Passenger transportation travel data processing system and method based on large model

Abstract

The invention discloses a passenger transportation data processing system and method based on a large model, and relates to the technical field of data processing. The method comprises the steps of collecting real-time passenger transport travel data, and performing fusion and standardization processing by using an entropy weight method and a time attenuation weight. The LSTM network is combined with the improved GARCH model to predict the passenger flow in the next week, and peak and valley periods are identified. And calculating a risk value and a confidence interval of the prediction error variance, and providing a detailed report. The scheduling suggestion optimization function considers factors such as total scheduling cost, time deviation penalty, passenger importance and the like, and adjusts to meet special requirements. And automatically or manually adjusting the operation plan according to the optimal scheduling suggestion, including increasing shifts, adjusting routes and the like. In the peak period, the special group is preferentially served; in the weekend or a specific activity day, special line service is added, and the operation time is prolonged. According to the method, the accuracy and real-time performance of passenger transportation data processing are improved, and the service quality and efficiency can be improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of data processing, and in particular relates to a passenger travel data processing system and method based on a large model. Background Art

[0002] With the acceleration of urbanization and the increase in population mobility, public transportation systems are facing unprecedented challenges. Traditional public transportation scheduling methods are often based on experience or fixed timetables, which are difficult to adapt to the dynamic changes in passenger flow in real time, resulting in low transportation efficiency, low passenger satisfaction, and waste of resources.

[0003] In order to meet these challenges, the industry has begun to explore the use of big data and artificial intelligence technologies to optimize public transportation systems. Among them, passenger travel data processing methods based on large models have become a popular research direction. This method can more accurately grasp the changing trends of passenger flow and the impact of special events by collecting and analyzing real-time data from multiple data sources (such as bus companies, subway systems, traffic management departments, weather forecast services, social media, etc.).

[0004] However, in practical applications, these real-time data often have problems such as inconsistent formats, redundant information, and noise interference, which require effective fusion and standardization. In addition, due to the time series characteristics of passenger flow, traditional forecasting methods often have difficulty accurately capturing its long-term trends and short-term fluctuations. Therefore, it is necessary to introduce advanced forecasting models, such as long short-term memory networks (LSTM) and generalized autoregressive conditional heteroskedasticity models (GARCH), to improve forecasting accuracy and stability.

[0005] On this basis, the present invention proposes a passenger travel data processing method based on a large model. This method realizes the intelligent scheduling and optimization of the public transportation system through data collection and standardization, LSTM and GARCH model analysis, risk assessment, scheduling suggestion optimization, and scheduling execution monitoring. This method not only improves transportation efficiency and passenger satisfaction, but also reduces operating costs and environmental impacts, and has important practical application value and promotion prospects. Summary of the invention

[0006] In order to overcome the shortcomings and deficiencies of the above-mentioned prior art, the first purpose of the present invention is to provide a passenger travel data processing system based on a large model; the second purpose of the present invention is to provide a passenger travel data processing system based on a large model.

[0007] The first object of the present invention adopts the following technical solution: A passenger travel data processing system based on a large model, including a data standardization module, a model analysis module, a risk assessment module, a scheduling suggestion optimization module, and a scheduling execution monitoring module; Data standardization module: used to collect real-time passenger travel data from multiple data sources, and use entropy weight method and time decay weight to fuse and standardize the data; Model analysis module: used to organize the integrated passenger travel data into a time series format, input it into the LSTM network for training and prediction, and output the passenger flow prediction results for each time period in the next week; analyze the prediction residuals and evaluate the prediction uncertainty and stability; Risk assessment module: used to calculate the risk value and confidence interval for each forecast error variance; Scheduling suggestion optimization module: used to construct a scheduling suggestion optimization function based on the predicted passenger status to obtain the best scheduling plan; adjust the scheduling plan considering special needs; Scheduling execution monitoring module: used to automatically or manually adjust the operation plan of public transportation according to the generated optimal scheduling suggestions; regularly monitor the implementation effect and dynamically adjust the scheduling strategy according to the actual situation.

[0008] The second object of the present invention adopts the following technical solution: The passenger travel data processing method based on a large model is used to realize a passenger travel data processing system based on a large model. The method flow is as follows: Step 1: Data collection and standardization, including collecting real-time data from multiple data sources, using API interfaces or direct database access to obtain data, and using entropy weight method and time decay weight to fuse and standardize the data; Step 2: Use the LSTM network combined with the improved GARCH model to analyze the fused data, predict the passenger flow in each time period in the next week, and identify potential peak and valley periods based on the prediction results; Step 3: Calculate the risk value and confidence interval for each forecast error variance, and provide a detailed report containing the risk value, forecast error variance and confidence interval; Step 4: Construct a scheduling suggestion optimization function based on the predicted passenger status to obtain the best scheduling plan, and adjust the scheduling plan considering special needs; Step 5: Execute the scheduling plan based on the generated optimal scheduling suggestion, including adjusting the operation plan of public transportation tools, increasing the frequency of trips, adjusting routes, etc., and regularly monitor the implementation effect, and dynamically adjust the scheduling strategy according to the actual situation.

[0009] Preferably, the data collection in step one includes collecting real-time data from bus companies, subway systems, traffic management departments, weather forecast services, and social media, including vehicle locations, number of passengers, route information, weather conditions, and special events.

[0010] Preferably, the data standardization process in step 1 uses the entropy weight method to calculate the amount of information of each feature in each data source to determine its importance, and introduces the time decay weight to adjust the importance of the feature. The fused data is calculated by a specific formula, as follows: The entropy weight method is used to calculate the information content of each feature in each data source. , to determine its importance: ; in, It is The feature in The probability distribution of samples; is the sample size; Introducing time decay weights To adjust the importance of features: ; in, is the difference between the data collection time and the current time; is the decay constant; is the number of features; The fused data Calculated by the following formula: ; in, It is Features It is The variance of the features; is the adjustment factor.

[0011] As a preferred method, in step 2, the fused passenger data is trained and predicted using a long short-term memory network, and the prediction residuals are analyzed in combination with an improved generalized autoregressive conditional heteroskedasticity model to evaluate the prediction uncertainty and stability. The formula is as follows: ; in, It's time The forecast error variance of is a constant term describing the baseline fluctuation; is the influence coefficient of the previous period’s forecast error; It's in time The prediction error of is the coefficient of the variance of the forecast error in the previous period; It is the adjustment parameter of the influence degree of nonlinear effect; is the nonlinear coefficient; is the index of the periodic term; is the number of periodic terms; is the frequency parameter related to the period; Based on the forecast results, potential peak and off-peak periods can be identified to provide a basis for subsequent scheduling.

[0012] Preferably, the risk assessment in step three includes calculating a risk value and a confidence interval for each forecast error variance to identify periods of greater uncertainty and provide a detailed report to decision makers or system administrators.

[0013] As a preference, the scheduling suggestion optimization function in step 4 considers the total scheduling cost, the basic cost of the plan, the weight factor of the time deviation penalty intensity, the importance weight of the passenger, the actual arrival time and the expected arrival time of the passenger, and adjusts the scheduling plan considering special needs; Among them, the scheduling suggestion optimization function is as follows: ; in, It's in time The total dispatch cost of is a set of scheduling schemes An element in It is a plan Basic costs; is the weight factor that adjusts the intensity of the time deviation penalty; Passenger Importance weight of It is in the plan Passengers The actual arrival time; Passenger expected arrival time; Special needs include the elderly and the disabled.

[0014] Preferably, executing the best scheduling plan in step five includes automatically or manually adjusting the operation plan of public transportation vehicles, increasing the frequency, adjusting the routes, and dynamically adjusting the scheduling strategy according to actual conditions to continuously optimize the service quality.

[0015] As a preference, during the morning rush hour from Monday to Friday, public transportation operation plans are automatically adjusted to include increasing frequency, giving priority to special groups, and dynamically adjusting routes; during weekends or specific event days, dedicated line services are added and operating hours are extended.

[0016] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are: 1. The present invention collects real-time data from multiple data sources, and uses entropy weight method and time decay weight to fuse and standardize the data, effectively solving the problem of wide data sources and different formats. This data processing method improves the accuracy and consistency of the data. At the same time, the collection and processing of real-time data enables the system to timely reflect the latest status of passenger travel; 2. The present invention analyzes and predicts the fused data by combining the long short-term memory network (LSTM) and the improved generalized autoregressive conditional heteroskedasticity model (GARCH). The LSTM network can capture the long-term dependencies in time series data, while the improved GARCH model can evaluate the uncertainty and stability of the prediction. This combined use not only improves the accuracy of passenger flow prediction, but also identifies potential peak and valley periods; 3. The present invention constructs a scheduling suggestion optimization function based on the predicted passenger status, and takes into account multiple factors such as the total scheduling cost, the basic cost of the plan, the weight factor of the time deviation penalty intensity, the importance weight of the passenger, and the actual arrival time and expected arrival time of the passenger. In addition, adjustments are made for special needs (such as priority for the elderly and the disabled). This scheduling suggestion optimization function not only makes the scheduling plan more in line with actual needs, but also can dynamically adjust the scheduling strategy according to actual conditions to continuously optimize the service quality. At the same time, the implementation of measures such as automatic or manual adjustment of the operation plan of public transportation, increase of flights, and adjustment of routes further improves the efficiency of passenger travel and passenger satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 The module diagram of the passenger travel data processing system based on the large model of the present invention is shown; Figure 2 A flow chart of the passenger travel data processing method based on a large model of the present invention is shown. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] In addition, the described features, structures or characteristics may be combined in one or more example embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the example embodiments of the present disclosure. However, those skilled in the art will appreciate that the technical solutions of the present disclosure may be practiced while omitting one or more of the specific details, or other methods, components, steps, etc. may be adopted. In other cases, well-known structures, methods, implementations or operations are not shown or described in detail to avoid obscuring various aspects of the present disclosure.

[0021] Embodiment 1:

[0022] See also Figure 1 As shown, the passenger travel data processing system based on a large model in this embodiment includes a data standardization module, a model analysis module, a risk assessment module, a scheduling suggestion optimization module and a scheduling execution monitoring module.

[0023] Data standardization module: used to collect real-time passenger travel data from multiple data sources, and use entropy weight method and time decay weight to fuse and standardize the data.

[0024] Model analysis module: used to organize the integrated passenger travel data into a time series format, input it into the LSTM network for training and prediction, and output the passenger flow prediction results for each time period in the next week; analyze the prediction residuals and evaluate the prediction uncertainty and stability.

[0025] Risk Assessment Module: used to calculate the risk value and confidence interval for each prediction error variance.

[0026] Scheduling suggestion optimization module: used to construct a scheduling suggestion optimization function based on the predicted passenger status to obtain the best scheduling plan; adjust the scheduling plan considering special needs (such as priority for the elderly and disabled).

[0027] Scheduling execution monitoring module: used to automatically or manually adjust the operation plan of public transportation according to the generated optimal scheduling suggestions; regularly monitor the implementation effect and dynamically adjust the scheduling strategy according to the actual situation.

[0028] The beneficial effects of this embodiment include: improving data processing efficiency and accuracy, accurately predicting passenger flow, effectively evaluating prediction risks, optimizing scheduling plans to meet special needs, and realizing automatic or manual adjustment of operation plans, ensuring continuous optimization of services, and improving the overall operational efficiency of public transportation.

[0029] Embodiment 2:

[0030] See also Figure 2 As shown, the passenger travel data processing method based on the large model of this embodiment has the following process: Step 1: Data collection and standardization.

[0031] S11. Collect real-time data from multiple data sources (e.g., bus companies, subway systems, transportation management departments, weather forecast services, social media), including but not limited to vehicle locations, passenger numbers, route information, weather conditions, special events (e.g., holidays, city events); Use API interface or direct database access to obtain data and ensure consistency in data format.

[0032] S12. Use entropy weight method and time decay weight to fuse and standardize data to ensure that the data format is consistent and valid.

[0033] The entropy weight method is used to calculate the information content of each feature in each data source. , to determine its importance: ; in, It is The feature in The probability distribution of samples; is the sample size.

[0034] Introducing time decay weights To adjust the importance of features, consider newer data to be more valuable than older data: ; in, is the difference between the data collection time and the current time; is the decay constant; is the number of features.

[0035] The fused data Calculated by the following formula: ; in, It is Features It is The variance of the features; is the adjustment factor.

[0036] Step 2: Use the LSTM network combined with the improved GARCH model to analyze the fused data and predict the passenger flow in each time period in the next week.

[0037] The fused passenger data is organized into a time series format and input into the long short-term memory network (LSTM) for training and prediction.

[0038] The design of LSTM network structure should consider the selection of hyperparameters such as the number of layers and the number of neurons in each layer, which can be tuned through grid search or Bayesian optimization.

[0039] The forecast residuals are analyzed in combination with the improved generalized autoregressive conditional heteroskedasticity model (GARCH) to assess forecast uncertainty and stability.

[0040] ; in, It's time The forecast error variance of is a constant term describing the baseline fluctuation; is the influence coefficient of the previous period’s forecast error; It's in time The prediction error of is the coefficient of the variance of the forecast error in the previous period; It is the adjustment parameter of the influence degree of nonlinear effect; is the nonlinear coefficient; is the index of the periodic term; is the number of periodic terms; is a frequency parameter related to the period.

[0041] Based on the prediction results, possible peak and off-peak periods can be identified to provide a basis for subsequent scheduling.

[0042] For example, the forecast results show that during the morning peak hours (7:00-9:00) from Monday to Friday, the passenger flow at downtown Station A and Station B will reach 10,000 and 8,000 people per hour respectively; while during the afternoon peak hours (17:00-19:00), the passenger flow at these two stations will be 9,000 and 7,500 people per hour respectively.

[0043] Step 3: Risk assessment.

[0044] The risk value and confidence interval were calculated for each forecast error variance, and it was found that the risk value of the Monday morning peak was the highest and the confidence interval was wide, indicating that the uncertainty during this period was large and additional resources might be needed.

[0045] Provide detailed reports including risk value, prediction error variance and confidence interval to decision makers or system administrators to help formulate more reasonable scheduling plans.

[0046] Step 4: Based on the predicted passenger status (such as passenger flow peak and valley), construct a scheduling suggestion optimization function to obtain the best scheduling solution. The scheduling suggestion optimization function is as follows: ; in, It's in time The total dispatch cost of is a set of scheduling schemes An element in It is a plan Basic costs; is the weight factor that adjusts the intensity of the time deviation penalty; Passenger Importance weight of It is in the plan Passengers The actual arrival time; Passenger expected arrival time.

[0047] Adjust scheduling plans to take into account special needs (such as priority for the elderly and people with disabilities) to ensure fairness and efficiency.

[0048] Step 5: Execute the best scheduling plan.

[0049] Based on the generated optimal scheduling recommendations, the operation plan of public transportation can be automatically or manually adjusted to increase frequency, adjust routes, etc.

[0050] Regularly monitor the implementation results, dynamically adjust the scheduling strategy according to actual conditions, and ensure continuous optimization of service quality.

[0051] Morning rush hour (7:00-9:00 am) Monday to Friday: Automatically adjust public transport operation plans: Increase frequency: During the morning rush hour, increase the frequency of trains between downtown Station A and Station B from every 5 minutes to every 3 minutes.

[0052] Prioritize services for special groups: Provide special service lanes or priority boarding opportunities for the elderly and people with disabilities to ensure fairness and efficiency.

[0053] Dynamically adjust routes: If you find that certain secondary stops have low passenger flow, you can consider temporarily canceling stops at these stops to speed up travel on the main route.

[0054] During weekends (especially when there are big events): Add dedicated line service: If the forecast shows that a large event is going to take place in a particular area, resulting in a surge in passenger traffic in that area, dedicated buses can be arranged in advance to run directly between the event location and major transportation hubs.

[0055] Extending operating hours: Considering that the event ends late, the operating hours of relevant lines can be appropriately extended to ensure that all passengers can return home safely.

[0056] The beneficial effects of this embodiment are as follows: by integrating multi-source data and applying LSTM and GARCH models, accurate prediction of passenger flow is achieved, and peak and valley periods are effectively identified. Based on this, the system can dynamically generate and execute the best scheduling plan, such as adjusting flights, optimizing routes, etc., which significantly improves operational efficiency and service quality. In addition, priority is given to the needs of special groups, ensuring fairness and humanity of services, and ultimately enhancing passenger satisfaction and the overall attractiveness of public transportation.

[0057] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

[0058] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A passenger travel data processing system based on a large model, characterized in that: The system includes a data standardization module, a model analysis module, a risk assessment module, a scheduling suggestion optimization module and a scheduling execution monitoring module; Data standardization module: used to collect real-time passenger travel data from multiple data sources, and use entropy weight method and time decay weight to fuse and standardize the data; Model analysis module: used to organize the integrated passenger travel data into a time series format, input it into the LSTM network for training and prediction, and output the passenger flow prediction results for each time period in the next week; Analyze forecast residuals to assess forecast uncertainty and stability; Risk assessment module: used to calculate the risk value and confidence interval for each forecast error variance; Scheduling suggestion optimization module: used to construct a scheduling suggestion optimization function based on the predicted passenger status to obtain the best scheduling plan; adjust the scheduling plan considering special needs; Scheduling execution monitoring module: used to automatically or manually adjust the operation plan of public transportation according to the generated optimal scheduling suggestions; regularly monitor the implementation effect and dynamically adjust the scheduling strategy according to the actual situation.

2. A passenger travel data processing method based on a large model, used to implement a passenger travel data processing system based on a large model as claimed in claim 1, characterized in that: The method flow is as follows: Step 1: Data collection and standardization, including collecting real-time data from multiple data sources, using API interfaces or direct database access to obtain data, and using entropy weight method and time decay weight to fuse and standardize the data; Step 2: Use the LSTM network combined with the improved GARCH model to analyze the fused data, predict the passenger flow in each time period in the next week, and identify potential peak and valley periods based on the prediction results; Step 3: Calculate the risk value and confidence interval for each forecast error variance, and provide a detailed report containing the risk value, forecast error variance and confidence interval; Step 4: Construct a scheduling suggestion optimization function based on the predicted passenger status to obtain the best scheduling plan, and adjust the scheduling plan considering special needs; Step 5: Execute the scheduling plan based on the generated optimal scheduling suggestion, including adjusting the operation plan of public transportation tools, increasing the frequency of trips, adjusting routes, etc., and regularly monitor the implementation effect, and dynamically adjust the scheduling strategy according to the actual situation.

3. The passenger travel data processing method based on a large model according to claim 2 is characterized in that: The data collection in step 1 includes collecting real-time data from bus companies, subway systems, transportation management departments, weather forecast services, and social media, including vehicle locations, number of passengers, route information, weather conditions, and special events.

4. The passenger travel data processing method based on a large model according to claim 2 is characterized in that: The data standardization process in step 1 uses the entropy weight method to calculate the amount of information of each feature in each data source to determine its importance, and introduces the time decay weight to adjust the importance of the feature. The fused data is calculated by a specific formula, as follows: The entropy weight method is used to calculate the information content of each feature in each data source. , to determine its importance: ； in, It is The feature in The probability distribution of samples; is the sample size; Introducing time decay weights To adjust the importance of features: ； in, is the difference between the data collection time and the current time; is the decay constant; is the number of features; The fused data Calculated by the following formula: ； in, It is Features It is The variance of the features; is the adjustment factor.

5. The passenger travel data processing method based on a large model according to claim 2 is characterized in that: In the second step, the fused passenger data is trained and predicted using a long short-term memory network, and the prediction residuals are analyzed in combination with an improved generalized autoregressive conditional heteroskedasticity model to evaluate the prediction uncertainty and stability. The formula is as follows: ； in, It's time The forecast error variance of is a constant term describing the baseline fluctuation; is the influence coefficient of the previous period’s forecast error; It's in time The prediction error of is the coefficient of the variance of the previous period’s forecast error; It is the adjustment parameter of the influence degree of nonlinear effect; is the nonlinear coefficient; is the index of the periodic term; is the number of periodic terms; is the frequency parameter related to the period; Based on the forecast results, potential peak and off-peak periods can be identified to provide a basis for subsequent scheduling.

6. The passenger travel data processing method based on a large model according to claim 2 is characterized in that: The risk assessment in step three includes calculating the risk value and confidence interval for each forecast error variance to identify periods of greater uncertainty and provide a detailed report to decision makers or system administrators.

7. The passenger travel data processing method based on a large model according to claim 2 is characterized in that: The scheduling suggestion optimization function in step 4 considers the total scheduling cost, the basic cost of the plan, the weight factor of the time deviation penalty intensity, the importance weight of the passenger, the actual arrival time and the expected arrival time of the passenger, and adjusts the scheduling plan considering special needs; Among them, the scheduling suggestion optimization function is as follows: ； in, It's in time The total dispatch cost of is a set of scheduling schemes An element in It is a plan Basic costs; is the weight factor that adjusts the intensity of the time deviation penalty; Passenger The importance weight of It is in the plan Passengers The actual arrival time; Passenger expected arrival time; Special needs include the elderly and the disabled.

8. The passenger travel data processing method based on a large model according to claim 2 is characterized in that: The execution of the optimal scheduling plan in step five includes automatically or manually adjusting the operation plan of public transportation vehicles, increasing the frequency of trips, adjusting routes, and dynamically adjusting the scheduling strategy according to actual conditions to continuously optimize service quality.

9. The passenger travel data processing method based on a large model according to claim 2 is characterized in that: During the morning rush hour from Monday to Friday, public transportation operation plans are automatically adjusted, including increasing frequency, giving priority to special groups, and dynamically adjusting routes; during weekends or specific event days, special line services are added and operating hours are extended.

Images