Method, system, device, and storage medium for public transportion scheduling

Abstract

The present disclosure provides a method, system, device and storage medium for the public transportation scheduling. The method includes acquiring comprehensive transportation information; determining a target transportation scheduling model from multiple transportation scheduling models, wherein different transportation scheduling models are used to predict public transportation scheduling information for different time durations; and inputting the comprehensive transportation information into the target transportation scheduling model to predict public transportation scheduling information for a corresponding time duration. The present disclosure improves flexibility and effectiveness of public transportation scheduling.

Description

Technical Field

[0001] This application relates to the field of Internet technology, specifically to a method, system, device, and storage medium for scheduling public transportation. Prior Technology

[0002] With the acceleration of urbanization and the increase in population density, urban public transportation systems are facing unprecedented challenges. The growing passenger demand and increasingly severe traffic congestion are forcing public transportation dispatch systems to transform towards more efficient and flexible service models. However, existing public transportation dispatch systems lack flexible forecasting mechanisms and often struggle to predict applicable dispatching schemes when faced with diverse traffic conditions, thus failing to effectively cope with various traffic situations. Summary of the Invention

[0003] This application provides a public transportation scheduling method, system, device, and storage medium. The public transportation scheduling method can accurately predict applicable scheduling schemes based on the actual needs of different traffic conditions, thereby improving the flexibility and effectiveness of public transportation scheduling.

[0004] In a first aspect, embodiments of this application provide a method for scheduling public transportation, the method comprising: Get comprehensive transportation information; A target traffic scheduling model is determined from multiple traffic scheduling models, and different traffic scheduling models are used to predict public transportation scheduling information for different time periods. The integrated traffic information is input into the target traffic scheduling model to predict public transportation scheduling information for the corresponding time period.

[0005] In some embodiments, the traffic scheduling model includes a short-term scheduling model, a medium-term scheduling model, and a long-term scheduling model. The short-term scheduling model is used to predict public transportation scheduling information for a first time period, the medium-term scheduling model is used to predict public transportation scheduling information for a second time period, and the long-term scheduling model is used to predict public transportation scheduling information for a third time period. The first time period is shorter than the second time period, and the second time period is shorter than the third time period. Determining the target traffic scheduling model from the multiple traffic scheduling models includes: Determine the time frame for the public transport scheduling information to be predicted; If the time length does not exceed the first time length, then the short-term scheduling model is determined to be the target traffic scheduling model; If the time length exceeds the first time length but does not exceed the second time length, then the intermediate scheduling model is determined to be the target traffic scheduling model. If the time length exceeds the second time length but does not exceed the third time length, then the long-term scheduling model is determined to be the target traffic scheduling model.

[0006] In some embodiments, if the target traffic scheduling model is determined to be the short-term scheduling model, then the integrated traffic information is input into the target traffic scheduling model to predict public transportation scheduling information for the corresponding time period, including: Obtain the first comprehensive traffic information corresponding to the fourth time length; The first integrated traffic information is input into the short-term scheduling model to predict the first public transportation scheduling information for the first time period.

[0007] In some embodiments, if the target traffic scheduling model is determined to be the medium-term scheduling model, then the integrated traffic information is input into the target traffic scheduling model to predict public transportation scheduling information for the corresponding time period, including: Obtain second comprehensive traffic information corresponding to the fifth time length, wherein the fifth time length is greater than the fourth time length; The second integrated traffic information is input into the medium-term scheduling model to predict the second public transportation scheduling information for the second time period; or The first public transportation scheduling information and the second integrated transportation information are input into the medium-term scheduling model to predict the second public transportation scheduling information for the second time period.

[0008] In some embodiments, if the target traffic scheduling model is determined to be the long-term scheduling model, then the integrated traffic information is input into the target traffic scheduling model to predict public transportation scheduling information for the corresponding time period, including: Obtain third comprehensive traffic information corresponding to the sixth time length, wherein the sixth time length is greater than the fifth time length; The third integrated transportation information is input into the long-term scheduling model to predict the third public transportation scheduling information for the third time period; or The second public transportation scheduling information and the third integrated transportation information are input into the long-term scheduling model to predict the third public transportation scheduling information for the third time period.

[0009] In some embodiments, the integrated traffic information includes real-time integrated traffic information and historical integrated traffic information, and the time length for determining the required predicted public transportation scheduling information includes: The real-time integrated traffic information and the historical integrated traffic information are analyzed to identify the fluctuation characteristics of the integrated traffic information. The fluctuation characteristics include at least one of the following: temporary fluctuations, periodic changes, and long-term trends. Based on the duration of the impact of the fluctuation characteristics on public transportation, the duration of the public transportation scheduling information to be predicted is determined.

[0010] In some embodiments, before determining the target traffic scheduling model from multiple traffic scheduling models, the method further includes: Collect comprehensive historical traffic information from multiple time periods; Machine learning is performed using historical comprehensive traffic information for each time period to obtain a traffic scheduling model corresponding to each time period.

[0011] In some embodiments, the step of performing machine learning on historical comprehensive traffic information for each time period to obtain a traffic scheduling model corresponding to each time period includes: The historical comprehensive traffic information for each time period is divided into training data set, validation data set, and test data set according to a preset ratio; Machine learning is performed using the training dataset to construct an initial traffic scheduling model corresponding to the time length. The initial traffic scheduling model is evaluated using the validation dataset, and its parameters are adjusted based on the evaluation results to optimize the model. The generalization performance of the optimized initial traffic scheduling model is evaluated using the test dataset. When the generalization performance of the optimized initial traffic scheduling model reaches the preset generalization capability requirement, the traffic scheduling model corresponding to the time length is obtained.

[0012] In some embodiments, inputting the integrated traffic information into the target traffic scheduling model to predict public transportation scheduling information for a corresponding time period includes: Relevant traffic information affecting public transportation is filtered from the comprehensive traffic information; The relevant traffic information is cleaned and standardized. The relevant traffic information after the cleaning and standardization processes are then subjected to dimensionality reduction processing. Feature extraction is performed on the dimensionality-reduced traffic information using at least one of the following methods: time series analysis, signal processing, and feature engineering, in order to extract derived traffic information. The derived traffic information is input into the target traffic scheduling model to predict public transportation scheduling information for the corresponding time period.

[0013] Secondly, embodiments of this application provide a public transportation dispatching system, the dispatching system comprising: Electronic devices for acquiring comprehensive traffic information, including Internet of Things (IoT) devices and mobile terminals; A server, connected to the IoT device and the mobile terminal, is used to execute the method described in any of the above-mentioned embodiments.

[0014] Thirdly, embodiments of this application provide a public transportation dispatching device, the device comprising: The acquisition module is used to obtain comprehensive traffic information. A determination module is used to identify a target traffic scheduling model from multiple traffic scheduling models, wherein different traffic scheduling models are used to predict public transportation scheduling information for different time periods. The prediction module is used to input the comprehensive traffic information into the target traffic scheduling model to predict public transportation scheduling information for a corresponding time period.

[0015] Fourthly, embodiments of this application provide a storage medium storing a computer program, which, when run on a computer, causes the computer to execute any of the methods described above.

[0016] This application's embodiments pre-define multiple traffic scheduling models. Different models can predict public transportation scheduling information for different time periods. Therefore, when facing different prediction needs, the most suitable model can be selected from among the multiple models as the target model. For example, when dealing with sudden temporary traffic congestion, a traffic scheduling model focused on predicting shorter time periods can be selected. Based on comprehensive traffic information, it can quickly output public transportation scheduling information for the next few minutes to several hours, providing an immediate and effective vehicle scheduling solution to quickly alleviate passenger flow pressure and solve the current emergency. As another example, when planning urban transportation from a longer-term perspective, a traffic scheduling model focused on predicting longer time periods can be selected. Based on comprehensive traffic information, it can predict public transportation scheduling information for the next few months to several years, which helps to better formulate longer-term vehicle scheduling and route adjustment strategies to meet the long-term optimization needs of public transportation. Therefore, this embodiment can flexibly select the appropriate target scheduling model from multiple traffic scheduling models when dealing with various traffic conditions and different forecasting needs, so as to predict public transportation scheduling information with strong targeting and high accuracy. Whether dealing with sudden immediate traffic problems or making long-term planning needs, it can be effectively solved, thereby greatly improving the flexibility and effectiveness of public transportation scheduling. Simple Explanation of the Diagram

[0017] To more clearly illustrate the technical solutions in the embodiments of this application, the illustrations used in the description of the embodiments will be briefly introduced below. Obviously, the illustrations described below are only some embodiments of this application. For those skilled in the art, other illustrations can be obtained from these illustrations without creative effort. Figure 1 is a flowchart illustrating a public transportation scheduling method provided in an embodiment of this application; Figure 2 is another flowchart illustrating the public transportation scheduling method provided in an embodiment of this application; Figure 3 is a flowchart illustrating the training process of the traffic scheduling model provided in the embodiments of this application; Figure 4 is another flowchart illustrating the training process of the traffic scheduling model provided in an embodiment of this application. Figure 5 is a schematic diagram of the process for extracting derived traffic features from comprehensive traffic information according to an embodiment of this application; Figure 6 is a schematic diagram of the public transportation dispatching system provided in an embodiment of this application; Figure 7 is a schematic diagram of the operation of the public transportation dispatching system provided in the embodiment of this application; Figure 8 is a schematic diagram of another scenario of the operation of the public transportation dispatching system provided in the embodiment of this application; Figure 9 is a schematic diagram of the structure of the public transportation dispatching device provided in the embodiment of this application; Figure 10 is another structural schematic diagram of the public transportation dispatching device provided in the embodiment of this application. Implementation

[0018] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the illustrations. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0019] Please refer to Figure 1, which is a flowchart illustrating a public transportation scheduling method according to an embodiment of this application. The specific flow of this public transportation scheduling method is as follows:

[0020] Get comprehensive transportation information on 101;

[0021] In 102, a target traffic scheduling model is determined from multiple traffic scheduling models, and different traffic scheduling models are used to predict public transportation scheduling information for different time periods.

[0022] In step 103, comprehensive traffic information is input into the target traffic scheduling model to predict public transportation scheduling information for the corresponding time period.

[0023] This embodiment can acquire comprehensive traffic information. For example, comprehensive traffic information can include multiple dimensions such as vehicle information, passenger information, environmental information, time information, and event information, providing comprehensive and rich data support for traffic scheduling models to predict public transportation scheduling information, thereby improving the accuracy of predictions. Vehicle information includes, for example, the vehicle's real-time location, speed, passenger status, whether the vehicle is in normal operation or malfunctioning, and the route the vehicle is traveling. Passenger information includes statistics on the number of passengers boarding and alighting at each station, passenger waiting time at each station, travel frequency, passenger origin station, destination station, expected travel time, and travel needs of special passengers such as the elderly, disabled, and pregnant women. Environmental information includes weather conditions, air quality, road conditions, road congestion, and road construction. Time information includes specific time points such as date, day of the week, and time period, as well as time characteristics such as whether it is a holiday, weekday, or weekend. Event information includes: arrangements for large-scale events such as concerts and sporting events, traffic control measures, traffic accidents, and natural disasters. This embodiment does not limit the specific content of the comprehensive traffic information; more multi-dimensional information that affects public transportation operation can be acquired as needed. It should be noted that comprehensive traffic information can include real-time comprehensive traffic information or comprehensive traffic information over a historical period.

[0024] This embodiment can acquire comprehensive traffic information in multiple ways. For example, IoT devices can be deployed on public transportation vehicles and at their stops. These IoT devices can capture rich data in real time, including the vehicle's real-time location, speed, passenger status, and passenger behavior data such as waiting times at various stops. The IoT devices can transmit the collected data to the server in a timely manner. Passengers can also conveniently input their personal travel needs, such as the originating station, destination, and expected travel time, through an application on their mobile terminal. This information is also uploaded to the server for processing and analysis via the application. The server can include a cloud server and / or a local server; this embodiment is not limited to this. Simultaneously, the server can integrate API interface technology to obtain real-time weather conditions, including environmental data such as temperature, rainfall, and wind speed, providing meteorological references for traffic scheduling. In terms of time, the precise time point of each data acquisition can be recorded, covering time characteristics such as date, day of the week, time period, and whether it is a holiday, providing a solid foundation for the time-series analysis of traffic information. Furthermore, information can be gathered directly from passengers or periodically from authoritative media and official websites to obtain information on important events such as large-scale event arrangements, traffic control measures, and government policies, all of which have a direct or indirect impact on traffic scheduling. This embodiment, through a series of information collection methods, can collect comprehensive multi-dimensional traffic information in real time.

[0025] This embodiment also pre-defines multiple traffic scheduling models. Different models can predict public transportation scheduling information for different time periods. Therefore, when facing different prediction needs, the most suitable model can be selected from among the multiple models as the target model. For example, when dealing with sudden temporary traffic congestion, a traffic scheduling model focused on predicting shorter time periods can be selected. Based on comprehensive traffic information, it can quickly output public transportation scheduling information for the next few minutes to several hours, providing an immediate and effective vehicle scheduling solution to quickly alleviate passenger flow pressure and solve the current emergency. As another example, when planning urban transportation from a longer-term perspective, a traffic scheduling model focused on predicting longer time periods can be selected. Based on comprehensive traffic information, it can predict public transportation scheduling information for the next few months to several years, which helps to better formulate long-term vehicle scheduling and route adjustment strategies to meet the long-term optimization needs of public transportation. Therefore, this embodiment not only improves the accuracy of prediction by comprehensively acquiring integrated traffic information, but also ensures that when dealing with various traffic conditions and different prediction needs, a suitable target scheduling model can be flexibly selected from multiple traffic scheduling models to predict targeted and highly accurate public transportation scheduling information. Whether dealing with sudden and immediate traffic problems or making long-term planning needs, it can be effectively solved, thereby greatly improving the flexibility and effectiveness of public transportation scheduling.

[0026] Please refer to Figure 2, which is another flowchart illustrating the public transportation scheduling method provided in this embodiment. The specific flow of this public transportation scheduling method can be as follows:

[0027] In section 201, you can obtain comprehensive transportation information.

[0028] This embodiment can acquire comprehensive traffic information, including multi-dimensional information such as vehicle information, passenger information, environmental information, time information, and event information. This provides comprehensive and rich data support for traffic scheduling models to predict public transportation scheduling information, thereby improving the accuracy of predictions. It should be noted that comprehensive traffic information can include real-time comprehensive traffic information or comprehensive traffic information over a historical period. The methods for acquiring comprehensive traffic information are as described above and will not be repeated here.

[0029] In step 202, determine the time frame for the public transport scheduling information that needs to be predicted.

[0030] In some embodiments, the traffic scheduling model may include a short-term scheduling model, a medium-term scheduling model, and a long-term scheduling model. The short-term scheduling model is used to predict public transportation scheduling information for a first time period, the medium-term scheduling model is used to predict public transportation scheduling information for a second time period, and the long-term scheduling model is used to predict public transportation scheduling information for a third time period. The first time period is shorter than the second time period, and the second time period is shorter than the third time period. The first, second, and third time periods can be adjusted as needed, and this embodiment does not impose limitations. For example, the short-term scheduling model can be used to predict public transportation scheduling information for the next few minutes to hours, the medium-term scheduling model can be used to predict public transportation scheduling information for the next few days to weeks, and the long-term scheduling model can be used to predict public transportation scheduling information for the next few months to years.

[0031] For example, the required timeframe for predicting public transport scheduling information can be determined based on the duration of the traffic conditions to be improved. For instance, when a large concert is held temporarily, potentially lasting several hours, it's necessary to immediately predict public transport scheduling information for the next few hours; in this case, the timeframe could be several hours. Similarly, when traffic or passenger flow changes during holidays or weekends, it's necessary to predict public transport scheduling information for those days; in this case, the timeframe could be one day to two weeks. Furthermore, when traffic or passenger flow experiences seasonal changes, it's necessary to predict public transport scheduling information for the following season; in this case, the timeframe could be several months.

[0032] In another example, the time frame for predicting public transportation scheduling information can be automatically determined based on integrated traffic information. This integrated traffic information includes real-time and historical integrated traffic information. Analysis of both data identifies fluctuation characteristics, including at least one of temporary fluctuations, periodic changes, or long-term trends. The time frame for predicting public transportation scheduling information is determined based on the duration of the impact of these fluctuation characteristics on public transportation.

[0033] By analyzing real-time and historical integrated traffic information, the fluctuation characteristics of integrated traffic information can be identified. If the fluctuation characteristic is identified as temporary, it may be due to short-term, non-periodic changes in traffic flow and / or passenger flow caused by sudden events such as traffic accidents, severe weather conditions (e.g., heavy rain), or large-scale events (e.g., concerts). These temporary fluctuations are sudden and short-lived, with an impact duration typically limited to the next few hours; therefore, the required public transportation scheduling timeframe can be considered to be several hours. If the fluctuation characteristic is identified as periodic, it refers to recurring changes in traffic flow and / or passenger flow within a specific time period. This periodicity reflects the regularity of people's daily travel habits, and the duration of its impact can be determined based on the length of the specific time period. For example, weekday morning and evening rush hours typically see a significant increase in traffic flow and / or passenger volume, with the impact lasting for several hours. Therefore, the required forecast duration for public transportation scheduling can be considered to be several hours. Weekend traffic flow and / or passenger volume may differ from weekdays, with the impact lasting for one to two days. Similarly, traffic flow and / or passenger volume during winter and summer vacations can also change compared to weekdays, with the impact lasting for one to two months. Therefore, the required forecast duration for public transportation scheduling can be considered to be one to two months. If the fluctuation characteristics are identified as long-term trends, reflecting the development trend or continuous direction of change in traffic flow and / or passenger volume, and with a longer duration of impact, the required forecast duration for public transportation scheduling can be considered to be the next few months to several years.

[0034] It should be noted that the traffic scheduling model in this embodiment is not limited to the division into short-term, medium-term, and long-term scheduling models. It can also be classified in other ways according to the time length. For example, the traffic scheduling model can be further subdivided into: real-time traffic scheduling model (responding quickly to real-time traffic changes), daily traffic scheduling model (considering traffic flow changes at different times of the day), weekly traffic scheduling model (planning based on the differences between weekdays and rest days within a week), monthly traffic scheduling model (considering seasonal changes or monthly event arrangements), or annual traffic scheduling model (making long-term plans for annual holidays, large-scale events, etc.).

[0035] In step 203, if the time length does not exceed the first time length, then the short-term scheduling model is determined as the target traffic scheduling model;

[0036] In step 204, retrieve the first comprehensive traffic information corresponding to the fourth time length;

[0037] In step 205, the first integrated transportation information is input into the short-term scheduling model to predict the first public transportation scheduling information for the first time period.

[0038] Once the required time frame for predicting public transportation scheduling information is determined, the time interval is further analyzed. If the time frame does not exceed the first time frame that the short-term scheduling model can predict, then the short-term scheduling model is selected as the target traffic scheduling model to respond to immediate traffic scheduling needs. The first time frame can be, for example, the next few minutes to several hours.

[0039] Obtain the first comprehensive traffic information corresponding to the fourth time length. The first comprehensive traffic information corresponding to the fourth time length can be the current real-time comprehensive traffic information, or the fourth time length can be equal to or close to the first time length. It can include the current real-time comprehensive traffic information and historical comprehensive traffic information close to the present, such as comprehensive traffic information from the last few minutes to several hours. Historical comprehensive traffic information can be extracted from a historical database.

[0040] The first comprehensive transportation information is input into a short-term scheduling model to predict the first public transportation scheduling information for the next two hours. Based on this information, traffic management departments can adjust public transportation schedules. For example, if the morning rush hour is from 7:00 AM to 9:00 AM, real-time vehicle locations, speeds, passenger loads, passenger numbers, and waiting times can be collected and input into the short-term scheduling model to predict public transportation scheduling for the next two hours. This allows for addressing challenges during the morning rush hour, such as increasing the number of vehicles and adjusting intervals at bus or subway stations with high passenger flow to shorten waiting times. It should be noted that traffic management departments can coordinate the scheduling of multiple public transportation modes, including buses, subways, taxis, carpooling, and ferries.

[0041] In step 206, if the time length exceeds the first time length but does not exceed the second time length, then the intermediate scheduling model is determined as the target traffic scheduling model;

[0042] In step 207, retrieve the second comprehensive traffic information corresponding to the fifth time length;

[0043] In step 208, the second integrated transportation information is input into the mid-term dispatch model to predict the second public transportation dispatch information for the second time period; or the first public transportation dispatch information and the second integrated transportation information are input into the mid-term dispatch model to predict the second public transportation dispatch information for the second time period.

[0044] When the time length of the public transportation scheduling information to be predicted exceeds the first time length but does not exceed the second time length, the medium-term scheduling model is determined as the target traffic scheduling model. For example, if the time to be predicted is from a few days to a few weeks, then the medium-term scheduling model is the most suitable choice to ensure the accuracy and practicality of the prediction.

[0045] Obtain the second comprehensive traffic information corresponding to the fifth time length, wherein the fifth time length is longer than the fourth time length and can be equal to or close to the second time length, such as including comprehensive traffic information from the past few days to the past few weeks, to provide sufficient information to support the medium-term scheduling model in predicting public transportation scheduling information for the next second time length.

[0046] For example, the second integrated transportation information can be directly input into the medium-term scheduling model to predict the second public transportation scheduling information for the next second time period.

[0047] In another example, to improve prediction accuracy, the prediction results from the short-term scheduling model—the first public transport scheduling information—can be input together with the second integrated traffic information into the medium-term scheduling model to predict the second public transport scheduling information for a second time period. This allows for a comprehensive consideration of short-term and medium-term traffic changes, as well as their interactions, resulting in more accurate and reliable predictions.

[0048] For example, if a large-scale sporting event is expected to be held within the next two weeks, attracting a large number of spectators, and a mid-term scheduling model is chosen to predict public transportation scheduling information for the next two weeks, then comprehensive traffic information from the past few days to the past few weeks needs to be collected. This information includes vehicle and passenger data such as vehicle location, speed, passenger status, number of passengers boarding and alighting, and passenger waiting time, as well as environmental information such as weather conditions, road conditions, and traffic control measures. In addition, event information such as the specific time and location of the sporting event and the expected number of spectators also needs to be collected. After collecting sufficient comprehensive traffic information, this second comprehensive traffic information is input into the mid-term scheduling model. Alternatively, the prediction results of the short-term scheduling model (i.e., public transportation scheduling information for the next few days) can be input together with the second comprehensive traffic information into the mid-term scheduling model to predict public transportation scheduling information for different time periods and areas within the next two weeks. Based on the public transportation scheduling information, vehicle deployment and route layout can be adjusted to ensure that public transportation services during the sporting event can meet the needs of spectators and participants.

[0049] In step 209, if the time length exceeds the second time length but does not exceed the third time length, then the long-term scheduling model is determined as the target traffic scheduling model;

[0050] In 210, retrieve the third comprehensive traffic information corresponding to the sixth time length;

[0051] In 211, the third integrated transportation information is input into the long-term scheduling model to predict the third public transportation scheduling information for the third time period; or the second public transportation scheduling information and the third integrated transportation information are input into the long-term scheduling model to predict the third public transportation scheduling information for the third time period.

[0052] When the time length for the public transportation scheduling information to be predicted exceeds the second time length but does not exceed the third time length, the long-term scheduling model is determined as the target traffic scheduling model. For example, if the time to be predicted is several months to several years, then the long-term scheduling model is the most suitable choice to ensure the accuracy and practicality of the prediction.

[0053] Obtain the third comprehensive transportation information corresponding to the sixth time length, where the sixth time length is longer than the fifth time length. The sixth time length can be equal to or close to the third time length, such as including comprehensive transportation information from the past few months to several years, to provide sufficient information to support the long-term scheduling model in predicting the third public transportation scheduling information for the next third time length.

[0054] For example, third-party integrated transportation information can be directly input into a long-term scheduling model to predict third-party public transportation scheduling information within a third time period.

[0055] In another example, to improve the accuracy of predictions, the prediction results of the medium-term scheduling model, i.e., the second public transportation scheduling information, can be input into the long-term scheduling model together with the third integrated transportation information so that the long-term scheduling model can make accurate predictions.

[0056] For example, the construction of a new shopping mall will impact traffic flow around it for several years. This timeframe exceeds the prediction range of a medium-term scheduling model (days to weeks) but does not exceed the prediction range of a long-term scheduling model (months to years). Therefore, a long-term scheduling model is chosen to predict and plan future public transportation services. Collecting comprehensive traffic information over a six-year timeframe (e.g., the past few years) is crucial. This information includes not only vehicle location, speed, passenger status, number of passengers boarding and alighting, and passenger waiting times, but also event information related to the new shopping mall's construction, such as construction time, location, expected size, and business distribution. This rich comprehensive traffic information provides a solid foundation for the long-term scheduling model, helping to predict public transportation scheduling in the years following the mall's completion. For instance, it allows for planning new public transportation routes, adjusting existing public transportation routes, increasing the frequency of existing routes to cope with increased passenger flow, or adjusting public transportation operating hours to match the mall's operating hours, thus meeting long-term transportation needs and providing convenient public transportation services.

[0057] It should be noted that there is a correlation between the first public transportation scheduling information generated by the short-term scheduling model, the second public transportation scheduling information generated by the medium-term scheduling model, and the third public transportation scheduling information generated by the long-term scheduling model. This is achieved through a progressive approach: using the results predicted by the short-term scheduling model as input features for the medium-term scheduling model, and then using the results predicted by the medium-term scheduling model as input features for the long-term scheduling model. The short-term scheduling model typically focuses on current and recent changes. Using this information as input features for the medium-term scheduling model demonstrates that the medium-term scheduling model better understands and predicts traffic trends over the next few days to weeks. Similarly, the results predicted by the medium-term scheduling model can further provide valuable information for long-term forecasting, thereby constructing a more coherent and accurate forecasting system.

[0058] In some embodiments, different traffic scheduling models are selected to predict effective public transportation scheduling information based on the characteristics of different traffic conditions. Therefore, before determining the target model from multiple traffic scheduling models based on real-time comprehensive traffic information, these traffic scheduling models need to be trained in advance. For example, please refer to Figure 3, which is a flowchart illustrating the traffic scheduling model training process provided in this embodiment. The specific process can be as follows:

[0059] In 301, historical comprehensive traffic information of multiple time periods was collected.

[0060] To train a model capable of predicting public transportation scheduling information over different time periods, it is necessary to collect historical comprehensive traffic information over multiple time periods. This information not only covers basic vehicle information (such as vehicle location, speed, and operating status) and passenger information (such as the number of passengers getting on and off the bus and passenger density), but also delves into environmental information (such as weather conditions, road conditions, and traffic control measures), time information (such as date, time period, and holidays), and various event information (such as special events, traffic accidents, and road construction).

[0061] Given that different traffic scheduling models are tasked with predicting traffic scheduling over different time spans—for example, short-term, medium-term, and long-term models—short-term models predict public transportation scheduling information for the next few minutes to hours, medium-term models for the next few days to weeks, and long-term models for the next few months to years, historical comprehensive traffic information matching the target prediction time span is selected as training data for the machine learning model during the training phase to ensure that the traffic scheduling model can accurately predict public transportation scheduling information for each time length.

[0062] It should be noted that the short-term, medium-term, and long-term scheduling models focus on different timeframes and prediction objectives, and may emphasize different types of comprehensive traffic information. Therefore, during training or prediction, the same or different types of information from the comprehensive traffic data can be used for training or prediction, depending on the actual situation. This embodiment does not impose any limitations on this. For example, the short-term scheduling model may focus more on specific traffic events, such as temporary traffic control or vehicle breakdowns; the medium-term scheduling model may focus more on information such as holidays and large-scale events; and the long-term scheduling model may focus more on government-implemented traffic policies, such as the impact of shared bicycles and other modes of transportation on public transportation.

[0063] In section 302, machine learning is performed on the historical comprehensive traffic information for each time period to obtain the traffic scheduling model corresponding to each time period.

[0064] Please refer to Figure 4, which is another flowchart illustrating the traffic scheduling model training process provided in this embodiment. Step 302 may include:

[0065] In 3021, historical comprehensive traffic information for each time period is divided into training data set, verification data set and test data set according to a preset ratio.

[0066] To ensure that the trained traffic scheduling model performs well on unseen data, when training a traffic scheduling model, historical comprehensive traffic information for a corresponding time period can be divided into a training data set, a validation data set, and a test data set according to a preset ratio. This embodiment does not limit the preset ratio; for example, it can be 8:1:1. This division can be achieved through random sampling or time series partitioning. The training data set is used to train the model, the validation data set is used for model tuning, and the test data set is used to finally evaluate the model's generalization ability.

[0067] In 3022, machine learning is used with the training dataset to build an initial traffic scheduling model corresponding to the time length.

[0068] After preparing the training dataset, a suitable machine learning model needs to be selected to learn from the dataset and build an initial traffic scheduling model. Machine learning models can include linear regression, decision trees, random forests, support vector machines, and neural networks. The same machine learning model can be used to train different traffic scheduling models, or different models can be selected based on their characteristics and advantages to achieve the best predictive performance. For example, neural networks or random forests are good at handling nonlinear relationships and complex patterns in time series data, making them suitable for predicting traffic flow in the next few hours, and therefore suitable for training short-term traffic scheduling models. Support vector machines (SVM) or random forests perform well in handling medium-term trends and periodic changes, making them suitable for predicting traffic demand in the next few days to weeks, and therefore suitable for training medium-term traffic scheduling models. Linear regression or time series models (such as ARIMA) are more stable and accurate in predicting long-term trends and seasonal changes, making them suitable for predicting traffic demand in the next few months to years, and therefore suitable for training long-term traffic scheduling models.

[0069] After selecting a machine learning model, model building can be performed before training to provide the necessary foundation and framework for model training. Model building can include feature selection, model initialization, loss function setting, and optimization algorithm selection. Feature selection determines which features will be used to train the model, helping the machine learning model capture the most important information in the data. Model initialization sets the initial parameters of the machine learning model, which will be adjusted during training. These parameters can include weights, biases, etc. Loss function setting defines a loss function to measure the accuracy of the model's predictions. The goal of training is to minimize the loss. Loss functions can include mean squared error (MSE) and cross-entropy. Optimization algorithm selection chooses an algorithm to update the model parameters to reduce the value of the loss function. Optimization algorithms can include gradient descent, stochastic gradient descent, Adam, etc.

[0070] After the model is built, the machine learning model is trained using the training dataset to minimize the loss function. The training process includes forward propagation, loss calculation, backpropagation, and iterative training. Forward propagation involves inputting the training dataset and calculating the predicted values ​​using the model; loss calculation involves calculating the error between the predicted and true values ​​based on the loss function; backpropagation involves calculating the gradient using the chain rule based on the error and updating the model parameters; iterative training involves repeating the forward and backpropagation processes until the loss function converges or a predetermined number of training rounds are reached.

[0071] In section 3023, the performance of the initial traffic scheduling model is evaluated using a validation dataset, and the parameters of the initial traffic scheduling model are adjusted based on the evaluation results to optimize the initial traffic scheduling model.

[0072] The validation set is a separate dataset from the training set, used to monitor model performance during training and to guide parameter adjustments. The trained model is then applied to the validation set, and the difference between the model's predictions and the actual results is calculated. Common evaluation metrics include root mean square error (RMSE), mean absolute error (MAE), and coefficient of determination (R²). The model's performance is analyzed based on the evaluation metrics. If the model performs poorly on the validation set, parameter adjustments may be necessary.

[0073] Furthermore, validation datasets can be used to evaluate model performance under different hyperparameter settings. By comparing the performance of different hyperparameter combinations on the validation dataset, the optimal hyperparameter combination can be selected, thereby improving the model's accuracy and generalization ability. Hyperparameters can include, for example, the learning rate, regularization parameter, and model complexity. Hyperparameter tuning methods can include grid search, random search, etc. In addition, cross-validation can be used to accurately evaluate the model's generalization ability, indicating whether the model is overfitting or underfitting. The three steps of validation dataset evaluation, hyperparameter tuning, and cross-validation work together to ensure high model performance and good generalization ability.

[0074] In 3024, the generalization performance of the optimized initial traffic scheduling model is evaluated using the test dataset. When the generalization performance of the optimized initial traffic scheduling model reaches the preset generalization capability requirement, the traffic scheduling model corresponding to the time length is obtained.

[0075] The test dataset is a dataset independent of the training and validation datasets. Analogous to the data the trained model will encounter in real-world applications, the purpose of the test dataset evaluation is to verify the model's performance on completely unseen data, i.e., the model's generalization ability. A model with good generalization ability can make accurate predictions on unseen data. The test dataset can be used to re-evaluate the optimized model. Commonly used evaluation metrics include accuracy, recall, F1 score, root mean square error (RMSE), and mean absolute error (MAE).

[0076] The optimized model is applied to a test dataset, and its performance on this dataset is analyzed to determine whether it meets the preset generalization performance requirements. If the model's generalization performance meets the preset requirements, it is considered effective and can be used for actual traffic scheduling; otherwise, it may be necessary to return to the model training and tuning phase for further optimization. The evaluation results from the test dataset provide a basis for the final deployment of the model, ensuring its practicality and effectiveness.

[0077] By evaluating the test dataset, when the generalization performance of the model is confirmed to meet the preset requirements, the trained model can be deployed to the actual public transportation scheduling system. When the generalization performance of multiple models meets the preset requirements, the best model can be selected from them based on the model's performance indicators and deployed as the traffic scheduling model. Public transportation scheduling information prediction and decision support can be realized through the API interface.

[0078] In some embodiments, when collecting raw integrated traffic information, before using it to train machine learning models or predict public transportation scheduling information, the raw integrated traffic information can be preprocessed to extract useful derived traffic features. This provides more useful input to the traffic scheduling model and effectively supports its training and prediction. Please refer to Figure 5, which is a flowchart illustrating the process of extracting derived traffic features from integrated traffic information according to an embodiment of this application. The specific process is as follows:

[0079] In section 401, relevant traffic information affecting public transportation is filtered from comprehensive traffic information.

[0080] In section 402, relevant traffic information is cleaned and standardized.

[0081] In section 403, the relevant traffic information after cleaning and standardization is subjected to dimensionality reduction processing;

[0082] In the 404 error, at least one of the following methods—time series analysis, signal processing, and feature engineering—is used to extract features from the dimensionality-reduced traffic information to obtain derived traffic information.

[0083] In step 405, derived traffic information is input into the target traffic scheduling model to predict public transport scheduling information for the corresponding time period.

[0084] By statistically analyzing a large amount of comprehensive traffic information, relevant traffic information that influences public transportation, such as passenger and vehicle traffic, can be filtered out. This relevant traffic information has a high correlation with public transportation. Statistical analysis methods include correlation analysis and analysis of variance, while filtering methods include Pearson correlation coefficient and chi-square test. Relevant traffic information can include vehicle operation data (such as vehicle location, speed, route, and arrival time), passenger boarding and alighting records, weather conditions (such as temperature, rainfall, and wind speed, which can affect passenger travel intentions), event information (such as holidays, large-scale events, and traffic control, which can significantly impact public transportation passenger flow), and time information (such as date, day of the week, and time of day, used to analyze the periodic patterns of passenger travel). This filtering process ensures that the data processed subsequently consists of key information that has a direct or indirect impact on public transportation scheduling.

[0085] Since the selected traffic information may contain missing values, outliers, duplicates, and variables of varying magnitudes, the data needs to be cleaned and standardized. For example, outliers can be identified by analyzing characteristics such as mean, variance, skewness, and kurtosis. The cleaning process mainly removes missing, outliers, and duplicates. For missing values, interpolation, imputation, or deletion methods can be used. For outliers, statistical methods (such as Z-scores and interquartile ranges) can be used to identify and process them. For duplicates, duplicate values ​​can be removed. Standardization aims to eliminate differences in magnitude between variables, standardizing or normalizing the data to make it comparable. Standardization, for example, converts the data to a standard normal distribution with a mean of 0 and a standard deviation of 1. Normalization scales the data to a preset range, such as [0,1].

[0086] Although the data has been filtered and cleaned to be relatively concise, dimensionality reduction is still necessary to further improve the model's training efficiency and prediction accuracy. This step can be achieved through dimensionality reduction techniques such as Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA). Dimensionality reduction can extract the feature combinations with the most information from the original features, thereby reducing feature dimensionality and model complexity.

[0087] For traffic information after dimensionality reduction, various methods can be used for feature extraction to extract useful derivative traffic information. For example, time series analysis, signal processing techniques, and feature engineering can be used to further extract and construct derivative traffic information with predictive value for public transportation scheduling. Comprehensive traffic information is usually time series data; therefore, time series analysis can reveal the trends and periodic patterns of passenger demand changes over time. For instance, methods such as moving averages, differencing, and Fourier transforms can be used to extract the periodic fluctuations in bus location, speed, and passenger numbers. Signal processing techniques, such as wavelet transforms and Fourier transforms, can extract the frequency domain features of the data, revealing derivative traffic features hidden in passenger behavior patterns. Feature engineering enhances the model's expressiveness by constructing new features, such as calculating passenger density and traffic flow change rates based on passenger boarding and alighting records.

[0088] The derived traffic information obtained through feature extraction, such as the trend characteristics, periodic fluctuations, and frequency domain characteristics of passenger demand patterns, as well as newly constructed features like passenger density and traffic flow change rates, will be input into the target traffic scheduling model. This information will explain the model's predictions of public transportation scheduling information for a specific future time period, such as passenger flow, vehicle scheduling demand, and route optimization suggestions. The combined application of these methods can significantly improve the accuracy and comprehensiveness of feature extraction.

[0089] Feature extraction can also be performed in various other ways. For example, machine learning algorithms and data acquisition techniques can be used, such as cluster analysis and frequent pattern mining, to extract features and identify passenger demand patterns. For instance, passenger travel demand patterns at different times and on different websites can be identified. Then, based on the identified travel demand patterns, derivative traffic information can be extracted, such as travel frequency during peak hours and the number of passengers on specific routes. The extracted derivative traffic information is then input into the target traffic scheduling model to demonstrate that the target traffic scheduling model predicts public transportation scheduling information and accurately predicts passenger flow in the future.

[0090] For example, association rule mining techniques, such as the Apriori algorithm and the FP-Growth algorithm, can be used to extract features to discover whether there are any regular relationships in passengers' travel behavior under different conditions, and optimize public transportation services accordingly.

[0091] Association rule mining techniques are used to discover interesting relationships between variables in large datasets, especially frequent co-occurrences (frequent itemsets) and causal relationships between variables. For example, the Apriori algorithm is one of the most classic association rule mining algorithms. Its core idea is to utilize the anti-monotonicity of frequent itemsets, meaning that if an itemset is frequent, then all its subsets are also frequent. Specific steps include generating candidate itemsets, calculating support, and generating new candidate itemsets. Generating candidate itemsets involves creating them from the dataset, initially consisting of all individual itemsets. Calculating support involves calculating the support of each candidate itemset, i.e., the frequency of its occurrence in the dataset; itemsets with support higher than a set threshold are retained as frequent itemsets. Generating new candidate itemsets involves creating new candidate itemsets based on frequent itemsets, merging two frequent itemsets to generate candidate itemsets containing more items, and continuously repeating the process of generating candidate itemsets and calculating support until no new frequent itemsets can be generated. The Apriori algorithm can be used to uncover passenger travel patterns within specific time periods, such as the rule that "a certain website has high demand during the morning rush hour".

[0092] For example, the FP-Growth algorithm is an improvement on the Apriori algorithm, aiming to improve the efficiency of association rule mining. The FP-Growth algorithm represents the dataset by constructing a frequent pattern tree (FP-Tree) and then mining frequent itemsets from the FP-Tree. The specific steps include constructing the FP-Tree and mining frequent itemsets. Constructing the FP-Tree involves scanning the dataset twice: the first scan calculates the support of each item, and the second scan inserts the item into the FP-Tree based on the support. Mining frequent itemsets involves performing pattern growth on the FP-Tree, recursively mining frequent itemsets from the tree. The advantage of the FP-Growth algorithm is that it reduces the number of candidate itemset generation and support calculations, thus improving mining efficiency. Through the FP-Growth algorithm, more complex travel demand patterns can be discovered, such as the coordinated demand across multiple websites within a specific time period.

[0093] The use of association rule mining technology can uncover the association rules between passenger travel behavior and other factors. Based on these association rules, passenger needs can be better met. Inputting these identified association rules into a target traffic scheduling model allows the model to predict public transportation scheduling information, enabling intelligent management of the public transportation scheduling system, improving operational efficiency, enhancing passenger satisfaction, and contributing to the sustainable development of urban transportation. For example, by identifying peak hours and high-demand stations, adjusting departure frequencies and optimizing route configurations can reduce passenger waiting time during peak hours and at these stations, improving travel efficiency. Furthermore, based on identified passenger travel preferences, route configurations can be optimized by adjusting or adding direct routes, reducing unnecessary stops, minimizing transfers and travel time, improving the travel experience, and increasing route efficiency. Additionally, association rules can be used to predict future demand changes and flexibly adjust public transportation capacity according to actual demand, ensuring that capacity matches passenger needs, avoiding empty or overloaded situations, and improving resource utilization. Finally, by rationally and dynamically allocating public transportation resources, operating costs can be reduced, and economic efficiency improved.

[0094] This embodiment, through in-depth understanding of the current state of public transportation, identifies existing problems and improvement opportunities, such as vehicle punctuality, passenger carrying efficiency, and passenger satisfaction. It proposes a smart public transportation dispatching system integrating IoT technology, cloud computing, data analysis, and user interaction. Please refer to Figure 6, which is a schematic diagram of the public transportation dispatching system provided in this embodiment. The public transportation dispatching system 500 includes an electronic device 510 and a server 520. The electronic device 510 is used to acquire comprehensive traffic information and includes an IoT device 511 and a mobile terminal 512. The server 520 is connected to the IoT device 511 and the mobile terminal 512, and is used to execute the public transportation dispatching method of any of the above embodiments. The mobile terminal 512 can be a mobile terminal such as a tablet computer or a smartphone. The server 520 can include a cloud server and / or a local server, and multiple traffic dispatching models can be preset on the server 520.

[0095] The server 520 can perform the following actions: acquiring comprehensive traffic information; determining a target traffic scheduling model from multiple traffic scheduling models, with different traffic scheduling models used to predict public transportation scheduling information for different time periods; and inputting the comprehensive traffic information into the target traffic scheduling model to predict public transportation scheduling information for the corresponding time periods.

[0096] In some embodiments, the traffic scheduling model includes a short-term scheduling model, a medium-term scheduling model, and a long-term scheduling model. The short-term scheduling model is used to predict public transportation scheduling information for a first time length, the medium-term scheduling model is used to predict public transportation scheduling information for a second time length, and the long-term scheduling model is used to predict public transportation scheduling information for a third time length. The first time length is shorter than the second time length, and the second time length is shorter than the third time length. In determining the target traffic scheduling model from multiple traffic scheduling models, the server 520 can perform the following: determine the time length of the public transportation scheduling information to be predicted; if the time length does not exceed the first time length, then determine the short-term scheduling model as the target traffic scheduling model; if the time length exceeds the first time length but does not exceed the second time length, then determine the medium-term scheduling model as the target traffic scheduling model; if the time length exceeds the second time length but does not exceed the third time length, then determine the long-term scheduling model as the target traffic scheduling model.

[0097] In some embodiments, the integrated traffic information includes real-time integrated traffic information and historical integrated traffic information. In determining the time length of the public transportation scheduling information to be predicted, the server 520 may perform the following: analyze the real-time integrated traffic information and historical integrated traffic information to identify the fluctuation characteristics of the integrated traffic information, the fluctuation characteristics including at least one of temporary fluctuation changes, periodic changes, and long-term trends; and determine the time length of the public transportation scheduling information to be predicted based on the duration of the impact of the fluctuation characteristics on public transportation.

[0098] In some embodiments, if the target traffic scheduling model is determined to be a short-term scheduling model, in the process of inputting comprehensive traffic information into the target traffic scheduling model to predict public transportation scheduling information of the corresponding time length, the server 520 may perform the following: obtain the first comprehensive traffic information corresponding to the fourth time length; input the first comprehensive traffic information into the short-term scheduling model to predict the first public transportation scheduling information of the first time length.

[0099] In some embodiments, if the target traffic scheduling model is determined to be a medium-term scheduling model, in the process of inputting comprehensive traffic information into the target traffic scheduling model to predict public transportation scheduling information for the corresponding time length, the server 520 may perform the following: obtain the second comprehensive traffic information corresponding to the fifth time length; input the second comprehensive traffic information into the medium-term scheduling model to predict the second public transportation scheduling information for the second time length; or input the first public transportation scheduling information and the second comprehensive traffic information into the medium-term scheduling model to predict the second public transportation scheduling information for the second time length.

[0100] In some embodiments, if the target traffic scheduling model is determined to be a long-term scheduling model, in the process of inputting comprehensive traffic information into the target traffic scheduling model to predict public transportation scheduling information for the corresponding time length, the server 520 may perform the following actions: obtaining the third comprehensive traffic information corresponding to the sixth time length; inputting the third comprehensive traffic information into the long-term scheduling model to predict the third public transportation scheduling information for the third time length; or inputting the second public transportation scheduling information and the third comprehensive traffic information into the long-term scheduling model to predict the third public transportation scheduling information for the third time length.

[0101] In some embodiments, before determining the target traffic scheduling model from multiple traffic scheduling models, the server 520 may perform the following: collect multiple historical comprehensive traffic information of different time lengths; perform machine learning on the historical comprehensive traffic information of each time length to obtain a traffic scheduling model corresponding to each time length.

[0102] In some embodiments, when machine learning is performed on historical comprehensive traffic information for each time period to obtain a traffic scheduling model corresponding to each time period, the server 520 can perform the following: dividing the historical comprehensive traffic information for each time period into a training dataset, a validation dataset, and a test dataset according to a preset ratio; performing machine learning on the training dataset to construct an initial traffic scheduling model corresponding to the time period; evaluating the performance of the initial traffic scheduling model using the validation dataset and adjusting the parameters of the initial traffic scheduling model based on the performance evaluation results to optimize the initial traffic scheduling model; evaluating the generalization performance of the optimized initial traffic scheduling model using the test dataset; and obtaining the traffic scheduling model corresponding to the time period when the generalization performance of the optimized initial traffic scheduling model reaches the preset generalization capability requirement.

[0103] In some embodiments, when inputting comprehensive traffic information into a target traffic scheduling model to predict public transportation scheduling information for a corresponding time period, the server 520 may perform the following: filtering relevant traffic information affecting public transportation from the comprehensive traffic information; cleaning and standardizing the relevant traffic information; performing dimensionality reduction processing on the cleaned and standardized relevant traffic information; extracting features from the dimensionality-reduced relevant traffic information using at least one of time series analysis, signal processing technology, and feature engineering to extract derived traffic information; and inputting the derived traffic information into the target traffic scheduling model to predict public transportation scheduling information for a corresponding time period.

[0104] For example, the deployment of the public transportation dispatch system 500 is described below. In this embodiment, when deploying the public transportation dispatch system 500, the Internet of Things (IoT) device 511 is selected and deployed. For example, IoT devices 511 are deployed on buses, bus stops, or other transportation facilities. IoT devices 511 may include GPS trackers, passenger counters, environmental sensors, etc., which can be used to collect vehicle information, passenger information, environmental information, etc., including the real-time location, speed, passenger capacity, and other operational information of vehicles, the number of passengers boarding and alighting at each stop, the waiting status of passengers at bus stops, and weather conditions, etc. This extensive information collection provides the public transportation dispatch system 500 with a comprehensive real-time perspective and uploads this information to the server 520.

[0105] To enable passengers to interact directly with the public transportation dispatch system 500, a user-friendly application can be developed on a mobile terminal 512, such as a smartphone. Passengers can input their travel needs in advance through the application, including pick-up location, destination, and estimated time. This information is uploaded to the server 520 via the application, providing the server 520 with richer data support for accurate dispatch prediction. In addition, passengers can also use the application to view vehicle arrival times or provide service quality feedback. The use of the application helps collect various passenger-related data and increases the interactivity between passengers and the public transportation dispatch system 500, allowing passengers to easily participate in the optimization of public transportation services.

[0106] The public transportation dispatch system 500 in this embodiment also establishes a powerful server 520. Server 520 can be a cloud server, where various comprehensive traffic information collected by IoT devices 511 and applications can be uploaded to server 520. For example, the collected information can be transmitted in real-time from the IoT devices 511 and mobile terminal 512 applications to server 520 via secure and efficient communication protocols (such as MQTT and HTTP). A database can be established on server 520 to store this large amount of information. This embodiment does not limit the type of database, as long as it has high availability and scalability and can meet the storage needs of large-scale data. For example, it can be a relational database (such as MySQL) or a non-relational database (such as MongoDB). Of course, server 520 can also obtain comprehensive traffic information in other ways, such as information collected periodically from authoritative media and official websites.

[0107] Server 520 is not only a data storage center but also the brain of data analysis and processing. Integrating advanced machine learning and artificial intelligence algorithms, Server 520 can predict public transportation scheduling information for a future period based on collected comprehensive traffic information and a pre-set traffic scheduling model. Server 520 can then send the predicted public transportation scheduling information to traffic management departments, which can then use this information to optimize traffic scheduling, such as departure frequency, routes, and stop locations.

[0108] After the initial deployment of the public transportation dispatch system 500, testing will be conducted, including unit testing, integration testing, and system testing. Based on the test results, necessary adjustments and optimizations will be made to ensure the stability and performance of each component and the overall system.

[0109] In addition, public transport drivers will be trained to familiarize themselves with the operation of the public transport dispatch system 500, ensuring their smooth use of the system. Simultaneously, publicity and education activities will be conducted to increase passenger awareness and willingness to use the public transport dispatch system 500. For example, based on this embodiment of the public transport dispatch system 500, after the server 520 collects comprehensive traffic information, it can instantly transmit information useful to drivers to the onboard equipment of vehicles approaching the platform. This allows drivers to know the number of passengers waiting at the platform and their destination information before arriving at the platform, enabling them to flexibly adjust their driving plans and stopping times to better meet passenger needs and improve service responsiveness. For instance, drivers can optimize stopping times and speeds based on real-time passenger data, reducing unnecessary waiting and empty runs, providing more punctual service, improving overall transportation efficiency, enhancing passenger satisfaction, and achieving a more efficient, environmentally friendly, and passenger-friendly public transport system.

[0110] After ensuring the stable operation of the public transportation dispatch system 500 and that all functions achieve the expected results, full deployment will be carried out. Simultaneously, a monitoring and maintenance mechanism for the public transportation dispatch system 500 will be established to monitor the operational status of public transportation and changes in passenger demand in real time, promptly identify and resolve problems, and ensure the long-term stable operation of the public transportation dispatch system 500. Furthermore, the public transportation dispatch system 500 in this embodiment will be continuously improved and optimized. Based on the operational status of the public transportation dispatch system 500 and feedback from passengers and drivers, data analysis and system optimization will be continuously performed to adapt to changes in passenger demand and technological advancements.

[0111] Through this series of closely linked and integrated steps, the goal is to achieve an efficient, flexible, and user-friendly intelligent public transportation dispatch system 500. This not only improves operational efficiency and passenger satisfaction but also contributes to the sustainable development of urban transportation. With continuous technological advancements and enhanced data analysis capabilities, the system's accuracy and effectiveness will be further improved, providing strong technical support and innovative impetus for the future development of public transportation systems.

[0112] Please refer to Figure 7, which is a schematic diagram of the operation of the public transportation dispatch system provided in the embodiment of this application.

[0113] The public transportation dispatch system in this embodiment combines Internet of Things technology, mobile terminal applications, and cloud data analysis to form a smart public transportation dispatch system 500, aiming to improve the efficiency of public transportation and passenger satisfaction.

[0114] The public transportation dispatch system 500 of this embodiment can collect various comprehensive traffic dispatch information in real time through the application of the Internet of Things (IoT) device 511 and the mobile terminal 512, such as vehicle information, passenger information, and environmental information. For example, the IoT device 511 in the vehicle or on the website can collect data such as vehicle location, speed, passenger capacity, and passenger waiting status in real time. Passengers can input their travel needs in advance through the application, including information such as pick-up location, destination, and expected pick-up time, and upload the collected data to the server 520.

[0115] Server 520 is a cloud platform capable of storing, processing, and analyzing massive amounts of data. Server 520 integrates machine learning algorithms and, based on the acquired comprehensive traffic scheduling information, can output public transportation scheduling information using any of the public transportation scheduling methods described above. Traffic management departments can optimize public transportation scheduling in real time based on the public transportation scheduling information provided by server 520, such as adjusting vehicle departure frequencies, planning optimal routes and stops, etc. The public transportation scheduling information can also be updated to IoT device 511, enabling IoT device 511 to receive updated scheduling information in real time and provide passengers with real-time public transportation information, such as vehicle arrival times and route changes, thereby enhancing the passenger travel experience.

[0116] The public transportation dispatch system also provides an instant feedback mechanism to collect real-time feedback from drivers and passengers, such as drivers reporting passenger boarding and alighting status and passengers evaluating their riding experience, and sending the real-time feedback to server 520. Passengers and drivers can submit their feedback through applications on mobile terminals 512 or IoT devices 511, and this feedback will be transmitted to server 520.

[0117] The machine learning algorithms integrated into Server 520 can continuously learn and optimize based on collected data and real-time feedback, improving the accuracy of predictions and ensuring that Public Transport Dispatch System 500 can dynamically adjust public transport dispatch information in a timely and accurate manner to adapt to changes in demand.

[0118] Through the implementation of the public transportation dispatch system 500 in this embodiment, the public transportation dispatch system will be transformed from a relatively static, manual-reliant model to a highly dynamic, data-driven service system. This transformation will bring advantages in many aspects, including convenience, efficiency, improved passenger satisfaction, adaptability and flexibility, and sustainable development, injecting new vitality and impetus into the future development of urban transportation systems. This transformation will bring the following significant advantages:

[0119] From the passenger's perspective, this embodiment significantly improves the efficiency and flexibility of public transportation scheduling, such as bus scheduling, by collecting and analyzing passenger travel needs in real time, reducing passenger waiting time, and more accurately meeting passenger needs, thereby greatly improving passenger satisfaction.

[0120] For example, please refer to Figure 8, which is a schematic diagram of another scenario of the operation of the public transportation dispatch system provided in this application embodiment. Passengers submit their travel needs to the public transportation dispatch system 500 through the interactive interface of the mobile terminal 512 application or the Internet of Things device 511, including information such as boarding location, destination, and expected boarding time. After receiving the passenger's travel needs, the public transportation dispatch system 500 analyzes the data in conjunction with other data sources (such as the real-time location, speed, and passenger capacity of buses), selects a suitable traffic dispatch model, predicts future passenger flow based on passenger demand and traffic conditions, and generates corresponding public transportation dispatch information. The traffic management department can use the public transportation dispatch information predicted by the public transportation dispatch system 500 to carry out actual public transportation vehicle dispatching work, such as adjusting vehicle departure frequency, planning routes, and selecting stops. According to the instructions of the traffic management department, vehicle drivers execute the dispatch plan and go to various stops to pick up and drop off passengers. By optimizing vehicle dispatching, passenger waiting time can be reduced, fuel consumption can be reduced, and faster public transportation services can be provided. Furthermore, the public transportation dispatch system 500 can continuously improve its predictive algorithms and dispatch schemes based on long-term collected data and feedback, enabling traffic management departments to respond to current passenger demands in a timely manner based on more accurate predictions and to improve the public transportation system in the long term.

[0121] Regarding drivers, in accordance with the public transportation dispatch plan, drivers' working hours and routes are reasonably arranged to ensure drivers' working hours, avoid drivers being overworked or idle, and ensure service quality. Moreover, before arriving at the platform, drivers can know the number of passengers waiting on the platform and their destination information through onboard equipment or other devices. Drivers can flexibly adjust their driving plans and stopping times to better meet passenger needs and improve service responsiveness.

[0122] In terms of alleviating urban traffic congestion, this embodiment can effectively reduce traffic pressure on main roads and improve the overall smoothness of urban traffic through more efficient public transportation scheduling. For example, during peak hours, the departure interval can be shortened and the number of trips can be increased to meet high demand; during off-peak hours, the departure interval can be extended and the number of trips can be reduced to lower operating costs.

[0123] In terms of environmental protection, this embodiment further reduces fuel consumption and greenhouse gas emissions by optimizing the operating efficiency of public transportation vehicles such as buses. Vehicles can be precisely scheduled according to actual demand, avoiding unnecessary empty runs and overloading, thus achieving a more environmentally friendly public transportation service.

[0124] Furthermore, the public transportation dispatching system in this embodiment possesses powerful data-driven decision-making capabilities. By collecting and analyzing a large amount of passenger travel data, transportation management departments can gain a deeper understanding of passenger behavior and demand patterns, providing a scientific basis and strong support for long-term transportation planning and improvement.

[0125] In summary, this technology not only significantly improves the transportation efficiency and passenger satisfaction of public transportation such as buses, but also has a positive impact on urban traffic congestion, environmental protection, and transportation planning. It is an innovative technology with broad application prospects and significant social benefits, and is expected to play an even more important role in future public transportation systems.

[0126] Please refer to Figures 9 and 10. Figure 9 is a structural schematic diagram of the public transportation dispatching device provided in an embodiment of this application, and Figure 10 is another structural schematic diagram of the public transportation dispatching device provided in an embodiment of this application. The public transportation dispatching device 600 may include an acquisition module 601, a determination module 602, and a prediction module 603.

[0127] The acquisition module 601 is used to: acquire comprehensive traffic information, which includes at least one of vehicle information, passenger information, environmental information, time information, and event information;

[0128] The determination module 602 is used to: determine the target traffic scheduling model from multiple traffic scheduling models, with different traffic scheduling models used to predict public transportation scheduling information for different time periods;

[0129] The prediction module 603 is used to input comprehensive traffic information into the target traffic scheduling model in order to predict public transportation scheduling information for a corresponding time period.

[0130] In some embodiments, the traffic scheduling model includes a short-term scheduling model, a medium-term scheduling model, and a long-term scheduling model. The short-term scheduling model is used to predict public transportation scheduling information for a first time length, the medium-term scheduling model is used to predict public transportation scheduling information for a second time length, and the long-term scheduling model is used to predict public transportation scheduling information for a third time length. The first time length is shorter than the second time length, and the second time length is shorter than the third time length. The determining module 602 can be used to: determine the time length of the public transportation scheduling information to be predicted; if the time length does not exceed the first time length, then determine the short-term scheduling model as the target traffic scheduling model; if the time length exceeds the first time length but does not exceed the second time length, then determine the medium-term scheduling model as the target traffic scheduling model; if the time length exceeds the second time length but does not exceed the third time length, then determine the long-term scheduling model as the target traffic scheduling model.

[0131] In some embodiments, the integrated traffic information includes real-time integrated traffic information and historical integrated traffic information. The determining module 602 can be used to: analyze the real-time integrated traffic information and historical integrated traffic information to identify the fluctuation characteristics of the integrated traffic information, the fluctuation characteristics including at least one of temporary fluctuation changes, periodic changes, and long-term trends; and determine the time length of the public transportation scheduling information to be predicted based on the duration of the impact of the fluctuation characteristics on public transportation.

[0132] In some embodiments, if the target traffic scheduling model is determined to be a short-term scheduling model, the prediction module 603 can be used to: obtain first comprehensive traffic information corresponding to the fourth time length; input the first comprehensive traffic information into the short-term scheduling model to predict the first public transportation scheduling information for the next first time length.

[0133] In some embodiments, if the target traffic scheduling model is determined to be a medium-term scheduling model, the prediction module 603 can be used to: obtain second integrated traffic information corresponding to the fifth time length; input the second integrated traffic information into the medium-term scheduling model to predict the second public transportation scheduling information in the next second time length; or input the first public transportation scheduling information and the second integrated traffic information into the medium-term scheduling model to predict the second public transportation scheduling information in the next second time length.

[0134] In some embodiments, if the target traffic scheduling model is determined to be a long-term scheduling model, the prediction module 603 can be used to: obtain the third integrated traffic information corresponding to the sixth time length; input the third integrated traffic information into the long-term scheduling model to predict the third public transportation scheduling information for the next third time length; or input the second public transportation scheduling information and the third integrated traffic information into the long-term scheduling model to predict the third public transportation scheduling information for the next third time length.

[0135] In some embodiments, the public transportation dispatching device 600 further includes a training module 604, which can be used to: collect multiple historical comprehensive traffic information of different time lengths; perform machine learning on the historical comprehensive traffic information of each time length to obtain a traffic dispatching model corresponding to each time length.

[0136] In some embodiments, the training module 604 can be used to: divide historical comprehensive traffic information for each time period into a training dataset, a validation dataset, and a test dataset according to a preset ratio; use the training dataset to perform machine learning on the machine learning model to construct an initial traffic scheduling model corresponding to the time period; use the validation dataset to evaluate the performance of the initial traffic scheduling model, and adjust the parameters of the initial traffic scheduling model according to the performance evaluation results to optimize the initial traffic scheduling model; use the test dataset to evaluate the generalization performance of the optimized initial traffic scheduling model, and when the generalization performance of the optimized initial traffic scheduling model reaches the preset generalization capability requirement, the traffic scheduling model corresponding to the time period is obtained.

[0137] In some embodiments, the prediction module 603 can be used to: filter relevant traffic information affecting public transportation from comprehensive traffic information; clean and standardize the relevant traffic information; perform dimensionality reduction on the cleaned and standardized relevant traffic information; extract features from the dimensionality-reduced relevant traffic information using at least one of time series analysis, signal processing technology, and feature engineering to extract derived traffic information; and input the derived traffic information into the target traffic scheduling model to predict public transportation scheduling information for the corresponding time period.

[0138] This application provides a computer-readable storage medium storing a computer program. When the computer program is executed on a computer, it causes the computer to execute the process in the public transportation scheduling method provided in this embodiment.

[0139] In the above embodiments, the descriptions of each embodiment have their own emphasis. For parts not described in detail in a certain embodiment, please refer to the detailed description of the public transportation scheduling method above, which will not be repeated here.

[0140] The public transportation dispatching device provided in this application embodiment belongs to the same concept as the public transportation dispatching method in the above embodiments. Any of the methods provided in the public transportation dispatching method embodiments can be run on the public transportation dispatching device. For details of its implementation process, please refer to the public transportation dispatching method embodiments, which will not be repeated here. Regarding the public transportation dispatching device of this application embodiment, its functional modules can be integrated into a single processing chip, or each module can exist physically separately, or two or more modules can be integrated into one module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium, such as read-only memory, a magnetic disk, or an optical disc.

[0141] It should be noted that, regarding the public transportation scheduling method of this application embodiment, those skilled in the art will understand that all or part of the process of implementing the public transportation scheduling method of this application embodiment can be accomplished by controlling related hardware through a computer program. The computer program can be stored in a computer-readable storage medium, such as memory, and executed by at least one processor. During execution, it can include the process of the embodiment of the public transportation scheduling method. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), random access memory (RAM), etc.

[0142] The above provides a detailed description of the public transportation scheduling method, device, storage medium, and mobile terminal provided in the embodiments of this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of this application. At the same time, those skilled in the art will find that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

[0143] 500: Public transportation dispatch system 510: Electronic equipment 520: Server 511: Internet of Things (IoT) devices 512: Mobile terminal 600: Public transportation dispatching device 601: Get Module 602: Determine Module 603: Prediction Module 604: Training Module 101-103: Steps 201-211: Steps 301-302: Steps 3021-3024: Steps 401-405: Steps

Claims

1. A method for scheduling public transportation, comprising: The process involves: acquiring comprehensive traffic information, including at least one of vehicle information, passenger information, environmental information, time information, and event information; determining a target traffic scheduling model from multiple traffic scheduling models based on the time length corresponding to the prediction target, wherein different traffic scheduling models are used to predict public transportation scheduling information for different time lengths, each traffic scheduling model corresponding to a different time length, and each traffic scheduling model is obtained through machine learning using historical comprehensive traffic information for the corresponding time length; and inputting the comprehensive traffic information into the target traffic scheduling model to predict public transportation scheduling information for the corresponding time length.

2. The public transportation scheduling method according to claim 1, wherein, The traffic scheduling model includes a short-term scheduling model, a medium-term scheduling model, and a long-term scheduling model. The short-term scheduling model is used to predict public transportation scheduling information within a first time period. The medium-term scheduling model is used to predict public transportation scheduling information within a second time period. The long-term scheduling model is used to predict public transportation scheduling information within a third time period. The first time period is shorter than the second time period, and the second time period is shorter than the third time period. Determining the target traffic scheduling model from multiple traffic scheduling models includes: determining the time period for the public transportation scheduling information to be predicted; if the time period does not exceed the first time period, then the short-term scheduling model is determined as the target traffic scheduling model; if the time period exceeds the first time period but does not exceed the second time period, then the medium-term scheduling model is determined as the target traffic scheduling model; if the time period exceeds the second time period but does not exceed the third time period, then the long-term scheduling model is determined as the target traffic scheduling model.

3. The public transportation scheduling method according to claim 2, wherein, If the target traffic scheduling model is determined to be the short-term scheduling model, then inputting the comprehensive traffic information into the target traffic scheduling model to predict public transportation scheduling information for the corresponding time length includes: obtaining the first comprehensive traffic information corresponding to the fourth time length; and inputting the first comprehensive traffic information into the short-term scheduling model to predict the first public transportation scheduling information for the first time length.

4. The public transportation scheduling method according to claim 3, wherein, If the target traffic scheduling model is determined to be the medium-term scheduling model, then inputting the integrated traffic information into the target traffic scheduling model to predict public transportation scheduling information for the corresponding time length includes: obtaining second integrated traffic information corresponding to the fifth time length, wherein the fifth time length is greater than the fourth time length; inputting the second integrated traffic information into the medium-term scheduling model to predict second public transportation scheduling information for the second time length; or inputting the first public transportation scheduling information and the second integrated traffic information into the medium-term scheduling model to predict second public transportation scheduling information for the second time length.

5. The public transportation scheduling method according to claim 4, wherein, If the target traffic scheduling model is determined to be the long-term scheduling model, then inputting the integrated traffic information into the target traffic scheduling model to predict public transportation scheduling information for the corresponding time length includes: obtaining third integrated traffic information corresponding to the sixth time length, wherein the sixth time length is greater than the fifth time length; inputting the third integrated traffic information into the long-term scheduling model to predict third public transportation scheduling information for the third time length; or inputting the second public transportation scheduling information and the third integrated traffic information into the long-term scheduling model to predict third public transportation scheduling information for the third time length.

6. The public transportation scheduling method according to claim 2, wherein, The integrated traffic information includes real-time integrated traffic information and historical integrated traffic information. The process of determining the time length of the public transportation scheduling information to be predicted includes: analyzing the real-time integrated traffic information and the historical integrated traffic information to identify the fluctuation characteristics of the integrated traffic information, wherein the fluctuation characteristics include at least one of temporary fluctuations, periodic changes, and long-term trends; and determining the time length of the public transportation scheduling information to be predicted based on the duration of the impact of the fluctuation characteristics on public transportation.

7. The public transportation scheduling method according to claim 1, wherein, Before determining the target traffic scheduling model from multiple traffic scheduling models, the method further includes: collecting multiple historical comprehensive traffic information of different time lengths; and performing machine learning on the historical comprehensive traffic information of each time length to obtain a traffic scheduling model corresponding to each time length.

8. The public transportation scheduling method according to claim 7, wherein, The step of performing machine learning on historical comprehensive traffic information for each time period to obtain a traffic scheduling model corresponding to each time period includes: dividing the historical comprehensive traffic information for each time period into a training dataset, a validation dataset, and a test dataset according to a preset ratio; performing machine learning on the training dataset to construct an initial traffic scheduling model corresponding to the time period; evaluating the performance of the initial traffic scheduling model using the validation dataset and adjusting the parameters of the initial traffic scheduling model based on the performance evaluation results to optimize the initial traffic scheduling model; evaluating the generalization performance of the optimized initial traffic scheduling model using the test dataset; and obtaining the traffic scheduling model corresponding to the time period when the generalization performance of the optimized initial traffic scheduling model reaches a preset generalization capability requirement.

9. The public transportation scheduling method according to claim 1, wherein, The step of inputting the comprehensive traffic information into the target traffic dispatch model to predict public transportation dispatch information for a corresponding time period includes: filtering relevant traffic information affecting public transportation from the comprehensive traffic information; cleaning and standardizing the relevant traffic information; performing dimensionality reduction processing on the cleaned and standardized relevant traffic information; extracting features from the dimensionality-reduced relevant traffic information using at least one of time series analysis, signal processing technology, and feature engineering to extract derived traffic information; and inputting the derived traffic information into the target traffic dispatch model to predict public transportation dispatch information for a corresponding time period.

10. A public transportation dispatching system, the dispatching system comprising: Electronic devices for acquiring comprehensive traffic information, including mobile terminals and Internet of Things (IoT) devices; A server, connected to the IoT device and the mobile terminal, is used to execute the method described in any one of requests 1 to 9.

11. A public transportation dispatching device, the device comprising: An acquisition module is used to acquire comprehensive traffic information, wherein the comprehensive traffic information includes at least one of vehicle information, passenger information, environmental information, time information, and event information; a determination module is used to determine a target traffic scheduling model from multiple traffic scheduling models based on the time length corresponding to the prediction target, wherein different traffic scheduling models are used to predict public transportation scheduling information for different time lengths, wherein each traffic scheduling model corresponds to a different time length, and each traffic scheduling model is obtained by machine learning using historical comprehensive traffic information for the corresponding time length; a prediction module is used to input the comprehensive traffic information into the target traffic scheduling model to predict public transportation scheduling information for the corresponding time length, wherein each traffic scheduling model corresponds to a different time length, and each traffic scheduling model is obtained by machine learning using historical comprehensive traffic information for the corresponding time length.

12. A storage medium storing a computer program that, when run on a computer, causes the computer to perform the method described in any one of requests 1 to 9.

Images