knowledge graph; Step S2.3, using the long short-term memory network LSTM to predict traffic flow and capture the spatiotemporal relationship of traffic flow data; Step S2.4, the traffic flow data is integrated with the information extracted from the urban traffic knowledge graph, and then input into the spatiotemporal multi-graph convolution module for training to extract spatiotemporal features; Step S2.5, output the traffic flow prediction value through the fully connected layer of the traffic flow prediction model, and evaluate the performance of the traffic flow prediction model on a real traffic flow dataset; Step S3: According to the traffic flow prediction value, the traffic light timing intelligent optimization algorithm based on the Grey Model model is used to intelligently adjust the traffic light control strategy, optimize the traffic flow distribution, and realize the dynamic optimization of the traffic network.

2. The method for optimizing a transportation network based on a large model according to claim 1, characterized in that: In step S1, traffic data is collected from three data sources: surveillance cameras, sensors and GPS, to obtain a comprehensive view of the traffic network; The missing values ​​in the data are processed by the mean filling method; the input data is normalized so that the feature values ​​of different dimensions have the same scale; the time series data is divided into appropriate time windows to provide contextual information; Extract key features from the preprocessed data, including traffic flow, speed, and accident information.

3. The method for optimizing a transportation network based on a large model according to claim 1, characterized in that: In step S3, the intelligent optimization algorithm for traffic light timing based on the Grey Model is used to intelligently adjust the traffic light control strategy, and the implementation steps are as follows: Step S3.1, collect historical traffic flow data and record them in units of 15 minutes; use the accumulation generation method to accumulate the original data once to form a new data sequence; Step S3.2, establish the Grey Model (1,1) model, and fit the model parameters by the least squares method; Step S3.3, comprehensively considering actual delay, effective green time and saturation factors, and establishing a fitness function; Step S3.4, using a genetic algorithm to optimize the timing of traffic lights, randomly generating a number of traffic light timing schemes as an initial population, and evaluating the fitness of each scheme according to a set objective function; Step S3.5: Generate a new generation of population through roulette wheel selection, crossover and mutation operations, and iterate until the custom-set termination condition is reached.

4. The method for optimizing a transportation network based on a large model according to claim 3 is characterized in that: The GreyModel-based intelligent optimization algorithm for traffic light timing has a built-in adaptive learning mechanism that can dynamically adjust the learning rate and model parameters according to real-time traffic conditions to adapt to changes in traffic flow.

5. A device for optimizing a traffic network based on a large model, characterized in that: Including data collection and feature extraction modules, Traffic flow prediction model and intelligent optimization algorithm module; The data collection and feature extraction module is responsible for collecting traffic flow monitoring data, weather information, and customized special event information related to traffic, and preprocessing the collected data, including missing value processing, normalization processing, and time window division of time series data; The self-attention mechanism of the Transformer model framework is used to process sequence data. The pre-processed data is trained through a combination of unsupervised learning and reinforcement learning to predict future traffic conditions. Traffic flow prediction model, responsible for traffic flow prediction based on LSTM and knowledge graph; The intelligent optimization algorithm module is responsible for intelligently adjusting the traffic light control strategy according to the traffic flow prediction value, using the traffic light timing intelligent optimization algorithm based on the Grey Model model, optimizing traffic flow distribution, and realizing dynamic optimization of the transportation network.

6. The device for optimizing a traffic network based on a large model according to claim 5, characterized in that: The data collection and feature extraction module collects traffic data from three data sources: surveillance cameras, sensors, and GPS to obtain a comprehensive view of the traffic network; The missing values ​​in the data are processed by the mean filling method; the input data is normalized so that the feature values ​​of different dimensions have the same scale; the time series data is divided into appropriate time windows to provide contextual information; Extract key features from the preprocessed data, including traffic flow, speed, and accident information.

7. The device for optimizing a traffic network based on a large model according to claim 5, characterized in that: The traffic flow prediction model performs traffic flow prediction, and the implementation steps are as follows: Step S2.1, collect traffic flow data, including historical traffic flow, road checkpoints, points of interest (POIs), and meteorological data, and design static and dynamic urban traffic knowledge graphs based on road traffic information and external factors to capture the semantic relationship between road traffic information and external factors; Step S2.2, using the relational evolution graph convolutional neural network RE-GCN to embed knowledge into the knowledge graph, inputting the urban traffic knowledge graph into the relational evolution graph convolutional neural network RE-GCN to achieve knowledge embedding; Step S2.3, using the long short-term memory network LSTM to predict traffic flow and capture the spatiotemporal relationship of traffic flow data; Step S2.4, the traffic flow data is integrated with the information extracted from the urban traffic knowledge graph, and then input into the spatiotemporal multi-graph convolution module for training to extract spatiotemporal features; Step S2.5: Output the traffic flow prediction value through the fully connected layer of the traffic flow prediction model, and evaluate the performance of the traffic flow prediction model on a real traffic flow dataset.

8. The device for optimizing a traffic network based on a large model according to claim 5, characterized in that: The intelligent optimization algorithm module uses the traffic light timing intelligent optimization algorithm based on the Grey Model model to intelligently adjust the traffic light control strategy. The implementation steps are as follows: Step S3.1, collect historical traffic flow data and record them in units of 15 minutes; use the accumulation generation method to accumulate the original data once to form a new data sequence; Step S3.2, establish the Grey Model (1,1) model, and fit the model parameters by the least squares method; Step S3.3, comprehensively considering actual delay, effective green time and saturation factors, and establishing a fitness function; Step S3.4, using a genetic algorithm to optimize the timing of traffic lights, randomly generating a number of traffic light timing schemes as an initial population, and evaluating the fitness of each scheme according to a set objective function; Step S3.5: Generate a new generation of population through roulette wheel selection, crossover and mutation operations, and iterate until the custom-set termination condition is reached.

9. The device for optimizing a traffic network based on a large model according to claim 6, characterized in that: The GreyModel-based intelligent optimization algorithm for traffic light timing has a built-in adaptive learning mechanism that can dynamically adjust the learning rate and model parameters according to real-time traffic conditions to adapt to changes in traffic flow.

10. A device for optimizing a transportation network based on a large model, characterized in that: The method comprises a memory and a processor; the memory is used to store a computer program, and the processor is used to implement the method according to any one of claims 1 to 4 when executing the computer program.

Images