Intelligent road network scheduling management method and system for public transportation

Abstract

The application discloses an intelligent road network scheduling management method and system for public transportation, relates to the technical field of traffic management, and comprises the following steps: collecting real-time traffic data to obtain real-time traffic data; simulating a road network of a target area based on digital twin technology to generate a road network model of the target area; obtaining a current vehicle scheduling scheme, combining real-time traffic data to perform simulation prediction, and generating a simulation prediction result; evaluating scheduling effect to generate a scheduling effect evaluation coefficient; generating an optimization instruction when the scheduling effect evaluation coefficient meets a preset scheduling effect evaluation coefficient; based on the optimization instruction, combining the road network model of the target area and the real-time traffic data, performing vehicle scheduling optimization, and generating an optimized vehicle scheduling scheme; and sending the optimized vehicle scheduling scheme to a vehicle scheduling unit to schedule public transportation vehicles in the target area. Thus, the technical effects of improving scheduling responsiveness, reducing scheduling cost, and optimizing management efficiency are achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of traffic management, in particular to an intelligent road network scheduling management method and system for public transportation. BACKGROUND

[0002] In urban traffic management, the scheduling management of public transportation is the key to improving traffic efficiency and service quality. The current public transportation scheduling management method mainly relies on fixed timetables and experience rules, which often cannot cope with complex traffic conditions and emergencies, and does not fully consider the cost factors of scheduling management. There are technical problems of untimely scheduling, lagging information feedback, and poor management effect. SUMMARY

[0003] The purpose of the present application is to provide an intelligent road network scheduling management method and system for public transportation. To solve the technical problems of untimely scheduling, lagging information feedback, and poor management effect in the prior art.

[0004] In view of the above technical problems, the present application provides an intelligent road network scheduling management method and system for public transportation.

[0005] In a first aspect, the present application provides an intelligent road network scheduling management method for public transportation, wherein the method comprises:

[0006] real-time traffic data of a target area is collected by a data collection system to obtain real-time traffic data;

[0007] A road network model of the target area is generated by simulating the target area based on digital twin technology;

[0008] A vehicle scheduling scheme of the current public transportation vehicle is obtained, and simulation prediction is performed in the target area road network model based on the real-time traffic data to generate a simulation prediction result;

[0009] The simulation prediction result is evaluated for scheduling effect to generate a scheduling effect evaluation coefficient;

[0010] When the scheduling effect evaluation coefficient meets a preset scheduling effect evaluation coefficient, an optimization instruction is generated;

[0011] Based on the optimization instruction, the target area road network model and the real-time traffic data are combined to perform vehicle scheduling optimization to generate an optimized vehicle scheduling scheme;

[0012] The optimized vehicle scheduling scheme is sent to a vehicle scheduling unit for scheduling of the public transportation vehicles in the target area.

[0013] In a second aspect, the present application further provides an intelligent road network scheduling management system for public transportation, wherein the system comprises:

[0014] a real-time collection module, configured to collect real-time traffic data of a target area by a data collection system, and obtain real-time traffic data;

[0015] a road surface simulation module, configured to simulate a road network of the target area based on a digital twin technology, and generate a target area road network model;

[0016] a current situation simulation and prediction module, configured to obtain a vehicle scheduling scheme of a current public transport vehicle, simulate and predict in the target area road network model in combination with the real-time traffic data, and generate a simulation and prediction result;

[0017] a scheduling evaluation module, configured to evaluate scheduling effect of the simulation and prediction result, and generate a scheduling effect evaluation coefficient;

[0018] an optimization triggering and instruction module, configured to generate an optimization instruction when the scheduling effect evaluation coefficient meets a preset scheduling effect evaluation coefficient;

[0019] a scheduling optimization output module, configured to perform vehicle scheduling optimization based on the optimization instruction in combination with the target area road network model and the real-time traffic data, and generate an optimized vehicle scheduling scheme;

[0020] a scheduling calling module, configured to send the optimized vehicle scheduling scheme to a vehicle scheduling unit, and perform public transport vehicle scheduling of the target area.

[0021] One or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0022] Real-time traffic data of a target area is collected by a data collection system, and real-time traffic data is obtained. A target area road network model is generated by simulating a road network of the target area based on a digital twin technology. A simulation and prediction result is generated by simulating and predicting in the target area road network model in combination with real-time traffic data. A scheduling effect evaluation coefficient is generated by evaluating scheduling effect of the simulation and prediction result. An optimization instruction is generated when the scheduling effect evaluation coefficient meets a preset scheduling effect evaluation coefficient. An optimized vehicle scheduling scheme is generated by performing vehicle scheduling optimization based on the optimization instruction in combination with the target area road network model and the real-time traffic data. The optimized vehicle scheduling scheme is sent to a vehicle scheduling unit, and public transport vehicle scheduling of the target area is performed. In this way, scheduling responsiveness is improved, scheduling cost is reduced, and management efficiency is optimized.

[0023] The above description is only a summary of the technical solutions of the present application. In order to more clearly illustrate the technical means of the present application, and then can be implemented according to the content of the specification, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the following specific embodiments of the present application are described. BRIEF DESCRIPTION OF DRAWINGS

[0024] The embodiments of the present application and the simple description described below are illustrated in combination with the drawings, and the drawings are described as follows:

[0025] Figure 1 The flowchart of the intelligent road network scheduling management method for public transportation of the present application;

[0026] Figure 2 The structure diagram of the intelligent road network scheduling management system for public transportation of the present application.

[0027] Explanation of reference signs: real-time acquisition module 11, road simulation module 12, current situation simulation and prediction module 13, scheduling evaluation module 14, optimization trigger and instruction module 15, scheduling optimization output module 16, scheduling calling module 17. DETAILED DESCRIPTION

[0028] The present application provides an intelligent road network scheduling management method and system for public transportation, which solves the technical problems of scheduling not in time, information feedback lag, and poor management effect in the prior art.

[0029] The overall idea adopted by the scheme in the technical embodiment of the present application to solve the above problems is as follows:

[0030] Firstly, the target area is subjected to real-time traffic data acquisition by a data acquisition system to obtain real-time traffic data. Then, the target area is subjected to road network simulation based on digital twin technology to generate a target area road network model. Then, a vehicle scheduling scheme of a current public transportation vehicle is obtained, and simulation and prediction are performed in the target area road network model in combination with the real-time traffic data to generate a simulation and prediction result. Further, scheduling effect evaluation is performed on the simulation and prediction result to generate a scheduling effect evaluation coefficient. When the scheduling effect evaluation coefficient meets a preset scheduling effect evaluation coefficient, an optimization instruction is generated. Based on the optimization instruction, the target area road network model and the real-time traffic data are combined to perform vehicle scheduling optimization to generate an optimized vehicle scheduling scheme. Finally, the optimized vehicle scheduling scheme is sent to a vehicle scheduling unit to realize scheduling of the public transportation vehicle in the target area. Further, the technical effect of improving scheduling responsiveness, reducing scheduling cost, and optimizing management efficiency is achieved.

[0031] For better understanding of the above technical solutions, the above technical solutions will be described in detail below in combination with the drawings and specific embodiments of the specification. It should be noted that the described embodiments are only part of the embodiments of the present application, not all embodiments of the present application, and it should be understood that the present application is not limited by the example embodiments described herein. Based on the embodiments of the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of protection of the present application. In addition, it should be noted that only the parts related to the present application are shown in the drawings for convenience of description, not all.

[0032] Embodiment one

[0033] As Figure 1 shown, the present application provides an intelligent road network scheduling management method for public transportation, which comprises:

[0034] S100: collecting real-time traffic data of the target area through a data collection system to obtain real-time traffic data;

[0035] Optionally, the data collection system refers to a collection system for acquiring real-time traffic data of the target area, including a data collection system independent of and connected to the target area traffic management system or a data collection system embedded in the target area traffic management platform.

[0036] Optionally, the data collection system embedded in the target area traffic management platform includes sensors, cameras, data transmission devices, data networks, etc.

[0037] Among them, the real-time traffic data includes traffic-related data collected in real time, including vehicle flow, vehicle speed, road conditions and other information. The real-time traffic data provides real and reliable basic data for subsequent traffic management, road planning and traffic flow analysis, etc.

[0038] Optionally, after obtaining the real-time traffic data, the real-time traffic data is preprocessed to improve data quality and accuracy, wherein the preprocessing method includes data cleaning, denoising, missing value processing, data conversion and normalization, etc.

[0039] Exemplarily, the real-time traffic data is preprocessed, including removing duplicate data, error data and outliers, ensuring the consistency and integrity of the data. Eliminate the noise in the data, make the data more smooth and reliable. Fill in the missing data, interpolation method, mean method, etc. can be used. Convert the data to make it suitable for subsequent analysis, such as converting the timestamp to date and time format. Standardize the data to make the data of different indicators comparable. Common methods include min-max normalization and Z-score normalization. Integrate data from multiple data sources to establish relationships between different types of data to obtain more comprehensive information. Based on principal component analysis and other technical methods, reduce the dimensionality of high-dimensional data to reduce data complexity and computational cost. Through the preprocessing steps, the quality and usability of real-time traffic data can be improved, providing a more reliable basis for subsequent data analysis and application.

[0040] S200: Simulate the road network of the target area based on digital twin technology to generate a target area road network model;

[0041] Optionally, the target area is simulated for road network, the purpose of which is to establish a digital twin simulation model of the target area road network, form a road network simulation environment of the target area, and then realize simulation and prediction in the digital space.

[0042] Exemplarily, a road network model of the target area is established by digital twin technology: first, collect geographic information data of the target area, including road network, traffic flow, road grade, traffic signals, etc. Then, use these data, combined with geographic information system (GIS) and other tools for data processing and modeling, to generate a digital road network model, which includes road network topology, traffic flow model, vehicle driving model, etc. Next, parameterize the road network model, convert different roads, traffic signals, vehicles, etc. into parameters in the digital model, configure the road network model, so that the road network model forms a model environment consistent with the target area, and thus obtains the target area road network model.

[0043] S300: Obtain the vehicle scheduling scheme of the current public transportation vehicle, combine the real-time traffic data, and simulate and predict in the target area road network model to generate a simulation and prediction result;

[0044] Optionally, the vehicle scheduling system of the interactive public transportation vehicle is used to obtain the current public transportation vehicle fleet scheduling scheme. The vehicle scheduling scheme of the public transportation vehicle includes vehicle departure time, route arrangement, etc.

[0045] Exemplarily, in a feasible implementation, firstly, real-time traffic data is combined into the target area road network model to initialize the target area road network model, thereby embodying the current traffic condition. Then, the vehicle scheduling scheme of the current public transport vehicle is taken as an input variable, and the road network model and the real-time traffic data are used for simulation prediction to predict the traffic condition in a future period of time, including congestion condition, vehicle flow, etc.

[0046] Optionally, before the simulation prediction in the target area road network model, a vehicle driving model of the target area is further configured, the vehicle driving model being used to describe the manner and rule of vehicle driving on the road network, and having an important influence on the accuracy and reliability of the simulation prediction result.

[0047] Exemplarily, configuring the vehicle driving model comprises: firstly, selecting a suitable vehicle driving model according to the actual situation and simulation demand of the target area. Commonly used vehicle driving models include micro-simulation models, macro-simulation models and hybrid simulation models, etc. Then, vehicle attributes and parameters are set, including vehicle type, vehicle speed, acceleration, braking distance, etc. The driving behavior of the vehicle in simulation is determined. Next, the driving rules of the vehicle are defined according to the actual traffic rules and road conditions, including the vehicle obeying traffic signals, vehicle lane changing rules, vehicle overtaking rules, etc. In addition, the vehicle driving mode in special situations, such as vehicle driving in emergency situations, road construction, etc. needs to be configured. By configuring the vehicle driving model of the target area, the driving condition of the vehicle on the road network can be more accurately simulated and predicted, thereby providing a scientific basis for public transport management and vehicle scheduling.

[0048] Optionally, the simulation prediction result is generated, including statistical analysis of the predicted road traffic efficiency, public transport vehicle traffic rate, public transport capacity, congestion duration and other road network operation data, storage of the above data, and formation of the simulation prediction result.

[0049] S400: scheduling effect evaluation is performed on the simulation prediction result to generate a scheduling effect evaluation coefficient;

[0050] Further, the simulation prediction result is subjected to scheduling effect evaluation to generate a scheduling effect evaluation coefficient, and step S400 comprises:

[0051] The simulation prediction result comprises passenger waiting time and vehicle late time;

[0052] The passenger waiting time and the vehicle late time are input into a time evaluation sub-network of a scheduling effect evaluation network to obtain a waiting time evaluation coefficient and a late time evaluation coefficient;

[0053] The waiting time evaluation coefficient and the late time evaluation coefficient are input into a coefficient connection sub-network of the scheduling effect evaluation network to obtain the scheduling effect evaluation coefficient.

[0054] Optionally, the simulation prediction result includes passenger waiting time and vehicle delay time, wherein the passenger waiting time can be divided into passenger average waiting time, passenger station average waiting time, passenger maximum waiting time, and 95% waiting time. The vehicle delay time can include vehicle average delay time, line vehicle average delay time, and interval vehicle average delay time.

[0055] Optionally, the scheduling effect evaluation network includes a time evaluation sub-network and a coefficient connection sub-network. The time evaluation sub-network is used to convert the passenger waiting time and the vehicle delay time into waiting time evaluation coefficients and delay time evaluation coefficients, so as to realize the evaluation of the scheduling effect in the waiting condition and the delay condition, and generate the waiting time evaluation coefficients and the delay time evaluation coefficients. The coefficient connection sub-network is used to comprehensively evaluate the waiting time evaluation coefficients and the delay time evaluation coefficients, and generate a scheduling effect evaluation coefficient. The scheduling effect evaluation coefficient provides a target parameter for subsequent optimization of the scheduling scheme.

[0056] Further, the passenger waiting time and the vehicle delay time are input into the time evaluation sub-network of the scheduling effect evaluation network, and the time evaluation sub-network is trained, and the training further includes:

[0057] Accessing a time evaluation record, wherein the time evaluation record includes a passenger waiting time evaluation record and a vehicle delay time evaluation record, the passenger waiting time evaluation record includes a sample passenger waiting time set and a sample waiting time evaluation coefficient set, and the vehicle delay time evaluation record includes a sample vehicle delay time set and a sample delay time evaluation coefficient set;

[0058] Based on the sample passenger waiting time set and the sample waiting time evaluation coefficient set, a first time evaluation channel is constructed;

[0059] Based on the sample vehicle delay time set and the sample delay time evaluation coefficient set, a second time evaluation channel is constructed;

[0060] The first time evaluation channel and the second time evaluation channel are integrated to obtain the time evaluation sub-network.

[0061] That is, by calling the time evaluation record, the training sample data of the first time evaluation channel and the second time evaluation channel are obtained. The time evaluation record includes a plurality of sample passenger waiting time data, sample waiting time evaluation coefficient data, sample vehicle late time data, and sample late time evaluation coefficient data of the target area. Through the time evaluation record, the first time evaluation channel and the second time evaluation channel are constructed and trained to form a time evaluation subnetwork, so that the time evaluation subnetwork has good adaptability to the target area, thereby improving the accuracy of the evaluation result and making it more practical, and then guiding the subsequent scheduling scheme optimization process.

[0062] Further, in some implementations, based on the sample passenger waiting time set and the sample waiting time evaluation coefficient set, the first time evaluation channel is constructed, and the method further includes:

[0063] Based on the neural network, a plurality of initial time evaluation channels are constructed;

[0064] The sample passenger waiting time set and the sample waiting time evaluation coefficient set are used to train the plurality of initial time evaluation channels to convergence, and a plurality of convergence times are obtained.

[0065] The initial time evaluation channel corresponding to the minimum value of the plurality of convergence times is obtained as the first time evaluation channel.

[0066] Optionally, a plurality of initial time evaluation channels are constructed simultaneously, each channel having different initial parameters, for example, including different network depth, network width, learning rate, initial weight, activation function, optimizer, and whether to configure cross-layer connection, etc.

[0067] Further, the passenger waiting time and the corresponding waiting time evaluation coefficient in the sample data set are used to train the plurality of initial time evaluation channels until the plurality of models converge to obtain a plurality of convergence times. Then, the time evaluation channel with the smallest convergence time is selected as the first time evaluation channel, thereby realizing the selection of the optimal initial time evaluation channel for subsequent scheduling effect evaluation.

[0068] Further, the waiting time evaluation coefficient and the late time evaluation coefficient are input into the coefficient connection subnetwork of the scheduling effect evaluation network to obtain the scheduling effect evaluation coefficient, and the method further includes:

[0069] The coefficient connection subnetwork includes a connection condition, wherein the connection condition includes the weight of the waiting time evaluation coefficient and the late time evaluation coefficient.

[0070] Based on the connection condition, the waiting time evaluation coefficient and the late time evaluation coefficient are weighted and summed to obtain the scheduling effect evaluation coefficient.

[0071] Optionally, based on the target area road network scheduling evaluation rule, the connection condition is set, specifically, including configuring corresponding weights for the waiting time evaluation coefficient and the late time evaluation coefficient, and the coefficient with a higher score in the target area road network scheduling evaluation rule has a larger weight.

[0072] S500: When the scheduling effect evaluation coefficient meets the preset scheduling effect evaluation coefficient, an optimization instruction is generated;

[0073] Optionally, the scheduling effect evaluation coefficient meets the preset scheduling effect evaluation coefficient, indicating that the scheduling effect of the current scheduling scheme is not good, and the scheduling scheme optimization needs to be performed. The preset scheduling effect evaluation coefficient is a threshold, reflecting the lower limit of the effect of the target area expected public transportation scheduling management.

[0074] S600: Based on the optimization instruction, the target area road network model and the real-time traffic data are combined to perform vehicle scheduling optimization, and an optimized vehicle scheduling scheme is generated;

[0075] The optimization instruction is used to activate the optimization of the vehicle scheduling scheme. The optimization instruction includes, for example, an optimization target, an optimization iteration number, an optimization termination condition, and an optimization output logic.

[0076] Optionally, the optimization target involves minimizing passenger waiting time, minimizing vehicle late time, maximizing public transportation coverage (ensuring that public transportation lines cover as many areas as possible to meet the travel needs of more passengers), minimizing public transportation line length (minimizing the length of public transportation lines to reduce vehicle travel distance and cost), and minimizing the number of transfers (minimizing the number of times passengers transfer between public transportation tools to improve the riding experience).

[0077] Further, in combination with the target area road network model and the real-time traffic data, vehicle scheduling optimization is performed to generate an optimized vehicle scheduling scheme, including:

[0078] Based on the vehicle scheduling scheme, a plurality of vehicle scheduling parameters are extracted, and a plurality of scheduling constraint spaces are generated;

[0079] Based on the plurality of scheduling constraint spaces, a first vehicle scheduling scheme is generated;

[0080] The target area road network model and the real-time traffic data are used to simulate and predict the first vehicle scheduling scheme, and a first simulation prediction result is generated;

[0081] The first simulation prediction result is evaluated for scheduling effect, and a first scheduling effect evaluation coefficient is generated;

[0082] When the first scheduling effect evaluation coefficient meets the preset scheduling effect evaluation coefficient, the first vehicle scheduling scheme is added to the candidate scheme set.

[0083] Cost prediction is performed on the candidate scheme set, and the optimized vehicle scheduling scheme is obtained according to the cost prediction result.

[0084] For example, first, a plurality of scheduling parameters are extracted from the vehicle scheduling scheme, and a plurality of scheduling constraint spaces are generated according to the parameters. For example, a vehicle interval constraint space determined based on the number of vehicles and the length of the line, a maximum speed and evaluation speed constraint space determined based on road safety, etc. Then, based on the plurality of scheduling constraint spaces, the current vehicle scheduling scheme is taken as a benchmark scheme, and a first vehicle scheduling scheme is generated by random variation. Then, the first vehicle scheduling scheme is simulated and predicted by using the target area road network model and real-time traffic data, and a first simulation prediction result is generated. Then, the first simulation prediction result is evaluated for scheduling effect, and a first scheduling effect evaluation coefficient is generated. If the first scheduling effect evaluation coefficient meets the preset scheduling effect evaluation coefficient, the first vehicle scheduling scheme is added to the candidate scheme set.

[0085] Optionally, the vehicle scheduling optimization also includes cost prediction on the candidate scheme set, and the optimized vehicle scheduling scheme is obtained according to the cost prediction result. For example, a cost prediction model is established based on historical data and relevant indicators, which can predict the cost of each candidate scheme. Then, for each scheme in the candidate scheme set, the cost prediction model is used to predict the cost, and then the optimized vehicle scheduling scheme with the lowest cost is selected according to the cost prediction result.

[0086] S700: The optimized vehicle scheduling scheme is sent to the vehicle scheduling unit for scheduling of public transport vehicles in the target area.

[0087] Optionally, after the optimized vehicle scheduling scheme is determined, the scheduling scheme is parsed and compiled to generate a corresponding machine language instruction set. The machine language instruction set includes a plurality of machine language instructions for controlling the public transport vehicles of the target area and the corresponding public transport indication devices and markers (bus stop boards, LED display screens, big data public transport vehicle information platforms, etc.). Each instruction in the plurality of machine language instructions has an address identifier to ensure transmission to the corresponding scheduling object.

[0088] In summary, the intelligent road network scheduling management method for public transport provided by the application has the following technical effects:

[0089] The real-time traffic data of the target area is collected through a data collection system to obtain real-time traffic data; a road network model of the target area is generated based on digital twin technology; a vehicle scheduling scheme of the current public transport vehicle is obtained, and simulation prediction is performed in the road network model of the target area in combination with the real-time traffic data to generate a simulation prediction result; scheduling effect evaluation is performed on the simulation prediction result to generate a scheduling effect evaluation coefficient; when the scheduling effect evaluation coefficient meets a preset scheduling effect evaluation coefficient, an optimization instruction is generated; based on the optimization instruction, vehicle scheduling optimization is performed in combination with the road network model of the target area and the real-time traffic data to generate an optimized vehicle scheduling scheme; and the optimized vehicle scheduling scheme is sent to a vehicle scheduling unit for scheduling of the public transport vehicle in the target area. Thus, the scheduling responsiveness is improved, the scheduling cost is reduced, and the management efficiency is optimized.

[0090] Embodiment Two

[0091] Based on the same concept as the intelligent road network scheduling management method for public transport in the embodiments, as shown in Figure 2 The application also provides an intelligent road network scheduling management system for public transport, which comprises:

[0092] A real-time collection module 11 is configured to collect real-time traffic data of a target area through a data collection system to obtain real-time traffic data.

[0093] A road surface simulation module 12 is configured to simulate a road network of the target area based on digital twin technology to generate a road network model of the target area.

[0094] A current situation simulation prediction module 13 is configured to obtain a vehicle scheduling scheme of a current public transport vehicle, perform simulation prediction in the road network model of the target area in combination with the real-time traffic data, and generate a simulation prediction result.

[0095] A scheduling evaluation module 14 is configured to perform scheduling effect evaluation on the simulation prediction result to generate a scheduling effect evaluation coefficient.

[0096] An optimization triggering and instruction module 15 is configured to generate an optimization instruction when the scheduling effect evaluation coefficient meets a preset scheduling effect evaluation coefficient.

[0097] A scheduling optimization output module 16 is configured to perform vehicle scheduling optimization in combination with the road network model of the target area and the real-time traffic data based on the optimization instruction to generate an optimized vehicle scheduling scheme.

[0098] A scheduling calling module 17 is configured to send the optimized vehicle scheduling scheme to a vehicle scheduling unit for scheduling of the public transport vehicle in the target area.

[0099] Further, the scheduling evaluation module 14 further includes:

[0100] The analysis calling unit is configured to call the simulation prediction result, which includes passenger waiting time and vehicle late time.

[0101] The multi-factor evaluation unit is configured to input the passenger waiting time and the vehicle late time into a time evaluation subnetwork of a scheduling effect evaluation network to obtain a waiting time evaluation coefficient and a late time evaluation coefficient.

[0102] The coefficient connection unit is configured to input the waiting time evaluation coefficient and the late time evaluation coefficient into a coefficient connection subnetwork of the scheduling effect evaluation network to obtain the scheduling effect evaluation coefficient.

[0103] Further, the multi-factor evaluation unit further includes:

[0104] The evaluation record unit is configured to call a time evaluation record, which includes a passenger waiting time evaluation record and a vehicle late time evaluation record, the passenger waiting time evaluation record including a sample passenger waiting time set and a sample waiting time evaluation coefficient set, and the vehicle late time evaluation record including a sample vehicle late time set and a sample late time evaluation coefficient set.

[0105] The first evaluation construction unit is configured to construct a first time evaluation channel based on the sample passenger waiting time set and the sample waiting time evaluation coefficient set.

[0106] The second evaluation construction unit is configured to construct a second time evaluation channel based on the sample vehicle late time set and the sample late time evaluation coefficient set.

[0107] The channel aggregation unit is configured to integrate the first time evaluation channel and the second time evaluation channel to obtain the time evaluation subnetwork.

[0108] Further, the coefficient connection unit further includes:

[0109] The connection configuration unit is configured to configure the coefficient connection subnetwork to include a connection condition, wherein the connection condition includes weights of the waiting time evaluation coefficient and the late time evaluation coefficient.

[0110] The evaluation calculation unit is configured to perform weighted summation on the waiting time evaluation coefficient and the late time evaluation coefficient based on the connection condition to obtain the scheduling effect evaluation coefficient.

[0111] Further, the first evaluation construction unit further includes:

[0112] The parallel construction unit is configured to construct a plurality of initial time evaluation channels based on a neural network.

[0113] The homologous training unit is configured to train the plurality of initial time evaluation channels to convergence by using the sample passenger waiting time set and the sample waiting time evaluation coefficient set, and obtain a plurality of convergence times.

[0114] The optimal selection unit is configured to obtain an initial time evaluation channel corresponding to a minimum value of the plurality of convergence times as the first time evaluation channel.

[0115] Further, the scheduling optimization output module 16 further comprises:

[0116] The scheduling constraint unit is configured to extract a plurality of vehicle scheduling parameters based on the vehicle scheduling scheme, and generate a plurality of scheduling constraint spaces.

[0117] The scheduling scheme generation unit is configured to generate a first vehicle scheduling scheme based on the plurality of scheduling constraint spaces.

[0118] The simulation prediction unit is configured to perform simulation prediction on the first vehicle scheduling scheme by using the target area road network model and the real-time traffic data, and generate a first simulation prediction result.

[0119] The effect evaluation unit is configured to perform scheduling effect evaluation on the first simulation prediction result, and generate a first scheduling effect evaluation coefficient.

[0120] The alternative storage unit is configured to add the first vehicle scheduling scheme to an alternative scheme set when the first scheduling effect evaluation coefficient meets the preset scheduling effect evaluation coefficient.

[0121] The cost screening unit is configured to perform cost prediction on the alternative scheme set, and screen the optimization vehicle scheduling scheme according to the cost prediction result.

[0122] It should be understood that the embodiments mentioned in the specification focus on their differences from other embodiments, and the specific embodiments in the foregoing embodiment one are also applicable to the intelligent road network scheduling management system for public transportation described in embodiment two. For the sake of brevity of the specification, no further expansion is made here.

[0123] It should be understood that the embodiments disclosed in the present application and the above description can enable those skilled in the art to implement the present application. Meanwhile, the present application is not limited to the aforementioned part of the embodiments. The embodiments mentioned in the present application are obvious modifications, combinations and substitutions, which also belong to the protection scope of the present application.

Claims

1. A method for intelligent road network dispatching management for public transportation, characterized in that, The method comprises the following steps: real-time traffic data of a target area is collected by a data collection system to obtain real-time traffic data; a road network model of the target area is generated by simulating the road network of the target area based on digital twinning technology; a vehicle scheduling scheme of a current public transport vehicle is obtained, and simulation prediction is performed in the road network model of the target area based on the real-time traffic data to generate a simulation prediction result; a scheduling effect evaluation coefficient is generated by evaluating the simulation prediction result; an optimization instruction is generated when the scheduling effect evaluation coefficient meets a preset scheduling effect evaluation coefficient; an optimized vehicle scheduling scheme is generated by performing vehicle scheduling optimization based on the optimization instruction, the road network model of the target area, and the real-time traffic data; the optimized vehicle scheduling scheme is sent to a vehicle scheduling unit to schedule the public transport vehicles in the target area; the simulation prediction result is evaluated to generate a scheduling effect evaluation coefficient, which comprises: the simulation prediction result includes passenger waiting time and vehicle delay time; the passenger waiting time and the vehicle delay time are input into a time evaluation subnetwork of a scheduling effect evaluation network to obtain a waiting time evaluation coefficient and a delay time evaluation coefficient; the waiting time evaluation coefficient and the delay time evaluation coefficient are input into a coefficient connection subnetwork of the scheduling effect evaluation network to obtain the scheduling effect evaluation coefficient; training the time evaluation subnetwork by inputting the passenger waiting time and the vehicle delay time into the time evaluation subnetwork of the scheduling effect evaluation network comprises: accessing a time evaluation record, wherein the time evaluation record includes a passenger waiting time evaluation record and a vehicle delay time evaluation record, the passenger waiting time evaluation record includes a sample passenger waiting time set and a sample waiting time evaluation coefficient set, and the vehicle delay time evaluation record includes a sample vehicle delay time set and a sample delay time evaluation coefficient set; constructing a first time evaluation channel based on the sample passenger waiting time set and the sample waiting time evaluation coefficient set; constructing a second time evaluation channel based on the sample vehicle delay time set and the sample delay time evaluation coefficient set; integrating the first time evaluation channel and the second time evaluation channel to obtain the time evaluation subnetwork.

2. The method of claim 1, wherein, constructing the first time evaluation channel based on the sample passenger waiting time set and the sample waiting time evaluation coefficient set comprises: constructing multiple initial time evaluation channels based on a neural network; training the multiple initial time evaluation channels to convergence using the sample passenger waiting time set and the sample waiting time evaluation coefficient set, and obtaining multiple convergence times; obtaining an initial time evaluation channel corresponding to the minimum value of the multiple convergence times as the first time evaluation channel.

3. The method of claim 2, wherein, inputting the waiting time evaluation coefficient and the delay time evaluation coefficient into the coefficient connection subnetwork of the scheduling effect evaluation network to obtain the scheduling effect evaluation coefficient comprises: the coefficient connection subnetwork includes a connection condition, wherein the connection condition includes the weights of the waiting time evaluation coefficient and the delay time evaluation coefficient. Based on the connection condition, the waiting time evaluation coefficient and the late time evaluation coefficient are weighted and summed to obtain the scheduling effect evaluation coefficient.

4. The method of claim 1, wherein, In combination with the target area road network model and the real-time traffic data, vehicle scheduling optimization is performed to generate an optimized vehicle scheduling scheme, including: Based on the vehicle scheduling scheme, a plurality of vehicle scheduling parameters are extracted, and a plurality of scheduling constraint spaces are generated. Based on the plurality of scheduling constraint spaces, a first vehicle scheduling scheme is generated. Using the target area road network model and the real-time traffic data, the first vehicle scheduling scheme is simulated and predicted to generate a first simulation prediction result. The first simulation prediction result is evaluated for scheduling effect to generate a first scheduling effect evaluation coefficient. When the first scheduling effect evaluation coefficient meets the preset scheduling effect evaluation coefficient, the first vehicle scheduling scheme is added to a candidate scheme set. The cost of the candidate scheme set is predicted, and the optimized vehicle scheduling scheme is obtained according to the cost prediction result.

5. The method of claim 1, wherein, Also includes: Evaluate the actual running effect of the public transportation vehicle scheduling at the preset time node; When the actual running effect does not meet the preset running effect, adjust the optimized vehicle scheduling scheme based on the actual running effect.

6. An intelligent road network dispatch management system for public transportation, characterized by, Includes: A real-time acquisition module is configured to acquire real-time traffic data of a target area through a data acquisition system; A road simulation module is configured to simulate a road network of the target area based on digital twinning technology to generate a target area road network model; A current simulation prediction module is configured to obtain a vehicle scheduling scheme of a current public transportation vehicle, and perform simulation prediction in the target area road network model in combination with the real-time traffic data to generate a simulation prediction result; A scheduling evaluation module is configured to evaluate the simulation prediction result for scheduling effect to generate a scheduling effect evaluation coefficient; The scheduling evaluation module is also configured to evaluate the simulation prediction result for scheduling effect to generate a scheduling effect evaluation coefficient, including: The simulation prediction result includes passenger waiting time and vehicle late time; The passenger waiting time and the vehicle late time are input into a time evaluation subnetwork of a scheduling effect evaluation network to obtain a waiting time evaluation coefficient and a late time evaluation coefficient; The waiting time evaluation coefficient and the late time evaluation coefficient are input into a coefficient connection subnetwork of the scheduling effect evaluation network to obtain the scheduling effect evaluation coefficient; The passenger waiting time and the vehicle late time are input into a time evaluation subnetwork of a scheduling effect evaluation network, and the time evaluation subnetwork is trained, including: Accessing a time evaluation record, wherein the time evaluation record includes a passenger waiting time evaluation record and a vehicle late time evaluation record, the passenger waiting time evaluation record includes a sample passenger waiting time set and a sample waiting time evaluation coefficient set, and the vehicle late time evaluation record includes a sample vehicle late time set and a sample late time evaluation coefficient set; construct a first time evaluation channel based on the set of sample passenger waiting time, the set of sample waiting time evaluation coefficients; construct a second time evaluation channel based on the set of sample vehicle late time, the set of sample late time evaluation coefficients; integrate the first time evaluation channel and the second time evaluation channel to obtain the time evaluation sub-network; an optimization trigger and instruction module, configured to generate an optimization instruction when the scheduling effect evaluation coefficient meets a preset scheduling effect evaluation coefficient; a scheduling optimization output module, configured to perform vehicle scheduling optimization based on the optimization instruction, the target area road network model and the real-time traffic data, and generate an optimized vehicle scheduling scheme; a scheduling calling module, configured to send the optimized vehicle scheduling scheme to a vehicle scheduling unit to perform public transportation vehicle scheduling in the target area.

Images