A fast connection adaptive scheduling system based on travel characteristics and response to heterogeneous transfer demands

Abstract

The present invention discloses a fast connection adaptive dispatching system based on the response of heterogeneous transfer demand based on travel characteristics, which relates to the technical field of traffic dispatching. The system includes a historical data processing unit, a prediction and analysis unit, a real-time monitoring unit and a response dispatching unit; the historical data processing unit is used to obtain historical transfer passenger flow data and extract the corresponding travel feature set; the prediction and analysis unit is used to train the demand prediction and analysis model using the travel feature set, and predict the transfer demand using the trained model; the real-time monitoring unit is used to collect the location, speed and passenger capacity information of public transportation in real time and monitor the passenger flow of transfer stations; the response dispatching unit is used to use genetic algorithms to perform dispatching simulation based on transfer demand and collected real-time data, and perform multi-mode traffic dispatching according to the optimal dispatching strategy. The present invention improves the efficiency and quality of traffic services and enhances the travel experience of passengers.

Description

Technical Field

[0001] The present invention relates to the technical field of traffic dispatching, and in particular to a fast connection adaptive dispatching system based on response to heterogeneous transfer demands of travel characteristics. Background Art

[0002] In recent years, the passenger collection and dispatching methods of domestic integrated transportation hubs have become increasingly diversified, with private cars and taxis being the main means of passenger transfer, supplemented by public transportation such as subways, light rail, and buses. The diverse passenger collection and dispatching methods often make integrated transportation hubs prone to passenger congestion and tight transportation capacity resources in daily operations. On the one hand, it will cause congestion inside and near the hub, and on the other hand, it will reduce the travel quality of passengers and even cause chaos. The reason for this phenomenon is the unreasonable allocation of capacity resources, especially during non-peak hours, the use of related facilities and the dispatch of capacity do not correspond to the needs of passengers, resulting in uneven capacity of each transfer mode, and some facilities and capacity resources are idle or underutilized.

[0003] The existing public transportation scheduling methods are unable to timely and accurately grasp the evolution characteristics of hub passenger flow in the temporal and spatial scales, and cannot make real-time adjustments and responses according to passenger demand. This may lead to low efficiency in public transportation scheduling, low data integration, and weak coordination and linkage capabilities between hubs. Summary of the invention

[0004] In view of this, the present application provides a fast-connection adaptive scheduling system based on travel characteristics and heterogeneous transfer demand response, which solves the technical problems of being unable to timely and accurately grasp the evolution characteristics of hub passenger flow in the temporal and spatial scales, and being unable to make real-time adjustments and responses according to passenger needs, which may lead to low efficiency in public transportation scheduling, low data integration, and weak coordination and linkage capabilities between hubs by utilizing multiple data sources such as passenger travel characteristics, historical passenger flow characteristics, and weather effect characteristics, as well as demand response methods.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] The present invention proposes a fast connection adaptive scheduling system based on travel characteristics heterogeneous transfer demand response, including: a historical data processing unit, a prediction analysis unit, a real-time monitoring unit and a response scheduling unit;

[0007] The historical data processing unit is used to obtain historical transfer passenger flow data in the integrated transportation hub, and extract corresponding travel feature groups from the historical transfer passenger flow data;

[0008] The prediction and analysis unit is used to construct and train a demand prediction and analysis model using the travel feature group, and use the trained demand prediction and analysis model to predict the transfer demand of each transportation mode in the future;

[0009] The real-time monitoring unit is used to collect the location, speed, and passenger capacity information of public transportation in real time and monitor the passenger flow information of the transfer station;

[0010] The response scheduling unit is used to perform scheduling simulation using a genetic algorithm based on the transfer demand predicted by the prediction and analysis unit and the real-time data collected by the real-time monitoring unit, calculate the optimal scheduling strategy, and perform multi-mode traffic scheduling according to the optimal scheduling strategy.

[0011] Furthermore, the historical data processing unit includes a historical data acquisition module and a data set construction module; the historical data acquisition module is used to acquire the historical transfer passenger flow data, and clean the data to remove duplicate, erroneous or invalid data to obtain the cleaned historical transfer passenger flow data; the data set construction module is connected to the historical data acquisition module, and the data set construction module is used to extract the travel characteristics of each passenger according to the cleaned historical transfer passenger flow data.

[0012] Furthermore, the data set construction module includes a group extraction module and a fusion module; the group extraction module is used to extract travel feature groups from the cleaned historical transfer passenger flow data; the extraction of travel feature groups includes: extracting each passenger's travel chain from the cleaned historical passenger flow, the travel chain is composed of multiple groups of entry and exit data, the entry and exit data include entry and exit stations, time and traffic mode, analyzing the transfer relationship in the travel chain according to the time interval between entry and exit, and the continuous entry and exit data belonging to the transfer relationship in each passenger's travel chain is combined with the passenger's personal label features as a group of travel feature groups; the fusion module is used to fuse the features of each travel feature group to obtain a target travel feature group.

[0013] Furthermore, the prediction and analysis unit includes a model construction and training module and a demand prediction module; the model construction and training module is used to construct a demand prediction and analysis model based on the Logit model, and use the target travel feature group to train the demand prediction and analysis model, optimize model parameters, and obtain a trained demand prediction and analysis model; the demand prediction module is used to use the trained demand prediction and analysis model to perform transfer demand prediction, and the transfer demand prediction includes: predicting the transfer volume of each station, drawing a transfer volume distribution map, thereby identifying popular transfer stations, and analyzing the passenger flow characteristics of popular transfer stations, the passenger flow characteristics include transfer type and transfer direction, and obtaining the spatiotemporal distribution characteristics of transfer demand.

[0014] Furthermore, the real-time monitoring unit includes a data acquisition module and a statistical module; the data acquisition module is used to obtain the real-time position, speed and passenger capacity information of public transportation through the GPS module, vehicle-mounted sensors and intelligent transportation network, and split it into time domain traffic data sets, and is also used to obtain real-time passenger flow data of transfer stations; the statistical module is used to count the passenger flow data of transfer stations, and according to the passenger flow data of transfer stations, count the number of passengers arriving, leaving and waiting at each station, and generate a real-time data set based on the statistical data.

[0015] Furthermore, the response scheduling unit includes a response calculation module and an adaptive scheduling module; the response calculation module is used to establish an objective function that minimizes the comprehensive passenger transfer time, passenger waiting time and public transportation vehicle departure cost, and uses a genetic algorithm to calculate the optimal scheduling strategy in real time based on the spatiotemporal distribution characteristics of the transfer demand predicted by the demand prediction and analysis model and the real-time data collected by the real-time monitoring unit; the adaptive scheduling module is used to adjust the operation plan of the public transportation vehicles in the hub according to the optimal scheduling strategy.

[0016] Furthermore, the real-time calculation of the optimal scheduling strategy using the genetic algorithm includes: defining the encoding of each solution in the genetic algorithm as In the form of, n is the number of bus departures per hour on the bus line to be scheduled, is the departure time, a group of departure times is an individual in the population, and each departure time is the chromosome of the individual; a random initial solution is generated, and the fitness function of the chromosome is calculated; and the optimal solution is obtained by iterative optimization, which is a scheduling strategy that minimizes the comprehensive cost of passenger transfer time, passenger waiting time and public transportation vehicle departure.

[0017] It can be seen from the above technical solution that the advantages of the present invention are:

[0018] 1. The present invention predicts the transfer passenger flow through a transfer demand prediction model, and utilizes multiple data sources such as passenger travel characteristics, historical passenger flow characteristics, and weather effect characteristics, as well as demand response methods, to monitor the overall operation status of the hub in real time, predict the number of passengers expected to arrive and depart at each hub connection mode, the share rate of various transportation modes, and the proportion of passenger flow characteristics, etc., and adaptively dispatches public vehicles based on the predicted data, thereby improving transfer efficiency.

[0019] 2. Through the collaborative work of multiple modules and technologies, efficient public transportation scheduling is achieved, which can improve the efficiency and quality of public transportation, reduce passengers' waiting time and number of transfers, and enhance passengers' travel experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The accompanying drawings, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0021] Figure 1 It is a structural schematic diagram of the fast connection adaptive scheduling system based on travel characteristics and heterogeneous transfer demand response of the present invention.

[0022] Figure 2 It is a structural schematic diagram of the historical data processing unit of the present invention.

[0023] Figure 3 It is a schematic diagram of the transfer relationship of the present invention.

[0024] Figure 4 This is a data feature extraction flow chart of the present invention. DETAILED DESCRIPTION

[0025] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments and the accompanying drawings. Here, the illustrative embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0026] The diverse passenger gathering and dispersing methods often make comprehensive transportation hubs prone to passenger congestion and tight transportation capacity resources in daily operation. On the one hand, it will cause congestion inside and near the hub, and on the other hand, it will reduce the quality of passengers' travel and even cause chaos. The reason for this phenomenon is the unreasonable allocation of capacity resources, especially during non-peak hours. The use of related facilities and the dispatch of capacity do not correspond to the needs of passengers, resulting in uneven carrying capacity of various land-side connection methods, and some facilities and capacity resources are idle or underutilized.

[0027] like Figure 1 and Figure 2 As shown, this embodiment provides a fast connection adaptive scheduling system based on travel characteristics and heterogeneous transfer demand response, which realizes efficient public transportation scheduling through the collaborative work of multiple modules and technologies and improves the travel experience of passengers. The adaptive scheduling system specifically includes: a historical data processing unit, a prediction and analysis unit, a real-time monitoring unit and a response scheduling unit.

[0028] The historical data processing unit is used to obtain historical transfer passenger flow data, and is also used to extract corresponding travel feature groups from the historical transfer passenger flow data.

[0029] like Figure 2As shown, the historical data processing unit includes a historical data acquisition module and a data set construction module; the historical data acquisition module is used to acquire historical transfer passenger flow data, clean the data, remove duplicate, erroneous or invalid data, and update the transfer data; the data set construction module is connected to the historical data acquisition module, and the data set construction module is used to extract the travel characteristics of each passenger based on the historical transfer passenger flow data.

[0030] Since there may be error factors such as equipment failure, signal interference and human error in the data collection process, and intelligent collection equipment such as GPS receivers will also be affected by urban high-rise buildings, tunnels and meteorological conditions, there will be duplicate and abnormal records in user travel data. Therefore, it is necessary to clean the data before analysis. Data cleaning is an important step in data preprocessing, which involves checking data consistency, processing invalid values ​​and missing values, etc. Data cleaning includes: deleting missing values ​​or filling them with methods such as average and median; deleting or correcting outliers. And randomly select a certain number of samples from the data source for verification to ensure the accuracy and effectiveness of the cleaning rules. And perform analysis and verification iteratively many times until a satisfactory cleaning result is obtained to ensure the consistency of the data format. For example, convert date and time data into a unified format, convert text data into an appropriate encoding format, etc.

[0031] In this embodiment, possible data sources include operation data of the public transportation system, passenger travel surveys, mobile payment records, GPS positioning data, etc., and historical transfer passenger flow data is collected from the selected data sources. The historical transfer passenger flow data includes the passengers' departure place, destination, transfer station, transfer time, etc.

[0032] Specifically, the data set construction module includes a grouping extraction module and a fusion module; the grouping extraction module is used to extract the starting point, destination, number of transfers, transfer station, transfer time and personal label features of each passenger, and analyze the extracted feature data with the influencing factors, and use the time series analysis method to analyze the changing trends and periodic laws of historical transfer data, and generate each travel feature group; the fusion module is connected to the grouping module, and the fusion module is used to fuse the features of each travel feature group to obtain the target travel feature group.

[0033] In this embodiment, when collecting data, the time difference T between the transaction time of the subway exit and entry card swiping record and the transaction time of the bus boarding card swiping record of the same payment account passenger is used to determine whether there is a relationship between the subway and the bus or the bus and the subway. Considering the rest, walking and waiting time that passengers may have during the transfer process, 25 minutes is selected as the transfer time threshold Tb for the subway to the bus. The relationship between the transfer time threshold and the determination of the transfer relationship is as follows: Figure 3As shown, when the interval between two consecutive trips of a passenger is less than the transfer time threshold, it is considered that there is a transfer relationship.

[0034] like Figure 4 As shown in Figure 1, data feature extraction includes three parts: data preparation, travel chain extraction and screening. (1) Data preparation: According to the transfer time threshold Tb, the transfer record information of passengers transferring from subway to bus or from bus to subway can be extracted from the card swiping data of passengers. Specifically, after cleaning the original data to obtain the passenger card swiping record data containing valid information fields, the card swiping record data is grouped by the card number field and sorted in ascending order according to the transaction time. The valid information fields include card number, transaction time, transaction type (subway or bus), entry and exit stations, etc.; (2) Travel chain extraction: traverse the card swiping record of each passenger, extract the travel chain in chronological order, extract the travel chain of passengers with the same card number, and check (3) Filtering: Filter out the travel chains that include both subway and bus from all travel chains. This is the travel chain information that may include transfer relationships. Calculate the difference in rail transit interval time in the travel chain. If it is less than the transfer time threshold, it is determined that there is a transfer relationship. The continuous rail transit entry and exit and bus boarding data records in the travel chain that includes the transfer relationship are filtered and stored as a group of transfer information. Collect and store the personal label features of the corresponding passengers based on passenger travel surveys and smart transportation methods. Repeat the above steps for all passengers' travel chains until all passengers' data are processed, transfer information extraction is completed, and each travel feature group is obtained.

[0035] In this embodiment, the time series analysis method includes trend analysis, seasonal analysis, etc. The time series analysis method can intuitively understand the changing trend and periodicity of historical transfer data. Provide a data basis for vehicle scheduling. Influencing factors include time, weather, holidays, etc. Personal tags include travel feature heterogeneity, which usually refers to the various feature differences shown by different passengers during the travel process. This difference can greatly affect the decision-making of passengers, such as which mode of transportation to choose, when to depart, which route to choose, etc. Travel feature heterogeneity specifically includes the following aspects: Personal characteristics: such as age, gender, occupation, income, family status, etc. These characteristics will affect passengers' travel needs, travel mode selection, etc. Travel purpose: such as work, school, shopping, entertainment, etc. Different travel purposes will affect passengers' travel time, travel distance, etc. Travel mode: such as walking, bicycle, public transportation, private car, etc. Different travel modes will affect passengers' travel costs, travel time, etc. Departure point and arrival point: Different departure points and arrival points will affect passengers' travel distance, travel time, etc. Departure time and arrival time: Different departure times and arrival times will affect passengers’ travel time cost, travel comfort, etc. Travel distance: Different travel distances will affect passengers’ travel mode choice, travel time, etc. Travel time: Different travel time will affect passengers’ travel comfort, travel cost, etc.

[0036] The feature fusion can be performed in a variety of ways, including but not limited to direct combination, mathematical transformation, machine learning methods, etc., to combine various features to create new features that can more comprehensively describe the data.

[0037] like Figure 1 As shown, the prediction and analysis unit is used to train the initial analysis model into a demand prediction and analysis model using the travel feature set, and use the demand prediction and analysis model to predict future transfer demand and identify the spatiotemporal distribution characteristics of the transfer demand.

[0038] The prediction and analysis unit includes a model building and training module and a demand prediction module; the model building and training module is used to build a demand prediction and analysis model based on the Logit model, and use the target travel feature group to train the demand prediction and analysis model, adjust the model parameters, and obtain a trained demand prediction and analysis model.

[0039] The demand forecasting module is used to forecast transfer demand based on the forecasted demand and input the forecast influencing factors into the trained demand forecasting and analysis model, conduct time series analysis on the transfer demand, count the transfer volume and passenger flow direction in each time period, count the transfer volume of each station based on the transfer information, draw a transfer volume distribution map, and thus identify popular transfer stations. The module also analyzes the passenger flow characteristics of popular transfer stations, including the transfer type and transfer direction, and obtains the spatiotemporal distribution characteristics of the transfer demand.

[0040] In this embodiment, the transfer types include subway to subway, subway to bus, etc. The spatiotemporal distribution characteristics include peak hours and popular transfer stations.

[0041] like Figure 3 As shown in the figure, the demand forecasting and analysis model based on the Logit model includes:

[0042] Step S1: Assume that the choice of transportation mode by passengers conforms to the normal distribution. Assume that for a transportation hub, the utility function of the passenger choosing the i-th transportation mode is U i , the probability that a passenger chooses the i-th mode of transportation is P(i);

[0043] Step S2: According to the factors affecting passengers' transfer, the Logit model includes three utility parts: F i represents the characteristic function of the ith mode of transportation, are independent and identically distributed random variables, that is, Changes with the passenger's personal behavior selection characteristics and personal label characteristics, is the environmental influencing factor, which includes weather factors and holiday factors. The characteristic function of the passenger choosing the i-th mode of transportation is: The utility function of the i-th mode of transportation is: in, represents the average waiting time for the i-th mode of transportation, represents the time of taking the i-th mode of transportation, represents the transfer time to the i-th mode of transportation, a represents the fare of the i-th mode of transportation, They are and the weight of a;

[0044] Step S3: The probability of a passenger choosing the i-th mode of transportation n is the total number of transportation modes, then the passenger flow transfer volume of different transportation modes at the hub is: i =Y*P(i), where B i is the transfer volume of the i-th mode of transportation in the hub, Y is the predicted total hub passenger traffic, and P(i) is the probability of choosing the i-th mode of transportation.

[0045] The total hub passenger traffic forecast is predicted using the multivariate regression method. Among them, m is the number of cells divided into the prediction area, B d is the transportation volume of the dth traffic zone, B d =a0+∑a k x dk , the variable value of the dth traffic zone is x dk , the regression coefficient of the kth variable is a k .

[0046] The main purpose of the urban comprehensive passenger transport hub is to organize the transfer passenger flow efficiently, quickly and safely. By analyzing and predicting the passenger flow distribution volume of the passenger hub and the passenger transfer volume between various modes of transportation, the adaptive scheduling and rapid connection of public transportation can be realized, so as to achieve the purpose of safe and rapid transfer. The demand prediction and analysis model for transfer in this embodiment adopts the Logit model, which can better reflect the diversity of personal characteristics and travel modes, and more effectively analyze traffic activities. The share rate of each passenger mode in the hub is predicted by the non-aggregate method. Among them, there are many factors that affect the utility value of passenger transfer, including economy, convenience, rapidity, comfort and safety. The model parameters and influencing factors can be trained and optimized according to the travel characteristics extracted from the historical transfer data. The environmental influencing factors include weather influencing factors, holiday influencing factors, etc. For example, for the holiday influencing factors, the number 0 represents working days, 1 represents weekends and national statutory holidays, and for the weather influencing factors, sunny days, rainy days and cloudy days are represented by the numbers 1, 0 and 0.5 respectively. The influencing factors under different states are calculated according to methods such as weighted summation.

[0047] The real-time monitoring unit is used to collect the location, speed and passenger capacity information of public transportation in real time and split it into time-domain traffic data sets. It is also used to monitor the passenger flow information at transfer stations.

[0048] like Figure 1 As shown, the real-time monitoring unit includes a data acquisition module and a statistical module; the data acquisition module is used to obtain the real-time position, speed and passenger capacity information of public transportation through the GPS module, vehicle-mounted sensors and intelligent transportation networks, and split it into time domain traffic data sets according to the time corresponding to the information, and is also used to obtain the passenger flow data of the transfer station; the statistical module is used to count the traffic data of the transfer station, and count the number of arriving, leaving and waiting passengers according to the passenger flow data of the transfer station to generate a real-time data set.

[0049] Obtain the departure time and route of public transportation through real-time data sets to adapt to real-time changing transfer needs.

[0050] The response dispatch unit includes a response calculation module and an adaptive dispatch module; the response calculation module is used to establish an objective function that minimizes the comprehensive passenger transfer time, regular passenger waiting time, and bus company departure cost. The solution of the objective function is a set of departure times with unequal intervals. The solution is obtained by a genetic algorithm, and the optimal dispatch strategy is calculated in real time based on the solution results; the adaptive dispatch module is used to automatically adjust the operation plan of public transportation vehicles based on the results of the response calculation module.

[0051] The optimal scheduling strategy using genetic algorithms in real time includes:

[0052] Step A: Encode the model. Suppose a bus line has a departure frequency of n per hour, and the corresponding set of sequential departure times are Then the encoding of each solution in the model is In the form of, a set of departure times is an individual in the population, and each departure time is the chromosome of an individual;

[0053] Step B: Generate the initial population. The public vehicles on the route are dispatched one by one, and random initial solutions can be generated one by one within the preset optimization period. Calculate the fitness function of the chromosome (the fitness function of this embodiment is the weighted sum of passenger transfer time, passenger waiting time and public transportation vehicle departure cost. The purpose of iterative calculation is to obtain the solution with the minimum weighted sum as the optimal solution);

[0054] Step C: Perform multiple iterations of selection, crossover, and mutation operations to obtain the optimal solution set.

[0055] The response dispatch unit is used to perform dispatch simulation based on transfer demand and collected real-time data using genetic algorithms, calculate the optimal dispatch strategy, and perform multi-modal traffic dispatch according to the optimal dispatch strategy.

[0056] The response dispatch unit includes a response calculation module and an adaptive dispatch module; the response calculation module is used to calculate the optimal dispatch strategy in real time using a genetic algorithm; the adaptive dispatch module is used to automatically adjust the operation plan of the public transportation vehicle according to the result of the response calculation module.

[0057] In this embodiment, the operation plan includes the departure time, route, capacity allocation, etc., to adapt to the real-time changing transfer demand. The dispatching strategy includes the departure time, driving route, stop points, etc. of the vehicle.

[0058] When performing adaptive dispatching, it can communicate with public buses in real time, send dispatching instructions and receive feedback information. It can control the driving status of public buses, such as adjusting speed and changing driving routes. In the passenger information service module, it can provide passengers with real-time bus information, such as vehicle location and estimated arrival time, through mobile phone apps and electronic bus stops. It can provide transfer suggestions at transfer stations to help passengers plan travel routes more efficiently. It also provides a user interface so that dispatchers can easily view real-time data, analyze forecast results and generate dispatch plans.

[0059] In another embodiment, a function that supports manual adjustment of the scheduling plan may be added to deal with emergencies or special circumstances.

[0060] In view of the complex and changeable characteristics of passenger flow and traffic flow in urban transportation hubs, insufficient organizational and scheduling capabilities, and low passenger travel efficiency, the present invention constructs a multi-modal transportation efficient collaborative organization and adaptive scheduling system to achieve real-time and accurate prediction of hub passenger flow and efficient scheduling of multi-modal transportation, improve the efficiency of passenger transfers between hubs, provide strong support for the hub's operational scheduling plans, and enhance the passenger experience.

[0061] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the embodiments of the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A fast connection adaptive scheduling system based on the response to heterogeneous transfer demand of travel characteristics, characterized by: include: Historical data processing unit, prediction analysis unit, real-time monitoring unit and response scheduling unit; The historical data processing unit is used to obtain historical transfer passenger flow data in the integrated transportation hub, and extract the corresponding target travel feature group from the historical transfer passenger flow data; The prediction and analysis unit is used to construct and train a demand prediction and analysis model using the target travel feature group, and use the trained demand prediction and analysis model to predict the transfer demand of each transportation mode in the future; The prediction and analysis unit includes a model building and training module and a demand prediction module; the model building and training module is used to build a demand prediction and analysis model according to the Logit model, and train the demand prediction and analysis model using the target travel feature group, optimize the model parameters, and obtain a trained demand prediction and analysis model; The demand prediction module is used to use the trained demand prediction and analysis model to predict the transfer demand. The transfer demand prediction includes: predicting the transfer volume of each station, drawing a transfer volume distribution map, thereby identifying popular transfer stations, and analyzing the passenger flow characteristics of popular transfer stations. The passenger flow characteristics include transfer type and transfer direction, and obtaining the spatiotemporal distribution characteristics of transfer demand; The real-time monitoring unit is used to collect the location, speed, and passenger capacity information of public transportation in real time and monitor the passenger flow information of the transfer station; The response dispatch unit is used to perform dispatch simulation using a genetic algorithm based on the transfer demand predicted by the prediction and analysis unit and the real-time data collected by the real-time monitoring unit, calculate the optimal dispatch strategy, and perform multi-mode traffic dispatch according to the optimal dispatch strategy; The response scheduling unit includes a response calculation module and an adaptive scheduling module; the response calculation module is used to establish an objective function that minimizes the comprehensive passenger transfer time, passenger waiting time and public transportation vehicle departure cost, and uses a genetic algorithm to calculate the optimal scheduling strategy in real time based on the spatiotemporal distribution characteristics of the transfer demand predicted by the demand prediction and analysis model and the real-time data collected by the real-time monitoring unit; the adaptive scheduling module is used to adjust the operation plan of the public transportation vehicles in the hub according to the optimal scheduling strategy; The predicted transfer volume of each station includes: Assume that passengers' choice of transportation mode follows a normal distribution; Calculate the utility function U of the i-th mode of transportation i : Among them, F i represents the utility of the i-th mode of transportation, are independent and identically distributed random variables, are environmental factors, including weather factors and holiday factors. represents the average waiting time, Indicates the time of riding. represents the transfer time, a represents the fare, They are and the weight of a; Predict the total passenger volume Y of the hub and calculate the transfer demand predicted for the i-th mode of transportation at the hub: B i =Y*P(i) Among them, B i is the predicted transfer demand of the i-th mode of transportation in the hub, Y is the predicted total passenger volume of the hub, P(i) is the probability of selecting the i-th mode of transportation, and n is the number of transportation modes in the hub; The total hub passenger traffic forecast is predicted using the multivariate regression method: Among them, m is the number of small areas divided into the prediction area, B d is the transportation volume of the dth small area; The real-time calculation of the optimal scheduling strategy using a genetic algorithm includes: Define the encoding of each solution in the genetic algorithm as In the form of, n is the number of bus departures per hour on the bus line to be scheduled, is the departure time, a group of departure times is an individual in the population, and each departure time is the chromosome of an individual; The optimal solution is obtained through iterative optimization, which is a scheduling strategy that minimizes the comprehensive costs of passenger transfer time, passenger waiting time, and public transportation vehicle dispatching.

2. The fast connection adaptive scheduling system based on the response to heterogeneous transfer demand of travel characteristics according to claim 1 is characterized in that: The historical data processing unit includes a historical data acquisition module and a data set construction module; The historical data acquisition module is used to acquire the historical transfer passenger flow data, and clean the data to remove duplicate, erroneous or invalid data to obtain cleaned historical transfer passenger flow data; The data set construction module is connected to the historical data acquisition module, and the data set construction module is used to extract the travel characteristics of each passenger according to the cleaned historical transfer passenger flow data.

3. The fast connection adaptive scheduling system based on the response to heterogeneous transfer demands of travel characteristics according to claim 2 is characterized in that: The data set construction module includes a grouping extraction module and a fusion module; The grouping extraction module is used to extract travel feature groups from the cleaned historical transfer passenger flow data; The extracted travel feature group comprises: Extract each passenger's travel chain from the cleaned historical passenger flow. The travel chain consists of multiple groups of entry and exit data, including entry and exit stations, time and traffic mode. Analyze the transfer relationship in the travel chain according to the time interval of entry and exit. Combine the continuous entry and exit data of each passenger's travel chain that belongs to the transfer relationship with the personal label characteristics of the passenger as a group of travel feature groups. The fusion module is used to fuse the features of each travel feature group to obtain a target travel feature group.

4. The fast connection adaptive scheduling system based on the response to heterogeneous transfer demands of travel characteristics according to claim 1 is characterized in that: The real-time monitoring unit includes a data acquisition module and a statistics module; The data acquisition module is used to obtain the real-time location, speed and passenger capacity information of public transportation through the GPS module, vehicle-mounted sensors and intelligent transportation network, and split it into time-domain traffic data sets, and is also used to obtain real-time passenger flow data at transfer stations; The statistical module is used to count the passenger flow data of the transfer station, and count the number of passengers arriving, leaving and waiting at each station according to the passenger flow data of the transfer station, and generate a real-time data set based on the statistical data.

Images