Artificial Intelligence-Based Rail Transit Dispatching and Analysis Methods and Systems

Abstract

This invention provides an artificial intelligence-based rail transit scheduling analysis method and system, belonging to the field of rail transit scheduling technology. First, a dynamic scheduling baseline for the rail transit system is established, including baseline operating characteristics, rule constraint characteristics, and scenario association characteristics. Next, real-time scheduling data is collected and compared with the dynamic scheduling baseline in multiple dimensions to generate a baseline deviation feature set. Then, a pre-trained scheduling collaborative evaluation model is invoked to perform dynamic correlation analysis on the baseline deviation feature set, generating a scheduling collaborative evaluation result. Based on the scheduling collaborative evaluation result, a multi-objective scheduling optimization model is constructed for collaborative optimization processing, generating an initial scheduling adjustment scheme set. Finally, the initial scheduling adjustment scheme set undergoes dynamic constraint verification processing to generate the final rail transit scheduling instruction, which triggers the scheduling control system to perform scheduling parameter update operations, thereby improving the efficiency, safety, and stability of the rail transit system operation.

Description

Technical Field

[0001] This invention relates to the field of rail transit scheduling technology, and more specifically, to a rail transit scheduling analysis method and system based on artificial intelligence. Background Technology

[0002] In the field of rail transit dispatching, traditional dispatching analysis methods have many limitations. On the one hand, some methods rely mainly on historical dispatching data and preset fixed rules for dispatching decisions. However, historical data can only reflect past operating conditions and cannot fully consider real-time changing factors, such as sudden passenger flow or equipment failure. Moreover, preset fixed rules are difficult to adapt to complex and ever-changing operating scenarios. When encountering special circumstances, these rules may not be able to provide effective dispatching solutions, resulting in a lack of flexibility and adaptability in dispatching decisions.

[0003] On the other hand, existing scheduling analysis methods often focus only on single-dimensional scheduling information, such as considering only the time or space dimensions, ignoring the correlation and mutual influence between different dimensions. Rail transit systems are complex integrated systems, with multiple dimensions such as time, space, and resources closely related. Single-dimensional analysis cannot comprehensively and accurately grasp the scheduling situation, easily leading to biased scheduling decisions and affecting the normal operation of the entire rail transit system. Summary of the Invention

[0004] In view of the aforementioned problems, and in conjunction with the first aspect of the present invention, embodiments of the present invention provide an artificial intelligence-based rail transit scheduling and analysis method, the method comprising:

[0005] A dynamic scheduling baseline for the rail transit system is established, which includes baseline operation characteristics constructed from historical normal scheduling data, rule constraint characteristics of a preset scheduling rule base, and scenario association characteristics of typical operating scenarios.

[0006] Real-time scheduling data of the rail transit system is collected, and the real-time scheduling data is compared and analyzed with the dynamic scheduling baseline in multiple dimensions to generate a baseline deviation feature set, which includes time dimension deviation features, spatial dimension deviation features and resource dimension deviation features.

[0007] The pre-trained scheduling coordination evaluation model is invoked to perform dynamic correlation analysis on the baseline deviation feature set, and a scheduling coordination evaluation result is generated. The scheduling coordination evaluation result includes deviation propagation path parameters, coordination conflict probability parameters, and resource adaptability parameters.

[0008] Based on the scheduling coordination evaluation results, a multi-objective scheduling optimization model is constructed. The deviation propagation path parameters, coordination conflict probability parameters, and resource adaptability parameters are coordinated and optimized through the multi-objective scheduling optimization model to generate an initial set of scheduling adjustment schemes.

[0009] The initial set of scheduling adjustment schemes is subjected to dynamic constraint verification. After verification, a final rail transit scheduling instruction is generated, which is used to trigger the rail transit scheduling control system to perform scheduling parameter update operations.

[0010] In another aspect, embodiments of the present invention also provide an artificial intelligence-based rail transit scheduling and analysis system, including a processor and a machine-readable storage medium connected to the processor. The machine-readable storage medium is used to store programs, instructions, or code, and the processor is used to execute the programs, instructions, or code in the machine-readable storage medium to implement the above-described method.

[0011] Based on the above, this invention establishes a dynamic scheduling baseline by constructing a benchmark operation feature based on historical normal scheduling data, rule constraint features of a preset scheduling rule base, and scenario association features of typical operating scenarios. Real-time scheduling data is collected and compared with the dynamic scheduling baseline in multiple dimensions to generate a baseline deviation feature set containing time-dimension deviation features, spatial-dimension deviation features, and resource-dimension deviation features. This accurately captures abnormal situations during the scheduling process, providing detailed data support for subsequent analysis. A pre-trained scheduling collaborative evaluation model is invoked to perform dynamic correlation analysis on the baseline deviation feature set, generating a scheduling collaborative evaluation result containing deviation propagation path parameters, collaborative conflict probability parameters, and resource suitability parameters. This deeply analyzes the intrinsic connections and impacts between deviations. Based on the scheduling collaborative evaluation result, a multi-objective scheduling optimization model is constructed for collaborative optimization, generating an initial scheduling adjustment scheme set. This comprehensively considers multiple optimization objectives to achieve overall optimization of the scheduling scheme. Finally, the initial scheduling adjustment scheme set is dynamically constrained and verified to generate the final rail transit scheduling instruction, ensuring the feasibility and effectiveness of the scheduling scheme and effectively improving the efficiency, safety, and stability of the rail transit system. Attached Figure Description

[0012] Figure 1 This is a schematic diagram of the execution flow of the rail transit scheduling and analysis method based on artificial intelligence provided in an embodiment of the present invention.

[0013] Figure 2 This is a schematic diagram of exemplary hardware and software components of an artificial intelligence-based rail transit scheduling and analysis system provided in an embodiment of the present invention. Detailed Implementation

[0014] The present invention will now be described in detail with reference to the accompanying drawings. Figure 1 This is a flowchart illustrating an artificial intelligence-based rail transit scheduling and analysis method according to an embodiment of the present invention. The following is a detailed description of the artificial intelligence-based rail transit scheduling and analysis method.

[0015] Step S110: Establish a dynamic scheduling baseline for the rail transit system. This dynamic scheduling baseline includes baseline operation characteristics constructed from historical normal scheduling data, rule constraint characteristics of a preset scheduling rule base, and scenario association characteristics of typical operating scenarios.

[0016] In this embodiment, a city rail transit network is used as the application scenario. This network encompasses multiple intersecting lines, connecting various major areas of the city, involving dozens of stations, hundreds of trains, and a centralized dispatch control center. Establishing a dynamic dispatch baseline is the fundamental step in the entire dispatch analysis method. Its core purpose is to construct a benchmark framework that reflects the normal operating state of the system, thereby accurately identifying deviations between the real-time operation and the benchmark state.

[0017] Step S111: Collect historical normal dispatch data of the rail transit system. This historical normal dispatch data includes historical train operation data, historical line status data, historical passenger flow data, and historical dispatch instruction data.

[0018] In this urban rail transit network, each train is equipped with GPS devices, speed sensors, and passenger flow counters in each carriage, recording the train's trajectory in real time, including its position coordinates every second; the exact time of each stop at a station, accurate to the second; departure time; and continuous speed change curves during travel. These collectively constitute the core content of historical train operation data. Along the line, track condition monitoring sensors are deployed beside the tracks to detect track smoothness, wear, etc.; the signaling system itself records its operating status, such as the display status of signal lights and the response time of signal switching; the power supply system has voltage and current monitoring devices. All this data is aggregated to form historical line status data. At each station, gate counters are installed at entrances and exits to record the number of people entering and exiting the station every hour; video analysis equipment is installed inside the stations to use image recognition technology to count the number of people remaining in the station at different times. This constitutes historical passenger flow data. The dispatch control center's command system automatically archives all issued dispatch commands, including the specific time of train departure, the station and duration of stops, and the specific speed adjustment values, forming historical dispatch command data.

[0019] Step S112: Perform scenario segmentation processing on the historical normal scheduling data. Based on the characteristics of the operating period, weather conditions, and holiday attributes, the historical normal scheduling data is divided into multiple typical operating scenario data units.

[0020] After data collection, the next stage is scenario segmentation. The segmentation of operational time periods is based on the travel patterns of urban residents. The morning peak period is defined as the peak travel time for commuters on weekday mornings, the evening peak period as the peak travel time for commuters leaving work on weekday evenings, the off-peak period as normal daytime hours outside of peak hours, and the nighttime period from late evening to early morning. Weather condition characteristics are obtained by connecting with historical meteorological data from the city's meteorological department, clearly distinguishing between different weather conditions such as sunny, cloudy, rainy, and snowy days. Rainy days are further subdivided into light rain, moderate rain, and heavy rain, and snowy days are similarly subdivided based on snowfall. Holiday attribute characteristics are determined based on national statutory holidays, weekends, and regular workdays.

[0021] The scenario segmentation adopts a multi-dimensional cross-combination approach, combining different values ​​of the above three features to form multiple typical operational scenario data units. For example, the "morning peak - light rain - weekday" scenario data unit contains all historical train operation data, historical line status data, historical passenger flow data, and historical dispatch instruction data during the morning peak period on weekdays when the weather is light rain; the "evening peak - moderate snow - weekend" scenario data unit contains various corresponding data during the evening peak period on weekends when there is moderate snow.

[0022] Step S113: Extract train punctuality rate features, interval running time features, station dwell time features, and resource occupancy rate features from each typical operation scenario data unit, and combine the extracted features to generate baseline operation features.

[0023] For each typical operational scenario data unit, feature extraction is performed using specialized data analysis tools. The extraction of train punctuality rate features involves analyzing the operation of all trains at all stops within that scenario. The difference between the actual arrival time and the planned arrival time for each train at each station is calculated. When the difference is within a preset allowable range, the train is considered punctual. The proportion of punctual arrivals to the total number of arrivals is then calculated. Finally, a weighted average of the punctuality rates for all trains and stations is taken to obtain the train punctuality rate feature for that scenario. This feature is presented as a multi-dimensional array, containing the punctuality rate distribution of different trains at different stations.

[0024] Extracting the interval running time feature involves dividing the line into multiple consecutive intervals. For each interval, the travel time of all trains passing through that interval is statistically analyzed to form a time distribution sequence, including the minimum, maximum, average, and frequency of occurrence of intervals with different durations. These data together constitute the interval running time feature.

[0025] The station dwell time feature is extracted by statistically analyzing the dwell time of all trains at each station in the scenario. This results in a feature sequence containing minimum, maximum, average, and distribution frequency values, reflecting the dwell time characteristics of different stations in the scenario.

[0026] Resource occupancy rate characteristics involve line resources and station platform resources. Line resource occupancy rate is the proportion of time that the line section is occupied by trains within a unit of time to the total time. Station platform resource occupancy rate is the proportion of time that the platform is occupied by trains within a unit of time, as well as the ratio of the number of people staying in the station to the maximum capacity of the platform. These data are integrated to form resource occupancy rate characteristics.

[0027] Finally, the extracted train punctuality rate features, interval running time features, station dwell time features, and resource occupancy rate features are combined in a set order to form a high-dimensional feature vector, which is the baseline operating feature corresponding to the data unit of this typical operating scenario.

[0028] Step S114: Obtain the preset scheduling rule base, extract the line capacity constraint rules, train minimum interval rules, station passenger flow carrying capacity rules and emergency scheduling priority rules from the preset scheduling rule base, and convert the extracted rules into rule constraint features.

[0029] The pre-set scheduling rule base is stored in the database of the dispatch control center. It is a digital form of a series of rules and regulations formulated based on the requirements of rail transit operation safety and efficiency. The line capacity constraint rules clearly stipulate the maximum number of trains that can be safely operated per hour on each line in different time periods. This number is determined comprehensively based on factors such as the track strength, signal system response speed, and platform length. For example, the maximum number of trains per hour on a certain trunk line is a certain value during peak hours and a smaller value during off-peak hours.

[0030] The minimum interval rule for trains specifies the minimum time interval that must be maintained between two trains running on the same line. This minimum time interval is determined based on factors such as the braking distance of the train and the reaction time of the signal system, and may have different values ​​for different lines and different time periods.

[0031] The station passenger flow carrying capacity rules determine the maximum number of passengers that each station can carry at different times based on the station's platform area, passageway width, number of entrances and exits, and other facility conditions. When the number of people in the station exceeds this number, flow restriction measures can be implemented.

[0032] The emergency dispatch priority rules detail the execution order of various dispatch instructions in different types of emergency situations. For example, in the event of a fire, the priority of the emergency evacuation instruction for trains is higher than that of the normal departure instruction; in the event of equipment failure, the dispatch instruction for maintenance trains has a higher priority than that for ordinary passenger trains.

[0033] When converting these rules into rule constraint features, feature encoding is used. For line capacity constraint rules, the maximum number of trains per line in each time period is converted into a numerical matrix, where rows represent lines, columns represent time periods, and matrix elements represent the corresponding maximum number of trains. Train minimum interval rules are converted into a list containing line identifiers and corresponding minimum interval values. Station passenger capacity rules are converted into a dictionary structure indexed by station identifiers and containing the maximum passenger capacity for different time periods. Emergency dispatch priority rules are converted into a priority sorting table, specifying the priority value for each instruction type.

[0034] Step S115: Analyze the correlation between baseline operating features and rule constraint features in each typical operating scenario data unit, calculate the matching degree parameter of the correlation, and based on the matching degree parameter, fuse the baseline operating features, rule constraint features, and scenario identification features of typical operating scenarios to generate a dynamic scheduling baseline.

[0035] In each typical operational scenario data unit, the correlation analysis between baseline operating characteristics and rule constraint characteristics is a meticulous, multi-step process. Taking the "morning peak-sunny day-weekday" scenario data unit as an example, the analysis first examines whether the train interval in the baseline operating characteristics falls within the range specified by the minimum train interval rule in the rule constraint characteristics, and statistically analyzes the proportion of trains whose actual operating intervals meet the minimum interval requirement in this scenario. Secondly, the analysis examines whether the number of trains in the baseline operating characteristics is within the maximum number of trains specified by the line capacity constraint rules, and calculates the distribution of the ratio between the actual number of trains and the maximum allowed number. Simultaneously, the analysis examines whether the station passenger flow data falls within the range of the station passenger flow carrying capacity rules. Through these analyses, a matching degree parameter reflecting the closeness of the correlation between the two is calculated. Then, based on this parameter, the baseline operating characteristics, rule constraint characteristics, and corresponding scenario identification characteristics are organically integrated to form the dynamic scheduling baseline components for this scenario. Finally, the baseline components of all scenarios are integrated to construct the dynamic scheduling baseline for the entire rail transit system.

[0036] Step S1151: Extract the feature vectors of the baseline operating features and the feature vectors of the rule constraint features from the typical operation scenario data unit, and convert them into vector representations of the same dimension. Then, calculate the cosine similarity between the baseline operating feature vector and the rule constraint feature vector, and use the cosine similarity value as the initial matching degree parameter.

[0037] For each typical operational scenario data unit, a feature vector is first extracted from the baseline operational features. This feature vector includes specific values ​​for train punctuality rate across different dimensions, various statistical indicators of interval running time, various statistical indicators of station dwell time, and various data on resource occupancy rates, with each feature serving as a dimension of the vector. Similarly, a feature vector is extracted from the rule constraint features, including various values ​​for line capacity constraints, specific values ​​for minimum train intervals, various values ​​for station passenger flow capacity, and values ​​for emergency dispatch priority.

[0038] When two vectors have different dimensions, feature expansion can be performed on the vector with fewer dimensions, for example, by adding new dimensions through interpolation, so that it has the same number of dimensions as the vector with more dimensions; or feature selection can be performed on the vector with more dimensions, retaining the dimensions related to the vector with fewer dimensions, thus unifying the dimensions of the two.

[0039] After unifying the dimensions, the cosine similarity between the two vectors is calculated. The calculation process involves first calculating the dot product of the two vectors, which is the sum of the product of the corresponding dimensions; then calculating the magnitudes of the two vectors, which is the square root of the sum of the squares of the values ​​in each dimension; finally, dividing the dot product by the product of the two magnitudes yields the cosine similarity value. This cosine similarity value ranges from -1 to 1, with values ​​closer to 1 indicating a higher degree of similarity. This cosine similarity value is used as a preliminary matching parameter.

[0040] Step S1152: Analyze the correlation strength between the train punctuality rate feature in the benchmark operation characteristics and the train minimum interval rule in the rule constraint characteristics, and calculate the compliance parameter of the punctuality rate with the interval rule.

[0041] When analyzing the correlation between train punctuality rate characteristics and the minimum train interval rule, the data for different punctuality rate intervals are first selected from the benchmark operating characteristics of train punctuality rate characteristics, such as intervals above 90%, 80%-90%, and 70%-80%. For each punctuality rate interval, the number of times the train interval conforms to the minimum train interval rule, as well as the total number of runs, are counted. Then, the ratio of the number of conformities to the total number of runs is calculated to obtain the conformity rate for that punctuality rate interval.

[0042] The compliance rate of all punctuality rate intervals is weighted and averaged according to the punctuality rate, with the weight being the proportion of the number of runs in each interval to the total number of runs. The result is the compliance parameter of punctuality rate with the interval rule, which reflects the degree of correlation between train punctuality rate and compliance with the minimum train interval rule.

[0043] Step S1153: Analyze the correlation strength between the resource occupancy rate feature in the baseline operating characteristics and the line capacity constraint rule in the rule constraint characteristics, and calculate the compliance parameter of the resource occupancy rate with the capacity rule.

[0044] When analyzing the correlation between resource occupancy rate characteristics and line capacity constraint rules, the line resource occupancy rate is divided into multiple intervals, such as below 60%, 60%-80%, and above 80%. For each resource occupancy rate interval, the number of times the actual number of trains operating on the line meets the maximum train quantity requirement in the line capacity constraint rules, and the total number of operations, are counted. The ratio of the number of compliances to the total number of operations is calculated to obtain the compliance rate for that resource occupancy rate interval.

[0045] Similarly, the compliance rate of all resource utilization rate intervals is weighted and averaged according to the proportion of the number of runs within the resource utilization rate interval to the total number of runs. This yields the compliance parameter of resource utilization rate with capacity rules, which reflects the degree of correlation between resource utilization rate and compliance with line capacity constraint rules.

[0046] Step S1154: The preliminary matching degree parameter, the on-time rate compliance parameter with the interval rule, and the resource occupancy rate compliance parameter with the capacity rule are weighted and summed to generate the comprehensive matching degree parameter.

[0047] To comprehensively evaluate the degree of matching between baseline operating characteristics and rule constraint characteristics, it is necessary to perform a weighted summation of the preliminary matching degree parameter, the compliance parameter of punctuality rate with the interval rule, and the compliance parameter of resource occupancy rate with the capacity rule. The weights are determined based on expert experience and statistical analysis of historical data. Since the minimum train interval rule and the line capacity constraint rule are directly related to operational safety, the compliance parameters of punctuality rate with the interval rule and resource occupancy rate with the capacity rule have relatively high weights, while the preliminary matching degree parameter has a relatively low weight.

[0048] The specific operation involves multiplying each parameter by its corresponding weight, then summing these three products to obtain the overall matching degree parameter. This overall matching degree parameter integrates matching information from multiple aspects, providing a more comprehensive reflection of the matching status between the baseline operating characteristics and the rule constraint characteristics.

[0049] Step S1155: If the comprehensive matching degree parameter is higher than the preset matching threshold, the baseline operation feature, rule constraint feature and the corresponding typical operation scenario identification feature are spliced ​​together to generate the scenario baseline unit of the typical operation scenario.

[0050] The preset matching threshold is determined based on the historical safe operation data and scheduling quality requirements of the urban rail transit network. This preset matching threshold has undergone multiple tests and adjustments to ensure that only benchmark operating features and rule constraint features with a high degree of matching can be used to construct scenario baseline units. When the calculated comprehensive matching degree parameter is higher than the preset matching threshold, it indicates that the benchmark operating features and rule constraint features have good consistency and adaptability under this typical operating scenario.

[0051] At this point, the feature vectors of the baseline operation features, the feature vectors of the rule constraint features, and the identification features of the typical operation scenario (such as "morning peak - sunny day - weekday") are converted into a unified feature format and then concatenated in the order of the baseline operation feature vector, the rule constraint feature vector, and the scenario identification features to form a longer feature vector, which is the scenario baseline unit of the typical operation scenario.

[0052] Step S1156: Classify and store the scenario baseline units of all typical operation scenarios according to scenario identification characteristics, and build a baseline library for dynamic scheduling baselines.

[0053] After generating scenario baseline units for all typical operational scenarios, they are classified according to various dimensions of the scenario identifier features. First, they are categorized according to the operational time period characteristics into categories such as morning peak, evening peak, off-peak, and nighttime; under each time period category, they are further categorized according to weather conditions into subcategories such as sunny, cloudy, rainy, and snowy; under each weather subcategory, they are further categorized according to holiday attributes into subcategories such as weekday, weekend, and statutory holidays.

[0054] After classification, the scene baseline units for each category are stored in the corresponding table in the database. The database adopts a distributed storage architecture to improve data access efficiency. At the same time, an index based on scene identifier features is built for the database, so that in subsequent queries, the corresponding scene baseline unit can be quickly located based on the current scene information, thereby constructing a baseline library for dynamically scheduled baselines.

[0055] Step S1157: Set the dynamic update cycle of the baseline library, recalculate the comprehensive matching degree parameter of each scene baseline unit in each update cycle, and replace the scene baseline units whose comprehensive matching degree parameter is lower than the preset threshold.

[0056] Considering that the operational status of urban rail transit networks changes over time, such as the addition of new lines, changes in the number of trains, and changes in passenger travel habits, the baseline database is set to be updated dynamically once a week. Within each update cycle, historical normal scheduling data from the previous week is first collected, and the comprehensive matching degree parameter of each typical operating scenario data unit is recalculated according to steps S111 to S1154.

[0057] For existing scenario baseline units, if the recalculated comprehensive matching parameter is lower than the preset matching threshold, it indicates that the scenario baseline unit can no longer accurately reflect the current operational status and needs to be replaced with a newly calculated scenario baseline unit whose comprehensive matching parameter is higher than the preset threshold. For newly added typical operational scenario data units, the comprehensive matching parameter is calculated according to the same process. If it meets the requirements, its corresponding scenario baseline unit is added to the baseline library, thereby ensuring that the baseline library can always accurately reflect the current operational status and rule requirements.

[0058] Step S120: Collect real-time scheduling data of the rail transit system, compare and analyze the real-time scheduling data with the dynamic scheduling baseline in multiple dimensions, and generate a baseline deviation feature set, which includes time dimension deviation features, spatial dimension deviation features and resource dimension deviation features.

[0059] During the real-time operation of the rail transit system, the dispatching and monitoring system continuously collects various real-time dispatching data, which is then transmitted to the dispatching analysis center in real time via a high-speed data transmission network. At the dispatching analysis center, the real-time dispatching data is meticulously compared with corresponding scenario baseline units matched from the baseline library of dynamic dispatching baselines. The differences between the two are analyzed from three key dimensions: time, space, and resources, identifying the sources of deviation and generating a baseline deviation feature set containing deviation information from these three dimensions.

[0060] Step S121: Collect real-time dispatch data through the rail transit dispatch monitoring system. This real-time dispatch data includes real-time train location data, real-time line occupancy status data, real-time station passenger flow data, and real-time dispatch instruction execution data.

[0061] The rail transit dispatching and monitoring system is a comprehensive system integrating various monitoring devices and data transmission protocols. Real-time train location data is collected every second by high-precision positioning equipment installed on the train, including the train's real-time latitude and longitude coordinates, the line identifier, and the section identifier. At the same time, the onboard computer on the train packages this location information with the train number identifier and the current timestamp, and then sends it to the dispatching center through a dedicated wireless communication network.

[0062] Real-time track occupancy status data is collected by sensors installed at the entrances and exits of track sections. Each sensor can detect the time when a train enters and leaves the section. Based on this time information, it is possible to calculate whether a train is currently occupying each track section and the duration of its occupation. The signaling system also uploads its judgment results on the track section occupancy status in real time as a supplement.

[0063] Real-time station passenger flow data is collected collaboratively by various devices within the station. At each entrance and exit, the turnstile system records every opening and closing action in real time and links it to the corresponding ticket information to count the number of people entering and exiting the station every minute. High-definition cameras are installed in the station hall and platform areas. These cameras are connected to an image analysis server, which processes the video streams in real time using human recognition algorithms to count the number of people remaining in different areas at different times, including the total number of people remaining in the station hall, the number of people remaining on each platform, and the number of people moving in the passageways. In addition, some stations have installed passenger flow sensors at escalators and stairwells to further assist in counting changes in passenger flow direction and density. All of this data constitutes the real-time station passenger flow data, which is transmitted to the dispatch center in real time through a dedicated data channel.

[0064] Real-time dispatch instruction execution data is collected by the dispatch center's instruction execution feedback system. When the dispatch center issues a dispatch instruction, such as adjusting train departure times or changing train platforms, the relevant train onboard systems and station control systems will provide real-time feedback on the instruction's reception status, start time, execution progress, and execution results. For example, after receiving a departure time adjustment instruction, the train can report whether the instruction was successfully received, whether the expected departure time was adjusted according to the instruction, and the deviation between the actual departure time and the instruction requirements. After receiving a passenger flow management instruction, the station can report the activation status of management measures, the status of station announcements, and the dispatch status of staff. This feedback information is summarized to form real-time dispatch instruction execution data.

[0065] Step S122: Based on the timestamp information of the real-time scheduling data, match the corresponding typical operation scenario data unit from the dynamic scheduling baseline, and obtain the baseline operation characteristics and rule constraint characteristics corresponding to the typical operation scenario data unit.

[0066] Each record in the real-time scheduling data contains precise timestamp information, accurate to the millisecond, recording the exact moment the data was collected. The scheduling analysis system first parses this timestamp information to determine the current date and specific time. Based on the date information, it queries a pre-set holiday database to determine whether the current date is a weekday, weekend, or statutory holiday, thus determining the holiday attribute characteristics. Based on the specific time, it compares it with pre-set operating period division standards to determine whether the current period is morning peak, evening peak, off-peak, or nighttime, thus determining the operating period characteristics. Simultaneously, the scheduling analysis system maintains a real-time connection with the city's meteorological service system to obtain current weather information, including weather type (sunny, cloudy, rainy, snowy, etc.) and the intensity of precipitation or snowfall, thus determining the weather condition characteristics.

[0067] After determining the above three characteristics, the scheduling analysis system combines the operating time period characteristics, weather condition characteristics, and holiday attribute characteristics into a current scenario identifier, such as "morning rush hour - moderate rain - weekday". Then, based on this scenario identifier, it searches the baseline database of the dynamic scheduling baseline to find the corresponding typical operating scenario data unit. The search process utilizes an index built based on the scenario identifier characteristics in the baseline database to quickly locate the matching unit. After finding the corresponding typical operating scenario data unit, the corresponding baseline operating characteristics and rule constraint characteristics are extracted from that unit as a benchmark reference for subsequent comparative analysis.

[0068] Step S123: Align the real-time train position data with the historical train position data in the reference operation features on the time axis, and calculate the difference between the real-time arrival time and the reference arrival time within the same line section as the time dimension deviation feature.

[0069] To ensure the comparability of real-time train location data with historical train location data, timeline alignment is required. First, the train's route information and the expected passage sequence for each section are extracted from the real-time train location data. Then, in the historical train location data based on baseline operating characteristics, historical travel records of the same train under similar date types and weather conditions are found, and the corresponding passage sequence for each section is extracted. The passage sequences of the two sets of data are compared to ensure they are completely identical routes. If there are any temporary detours or other issues, the historical data needs to be re-filtered and matched.

[0070] After determining the matching historical data, the starting point of each track section is used as the reference point for the timeline. The time when real-time trains enter each track section is aligned with the time when historical trains enter the same track section. For example, the time when a train enters the first track section is used as the baseline zero point. The offset of the entry time of real-time trains and historical trains in each subsequent track section relative to this baseline zero point is calculated. The timeline alignment is achieved in this way.

[0071] After alignment, for the same line section, the real-time arrival time of the train at the end of that section and the baseline time of the historical train at the end of that section are found, and the difference between these two times is calculated. If the real-time arrival time is later than the baseline arrival time, the difference is positive; if the real-time arrival time is earlier than the baseline arrival time, the difference is negative. These differences for all line sections are arranged in order of line section sequence to form a time-dimensional deviation characteristic, reflecting the deviation of the train from the baseline operating state in the time dimension.

[0072] Step S1231: Extract real-time train location data from real-time dispatch data, and parse the train identification information, line section identification information, and arrival timestamp information in the real-time train location data.

[0073] From the overall real-time scheduling data stream, all records marked as real-time train location data are filtered out using data filtering rules. Each record contains a header that clearly identifies the data type as train location data. Each train location data record is parsed to extract the train identification information. This train identification is a unique string composed of the train's line code, train number, etc., which uniquely identifies a specific train.

[0074] The system extracts the track section identification information, which consists of a track code and a section number. The track code represents the track the train is currently on, and the section number represents the specific section on that track. For example, the section numbers on a track increase sequentially from the starting point, with each number corresponding to a continuous track section. Simultaneously, the system extracts the arrival timestamp information, accurate to the second, recording the moment the train completely enters the end point of the track section—that is, the time when both the front and rear of the train have passed the signal marker indicating the end of the section.

[0075] Step S1232: Based on train identification information and line section identification information, match the corresponding historical train location data from the baseline operating features and extract historical arrival timestamp information.

[0076] Using the parsed train identification information, a search is performed in the historical train location data of the baseline operating characteristics to find all historical operation records of the train over a past period. Then, based on the line section identification information, records containing the same line section identification are filtered out, that is, records in which the train has traveled through the same line section.

[0077] For the selected historical records, further matching is performed based on current scene characteristics. Priority is given to historical records that are the same as or similar to the characteristics of the current operating period, weather conditions, and holiday attributes to ensure the reference value of historical data. From the historical records with the highest matching degree, the historical arrival timestamp information of the train passing through the end of the corresponding line section is extracted. This timestamp is also accurate to the second and is consistent with the format of the real-time arrival timestamp.

[0078] Step S1233: Convert the real-time arrival timestamp information and the historical arrival timestamp information into time values ​​in the same time coordinate system.

[0079] Both real-time and historical arrival timestamps contain time elements such as year, month, day, hour, minute, and second. To calculate the time difference, they need to be converted to time values ​​in the same time coordinate system. The conversion method involves converting the timestamp information into the total number of seconds counted from 00:00:00 on the current day.

[0080] For example, if a real-time arrival timestamp is 8:10:05 AM on a weekday, the converted time value is the total number of seconds from midnight to that moment, which is eight hours multiplied by 3,600 seconds per hour, plus ten minutes multiplied by 60 seconds per minute, plus five seconds to obtain the corresponding total number of seconds. Historical arrival timestamps are converted in the same way to obtain the corresponding total number of seconds. Through the above conversion, the real-time arrival time and historical arrival time are placed on the same time coordinate system, which facilitates subsequent difference calculations.

[0081] Step S1234: Calculate the difference between the real-time arrival time value and the historical arrival time value to obtain the time deviation value of a single line section.

[0082] In the same time coordinate system, the time deviation value for a single line section is obtained by subtracting the historical arrival time value from the real-time arrival time value. If the deviation value is positive, it means that the real-time train arrives at the end of the line section later than the historical reference time, indicating a delay; if the deviation value is negative, it means that the real-time train arrives earlier than the historical reference time, indicating an advance; if the deviation value is zero or close to zero, it means that the real-time arrival time is basically consistent with the historical reference time.

[0083] For each section of the line, the corresponding time deviation value is calculated according to the above method. These time deviation values ​​reflect the time deviation of the train in different sections.

[0084] Step S1235: Perform sequence analysis on the time deviation values ​​of the same train in continuous track sections, identify the changing trend of the time deviation values, and calculate the trend change rate parameter.

[0085] The time deviation values ​​of the same train in consecutive track sections are arranged according to the travel order of the track sections to form a time deviation sequence. This sequence is then visualized to plot a curve showing the change in deviation values ​​as a function of the track sections. The trend of change is identified by observing the direction of the curve. For example, a continuously rising curve indicates that train delays are worsening; a continuously falling curve indicates that trains are arriving ahead of schedule; and a fluctuating curve without a clear pattern indicates that the time deviation is unstable.

[0086] To quantify the aforementioned trend, a trend change rate parameter is calculated. Specifically, the continuous line intervals are divided into multiple adjacent interval pairs. For each interval pair, the difference between the time deviation value of the subsequent interval and the time deviation value of the preceding interval is calculated; this difference reflects the amount of change in the deviation value. Then, the average of these changes is used as the core indicator of the trend change rate parameter, while the standard deviation of the changes is calculated to reflect the stability of the trend. Combining the average and standard deviation forms the trend change rate parameter, comprehensively reflecting the changing trend and fluctuations of the time deviation value.

[0087] Step S1236: Combine the time deviation value and trend change rate parameter of a single line interval to generate a time dimension deviation feature, which includes the interval time deviation sequence and the deviation trend change sequence.

[0088] Arrange all the time deviation values ​​of individual line sections in the order of the line sections the train travels to form a time deviation sequence. Each element in the time deviation sequence corresponds to the time deviation value of a line section, which can intuitively show the distribution of the time deviation of the train on the entire route.

[0089] Simultaneously, the trend change rate parameters are arranged in the order of the corresponding interval pairs to form a deviation trend change sequence. For example, for n line intervals, n-1 interval pairs of trend change rate parameters can be formed. These parameters are arranged in the order of the interval pairs to form a deviation trend change sequence, reflecting the changing trend of time deviation between different intervals.

[0090] By combining the interval time deviation sequence and the deviation trend change sequence, that is, by splicing the two sequences together in sequence, a time dimension deviation feature is formed, which comprehensively reflects the deviation of the train in the time dimension and its changing trend.

[0091] Step S124: Compare the real-time line occupancy status data with the line capacity constraint rules in the rule constraint features, and identify the deviation value between the current line interval occupancy rate and the maximum occupancy rate allowed by the rule as the spatial dimension deviation feature.

[0092] First, occupancy information for each track section is extracted from real-time track occupancy status data, including the number of trains currently operating in that section, the start time of each train's occupation of the section, and its estimated departure time. Based on this information, the occupancy rate of each track section at the current moment is calculated. This is done by dividing the total length of trains currently occupying the section by the total length of the section, and then combining this with the train speed and section length to convert it into an occupancy percentage per unit time. For example, if the total length of a track section is a fixed value, and two trains are currently in that section, the ratio of their total length to the total section length is a certain percentage. Combined with the average time it takes for trains to pass through the section, the proportion of the section occupied per unit time is calculated, which is the current track section occupancy rate.

[0093] The maximum allowable occupancy rate of the corresponding line section is extracted from the line capacity constraint rules. This maximum allowable occupancy rate is determined based on factors such as the line design standards, signal system capabilities, and safety redundancy. Different line sections and different time periods may have different values. For example, during peak hours, the maximum allowable occupancy rate may be set higher to improve transportation efficiency, while it may be relatively lower during off-peak hours.

[0094] The current occupancy rate of a line section is compared with the maximum allowed occupancy rate, and the difference between the two is calculated as the current occupancy rate minus the maximum allowed occupancy rate. If the difference is positive, it means that the current occupancy rate exceeds the maximum allowed value, indicating excessive space occupancy. If the difference is negative, it means that the current occupancy rate is within the allowed range and has some redundancy. If the difference is zero, it means that the current occupancy rate has just reached the maximum allowed value. The above deviation values ​​of all line sections are arranged in order of line section to form spatial dimension deviation characteristics.

[0095] Step S125: Match the real-time station passenger flow data with the historical passenger flow data in the baseline operation characteristics for different time periods. Combine the resource allocation information in the real-time dispatch instruction execution data to calculate the adaptation deviation value between the current resource configuration and passenger flow demand as the resource dimension deviation feature.

[0096] First, the real-time station passenger flow data is divided into time periods, with each day divided into multiple consecutive one-hour periods. The number of people entering, exiting, and the average number of people remaining in the station within each period are calculated to form real-time passenger flow time period data. From the historical passenger flow data in the baseline operating characteristics, historical passenger flow data with the same characteristics as the current scenario (operating hours, weather conditions, holiday attributes) are found. Using the same time period division method, the historical number of people entering, exiting, and the historical average number of people remaining in the station for each period are calculated to form historical passenger flow time period data.

[0097] Matching real-time passenger flow data with historical passenger flow data involves comparing passenger flow data for the same time period and calculating the ratios of real-time entry numbers to historical entry numbers, real-time exit numbers to historical exit numbers, and real-time average number of people remaining in the station to historical average number of people remaining in the station. These ratios reflect the degree of difference between the current passenger flow and the historical baseline passenger flow, constituting the characteristics of passenger flow demand.

[0098] Resource allocation information is extracted from real-time dispatch instruction execution data, including the number of staff at each station in each time period, the number of turnstiles opened, the number of escalators in operation, the number of platforms the train stops at, and the stopping time. This resource allocation information is compared with historical baseline resource configuration information (extracted from baseline operating characteristics), and the ratio of the current resource configuration to the historical baseline resource configuration is calculated, such as the ratio of the current number of staff to the historical baseline number of staff, and the ratio of the current number of turnstiles opened to the historical baseline number of turnstiles opened, etc., which constitute the resource allocation characteristics.

[0099] To calculate the mismatch between current resource allocation and passenger flow demand, the ratios of each item in the resource allocation feature are compared with the corresponding ratios in the passenger flow demand feature. The difference for each corresponding item is calculated, i.e., the resource allocation ratio minus the passenger flow demand ratio. For example, subtracting the real-time entry ratio from the current staff number ratio yields the mismatch between staff allocation and entry passenger flow demand; subtracting the real-time exit ratio from the current number of open turnstiles yields the mismatch between turnstile allocation and exit passenger flow demand, and so on. All these mismatch values ​​are combined in a predetermined order to form the resource dimension deviation feature.

[0100] Step S126: The time dimension deviation features, spatial dimension deviation features, and resource dimension deviation features are associated and integrated according to the preset dimension weights to generate a baseline deviation feature set.

[0101] The preset dimension weights are determined based on the actual needs and historical experience of rail transit scheduling, and are set through a combination of expert evaluation and data analysis. Time dimension deviation characteristics directly affect train punctuality and operational efficiency, and are given a higher weight during peak hours; spatial dimension deviation characteristics relate to the safe operation and capacity utilization of the line, and are given a relatively higher weight when the line load is high; resource dimension deviation characteristics affect the effectiveness of passenger flow management and passenger experience, and are given a higher weight at stations with high passenger flow.

[0102] The process of correlation and integration involves first converting the time-dimension, spatial-dimension, and resource-dimension deviation features into standardized feature vectors, with each element of the feature vector being a dimensionless ratio or deviation value. Then, each element of each feature vector is multiplied by its corresponding dimension weight to obtain a weighted feature vector. Finally, the three weighted feature vectors are concatenated in the order of time, space, and resource dimensions to form a baseline deviation feature set, containing deviation information and their weighted effects across the three dimensions.

[0103] Step S130: Call the pre-trained scheduling collaborative evaluation model to perform dynamic correlation analysis on the baseline deviation feature set and generate scheduling collaborative evaluation results. The scheduling collaborative evaluation results include deviation propagation path parameters, collaborative conflict probability parameters, and resource adaptability parameters.

[0104] The scheduling coordination evaluation model is a deep learning-based neural network model specifically designed to analyze the correlations between various deviation features in rail transit scheduling. Trained on a large amount of historical baseline deviation feature data and corresponding scheduling evaluation results, this model can automatically learn the hidden correlations and transmission patterns between deviation features. When the model is invoked, the baseline deviation feature set is taken as input. The scheduling coordination evaluation model uses its internal multi-layer neural network to extract features and perform correlation analysis, outputting scheduling coordination evaluation results that include deviation transmission path parameters, coordination conflict probability parameters, and resource suitability parameters.

[0105] Step S131: Input the baseline deviation feature set into the feature preprocessing layer of the scheduling collaborative evaluation model, and standardize the time dimension deviation features, spatial dimension deviation features and resource dimension deviation features to generate standard deviation feature vectors.

[0106] The baseline bias feature set first enters the feature preprocessing layer of the scheduling collaborative evaluation model. The main function of this feature preprocessing layer is to standardize the input features to eliminate differences in dimensions and numerical ranges between features of different dimensions, thereby improving the analytical accuracy of the model. For each element in the time dimension bias feature, the mean-variance standardization method is used, that is, each element is subtracted from the mean of the time dimension bias feature and then divided by the standard deviation of the time dimension bias feature.

[0107] Similarly, each element in the spatial dimension bias feature and the resource dimension bias feature is standardized by subtracting its mean and dividing by its standard deviation. After standardization, the bias features of the three dimensions are converted into standard feature vectors with a mean of zero and a standard deviation of one. These three standard feature vectors are then concatenated in the order they were input to form a unified standard deviation feature vector, which is used as the input to the next layer of the model.

[0108] Step S132: By using the graph structure construction module of the scheduling collaborative evaluation model, a deviation correlation graph structure is constructed with the standard deviation feature vector as the node and the correlation strength between deviation features of each dimension as the edge weight.

[0109] After the standard deviation feature vector enters the graph structure construction module, the graph structure construction module first treats each element in the standard deviation feature vector as an independent node. Each node represents a specific deviation feature element, such as the time deviation value of a certain line section, the resource adaptation deviation value of a certain station, etc.

[0110] Then, the association strength between any two nodes is calculated, serving as the weight of the edge connecting them. The association strength is calculated based on the co-occurrence frequency and the similarity of the changing trends of the deviation feature elements represented by the two nodes in historical data. Specifically, the proportion of times two deviation feature elements simultaneously exhibit non-zero deviations in the historical data is counted as the co-occurrence frequency; the correlation coefficient of the change curves of the two deviation feature elements in the historical data is calculated as the changing trend similarity. The co-occurrence frequency and the changing trend similarity are then weighted and summed according to a predetermined ratio to obtain the association strength between the two nodes, i.e., the edge weight.

[0111] Based on node and edge weights, an undirected deviation correlation graph structure is constructed. Each node in the graph corresponds to a deviation feature element, and the weight of each edge corresponds to the correlation strength between two deviation feature elements. This deviation correlation graph structure can intuitively demonstrate the interrelationships between different deviation feature elements. For example, the train arrival delay element in the time dimension deviation feature may have a strong correlation with the line section congestion element in the spatial dimension deviation feature; therefore, in the graph, there would be a high-weighted edge between the nodes representing these two elements. Similarly, the insufficient station staff element in the resource dimension deviation feature may be closely related to the extended train dwell time caused by slow passenger boarding and alighting in the time dimension deviation feature; correspondingly, the edge weight between them would also be high.

[0112] Step S133: Use the path mining algorithm of the scheduling collaborative evaluation model to perform path search processing on the deviation association graph structure, identify the key transmission path from the initial deviation node to the affected node, and extract the path length parameter, node influence intensity parameter and path transmission efficiency parameter as deviation transmission path parameters.

[0113] The path mining algorithm in the scheduling and collaborative evaluation model is a tool specifically designed to find deviation propagation paths within a deviation correlation graph structure. Initial deviation nodes are those whose deviation values ​​exceed a preset deviation threshold; these nodes are the starting points for deviation propagation. Influencing nodes are those nodes that significantly impact the overall operational efficiency or safety of the rail transit system, such as passenger flow-related nodes at key transfer stations and section occupancy-related nodes on main lines. Using the path mining algorithm, starting from the initial deviation nodes, all possible paths leading to the influencing nodes are explored. Then, key propagation paths are selected, and relevant parameters are extracted.

[0114] Step S1331: Identify all initial deviation nodes from the deviation correlation graph structure. Initial deviation nodes are nodes whose deviation values ​​exceed a preset deviation threshold in the time dimension deviation feature, spatial dimension deviation feature, or resource dimension deviation feature.

[0115] The preset deviation thresholds are determined based on the operational standards and historical data of the rail transit system. Different types of deviation feature nodes will have different preset deviation thresholds. For example, for train arrival time deviation nodes in the time dimension deviation features, the preset deviation threshold may be set to a time length. When the difference between the actual arrival time and the reference arrival time exceeds this length, the node is identified as an initial deviation node. For line section occupancy rate deviation nodes in the spatial dimension deviation features, the preset deviation threshold may be a percentage. When the difference between the actual occupancy rate and the reference occupancy rate exceeds this percentage, the node becomes an initial deviation node. By comparing the deviation value of each deviation feature node with the corresponding preset deviation threshold, all initial deviation nodes are filtered out.

[0116] Step S1332: Starting from each initial deviation node, use the depth-first search algorithm to perform path search in the deviation association graph structure and record the propagation path from the initial deviation node to all other nodes.

[0117] Depth-first search (DFS) is an algorithm that traverses the nodes of a graph in a depth-first manner. Starting from an initial deviation node, it searches as deep as possible along a path until it cannot proceed further. Then, it backtracks to the previous node and chooses another unexplored path to continue the search. During the search, it meticulously records every path from the initial deviation node that leads to all other nodes, including the nodes traversed and the order of connections between them. For example, starting from the initial deviation node "Train A is delayed at station X," a possible path is "Train A is delayed at station X → Subsequent train B at station X is forced to slow down → Train B arrives at station Y late → Passengers are stranded at station Y → Transfer train C at station Y is delayed." The algorithm will record this entire path.

[0118] Step S1333: Calculate the sum of edge weights for each transmission path and use the sum of edge weights as the path influence strength parameter. The larger the path influence strength parameter, the more significant the transmission influence of the transmission path.

[0119] Each propagation path consists of a series of nodes and edges. The weights of the edges reflect the strength of the association between two adjacent nodes, that is, the degree of influence of the deviation propagating from one node to another. For each recorded propagation path, the weights of all edges on the path are added together, and the sum is the path influence strength parameter. The larger this parameter value, the more significant the impact of the deviation propagation along this path. For example, if the sum of the edge weights on a path has a large value, it means that the deviation propagation along that path will have a significant impact on subsequent nodes.

[0120] Step S1334: Count the number of nodes in each transmission path and use the number of nodes as the path length parameter.

[0121] The path length parameter is an indicator that measures the length of a propagation path. It is determined by counting the number of nodes in each propagation path. The more nodes there are, the larger the path length parameter, indicating that the deviation needs to pass through more nodes to be propagated. For example, a propagation path with 5 nodes has a path length parameter of 5, while a propagation path with 3 nodes has a path length parameter of 3.

[0122] Step S1335: Calculate the ratio of the path influence intensity parameter to the path length parameter, and use this ratio as the path transmission efficiency parameter. The higher the path transmission efficiency parameter, the higher the influence transmission efficiency per unit path length.

[0123] The path transmission efficiency parameter reflects the efficiency of deviation transmission along the transmission path. During calculation, the path influence intensity parameter for each transmission path is divided by the path length parameter; the result is the path transmission efficiency parameter. If a path has a large path influence intensity parameter and a relatively small path length parameter, its path transmission efficiency parameter will be high, indicating a stronger impact of deviation transmission per unit path length. For example, if a path has a certain influence intensity parameter and a path length parameter of 3, its path transmission efficiency parameter is that value divided by 3; if another path has the same influence intensity parameter but a path length parameter of 5, its path transmission efficiency parameter is also that value divided by 5. Clearly, the former is more efficient.

[0124] Step S1336: Sort all conduction paths from high to low according to the path influence intensity parameter, and select a preset number of conduction paths as key conduction paths.

[0125] The preset number is determined based on the complexity of the rail transit system and the analysis requirements, ensuring that the selected critical transmission paths cover those deviation transmission paths that have the most significant impact on the system. After ranking all transmission paths according to the path influence intensity parameter, the top-ranked transmission paths have a more significant impact on the system. The preset number of transmission paths selected before ranking are identified as critical transmission paths, which are the focus of subsequent analysis and optimization.

[0126] Step S1337: Extract the path length parameter, node influence intensity parameter and path transmission efficiency parameter of each path from the critical transmission path, and combine them to generate deviation transmission path parameters.

[0127] For each critical transmission path, its path length parameter, node influence intensity parameter (i.e. path influence intensity parameter), and path transmission efficiency parameter are extracted. These parameters are then combined in a set order to form a deviation transmission path parameter that contains the parameters of each of the multiple critical transmission paths. This parameter comprehensively reflects information such as the length, influence intensity, and transmission efficiency of the critical deviation transmission path.

[0128] Step S134: Calculate the cumulative impact value of deviation characteristics on each key transmission path based on the deviation transmission path parameters, input the cumulative impact value into the collaborative conflict probability calculation module, and generate collaborative conflict probability parameters by combining the rule constraint features of the preset scheduling rule base.

[0129] First, for each critical transmission path, the cumulative impact value of the deviation characteristics on that path is calculated based on the deviation values ​​of each node and the weights of the edges between nodes. During the calculation, the deviation value of each node is transmitted and accumulated according to the weights of the edges between it and the next node, ultimately yielding the cumulative impact value of the entire path. Then, these cumulative impact values ​​are input into the collaborative conflict probability calculation module. This module, combined with the rule constraint features in the preset scheduling rule base, analyzes the potential conflicts between different scheduling stages caused by these cumulative impacts. For example, if the cumulative impact value of a critical transmission path is large, it may cause a conflict between train departure plans and line capacity. The module calculates the probability of such a conflict occurring based on historical data and rule constraints; the set of these probability values ​​constitutes the collaborative conflict probability parameter.

[0130] Step S135: Call the resource adaptability analysis module of the scheduling collaborative evaluation model, match and analyze the resource dimension deviation characteristics with the resource configuration information in the real-time scheduling data, and calculate the matching ratio parameter between the resource supply and the deviation repair demand as the resource adaptability parameter.

[0131] The resource fit analysis module is specifically designed to analyze the degree of fit between resource allocation and deviation correction requirements. Resource-dimensional deviation characteristics reflect the discrepancy between the current resource allocation and the baseline resource allocation. Real-time scheduling data includes resource allocation information such as the number of trains, the number of station staff, platform space size, and the number of security screening devices. By matching the resource-dimensional deviation characteristics with this resource allocation information, the required quantity of various resources for correcting deviations is determined. This quantity is then compared with the current actual resource supply to calculate the ratio between the supply and demand for each resource. The set of these ratio parameters constitutes the resource fit parameters. The closer the ratio is to 1, the higher the resource fit.

[0132] Step S136: Integrate deviation propagation path parameters, collaborative conflict probability parameters, and resource adaptability parameters to generate scheduling collaborative evaluation results.

[0133] The deviation propagation path parameters obtained in step S1337, the coordination conflict probability parameters obtained in step S134, and the resource adaptability parameters obtained in step S135 are integrated and combined according to a set format to form a complete scheduling coordination evaluation result. This scheduling coordination evaluation result comprehensively reflects the deviation propagation path information, the probability of possible coordination conflicts, and the resource adaptability.

[0134] Step S140: Construct a multi-objective scheduling optimization model based on the scheduling coordination evaluation results. Through the multi-objective scheduling optimization model, coordinate optimization is performed on the deviation propagation path parameters, coordination conflict probability parameters, and resource adaptability parameters to generate an initial set of scheduling adjustment schemes.

[0135] The multi-objective scheduling optimization model is a mathematical model that can handle multiple optimization objectives simultaneously. Based on the scheduling coordination evaluation results, it takes the deviation propagation path parameters, coordination conflict probability parameters, and resource adaptability parameters as optimization objects, sets corresponding optimization objectives and constraints, and obtains multiple possible scheduling adjustment schemes through solving, forming an initial set of scheduling adjustment schemes.

[0136] Step S141: Using the deviation propagation path parameters, collaborative conflict probability parameters, and resource adaptability parameters in the scheduling and coordination evaluation results as optimization objectives, construct the objective function of the multi-objective scheduling optimization model. The objective function includes the objectives of shortening the deviation propagation path, reducing the collaborative conflict probability, and improving resource adaptability.

[0137] The objective function defines the directions for optimization. The objective of shortening the deviation propagation path aims to reduce the length of critical propagation paths and decrease the propagation range of deviations by adjusting the scheduling scheme. The objective of reducing the probability of coordination conflicts aims to reduce the probability of conflicts between different scheduling stages, thereby improving system stability. The objective of improving resource adaptability aims to improve the fit between resource supply and deviation repair needs, ensuring effective resource utilization. These three objectives together constitute the objective function of the multi-objective scheduling optimization model.

[0138] Step S142: Extract line capacity constraint rules, train minimum interval rules, and station passenger flow carrying capacity rules from the preset scheduling rule base as constraints for the multi-objective scheduling optimization model.

[0139] Constraints are the restrictions that a multi-objective scheduling optimization model must adhere to, ensuring that the optimized scheduling scheme is feasible in actual operation. Line capacity constraints limit the maximum number of trains that can operate on each line per unit time, preventing line overload; minimum train interval constraints ensure sufficient safe distance and time intervals between trains on the same line; station passenger capacity constraints ensure that the number of passengers within a station does not exceed its capacity, avoiding safety accidents. Incorporating these rules as constraints into the model ensures that the optimization process proceeds within a reasonable range.

[0140] Step S143: Use the path length parameter and node influence intensity parameter in the deviation propagation path parameters as input variables for the deviation propagation path shortening target, and set the path length shortening weight coefficient.

[0141] The path length parameter directly reflects the length of the propagation path, while the node influence strength parameter reflects the magnitude of the propagation impact along the path. Using these two parameters as input variables for shortening the propagation path means that during optimization, priority can be given to shortening longer paths with greater impact. The path length shortening weight coefficient is set according to the degree of impact of different paths on the overall system operation; for paths with greater impact, the weight coefficient is set higher to ensure they receive more attention during optimization.

[0142] Step S144: Use the cooperative conflict probability parameter as the input variable for the cooperative conflict probability reduction target, and set the conflict probability reduction weight coefficient.

[0143] The cooperative conflict probability parameter contains the probability values ​​of various possible scheduling conflicts. Using it as an input variable for the cooperative conflict probability reduction objective means minimizing these probability values ​​during the optimization process. The conflict probability reduction weighting coefficient is determined based on the severity of different conflicts. For conflicts that may lead to serious consequences, such as those related to the risk of train rear-end collisions, the weighting coefficient is set higher to prioritize reducing their probability of occurrence.

[0144] Step S145: Use the resource adaptability parameter as the input variable for the resource adaptability improvement target, and set the adaptability improvement weight coefficient.

[0145] Resource suitability parameters reflect the matching ratio between the supply and demand of various resources. Using this parameter as an input variable for improving resource suitability aims to optimize scheduling schemes to make these ratios more reasonable. The suitability improvement weighting coefficient is set according to the importance of the resource. For critical resources, such as train capacity during peak hours, the weighting coefficient is set higher to prioritize improving their suitability.

[0146] Step S146: Solve the objective function under constraints using a multi-objective optimization algorithm to generate multiple sets of candidate optimization solutions, including train timetable adjustment parameters, line resource allocation adjustment parameters, and station passenger flow management adjustment parameters.

[0147] The multi-objective optimization algorithm employs advanced algorithms such as non-dominated sorting genetic algorithms to solve the objective function while satisfying constraints. The algorithm iteratively optimizes by simulating selection, crossover, and mutation processes in biological evolution, generating multiple sets of candidate optimization solutions. Each set of candidate solutions includes specific train timetable adjustment parameters, such as adjusting train departure times and changing station stops; line resource allocation adjustment parameters, such as adjusting the number of trains allocated to different lines; and station passenger flow management adjustment parameters, such as increasing the number of station staff and adjusting the opening status of entrances and exits.

[0148] For example, step S1461: Initialize the population parameters of the multi-objective optimization algorithm, and set the population size, maximum number of iterations and crossover mutation probability.

[0149] The settings of population parameters have a significant impact on the performance of multi-objective optimization algorithms. Population size refers to the number of candidate solutions initially generated. An excessively large population size may increase computational cost, while an excessively small population size may lead to insufficient searching. An appropriate population size should be set based on the complexity of the problem. The maximum number of iterations is the maximum number of steps the algorithm can take, ensuring that the algorithm completes the search within a reasonable timeframe. The crossover probability is the probability that two parent individuals will produce offspring during the crossover operation; the mutation probability is the probability that an individual's genes will mutate. The settings of these two probabilities need to balance the algorithm's exploration ability and convergence speed, and are usually determined through multiple trials to find suitable values.

[0150] Step S1462: Use the train timetable adjustment parameters, line resource allocation adjustment parameters, and station passenger flow management adjustment parameters as optimization variables, and encode the optimization variables to generate the initial population individuals.

[0151] The optimization variables are the parameters that the multi-objective optimization algorithm needs to optimize. Train timetable adjustment parameters, line resource allocation adjustment parameters, and station passenger flow management adjustment parameters are converted into an encoding format that the algorithm can process, such as binary encoding or real number encoding. For example, the adjustment amount of train departure time is encoded as a binary number of a set length, with each bit representing a different adjustment magnitude. Through the above encoding process, an initial population of individuals is generated, with each individual corresponding to a set of possible combinations of optimization variables.

[0152] Step S1463: Calculate the objective function value for each individual in the initial population. The objective function value includes the value of shortening the deviation transmission path, the value of reducing the probability of cooperative conflict, and the value of improving resource suitability.

[0153] For each individual in the initial population, its corresponding optimization variables are substituted into the objective function of the multi-objective scheduling optimization model to calculate the following: the deviation propagation path shortening value (the length shortened compared to the original path); the cooperative conflict probability reduction value (the extent of the reduction in conflict probability); and the resource fit improvement value (the degree of improvement in resource fit). These objective function values ​​reflect the performance of the scheduling scheme corresponding to that individual in achieving the optimization objectives.

[0154] Step S1464: Perform non-dominated sorting of individuals in the initial population based on the objective function value, determine the Pareto level of each individual, and calculate the crowding distance of the individuals.

[0155] Non-dominated ranking is a method used in multi-objective optimization to distinguish the superiority of individuals. An individual is said to dominate another individual if it is not inferior to another individual in all objective functions and is superior to another individual in at least one objective function. Non-dominated ranking divides the initial population into different Pareto ranks, with individuals in lower ranks being better. Crowding distance measures the density of individuals within the same Pareto rank; a larger distance indicates a sparser distribution of individuals in the solution space and better diversity. When calculating crowding distance, for each objective function, individuals within that rank are ranked according to their objective function values. The crowding distance between the two ends of the group is set to infinity, and the crowding distance between the middle individuals is the sum of the differences in the objective function values ​​of their two adjacent individuals.

[0156] Step S1465: Select superior individuals based on Pareto level and crowding distance to enter the next generation of the population, and generate new population individuals through crossover and mutation operations.

[0157] During the selection process, individuals with lower Pareto scores are prioritized. Within the same Pareto score, individuals with larger crowding distances are selected to ensure the quality and diversity of the population. Crossover involves exchanging the codes of two selected individuals according to a predetermined method to generate a new individual. For example, single-point crossover in binary encoding randomly selects a crossover point and swaps the codes of the two individuals after that point. Mutation randomly alters the codes of individuals, such as bit flipping in binary encoding, changing a 0 to a 1 or a 1 to a 0, to increase population diversity and prevent the algorithm from getting trapped in local optima.

[0158] Step S1466: Perform constraint condition verification on the new population individuals and remove individuals that do not meet the line capacity constraint rules, train minimum interval rules, and station passenger flow carrying capacity rules.

[0159] Newly generated individuals in the population may not meet the constraints and need to be validated. The scheduling scheme corresponding to each new individual is compared with the line capacity constraint rules, train minimum interval rules, and station passenger flow carrying capacity rules to check for violations. For example, it is checked whether the adjusted number of trains exceeds the line capacity limit, whether the interval between trains is less than the minimum interval requirement, and whether the station passenger flow exceeds its carrying capacity. Individuals that do not meet the constraints are removed, and only those that meet the requirements are retained for the next iteration.

[0160] Step S1467: Repeat the objective function value calculation, non-dominated sorting, selection, crossover mutation, and constraint verification steps until the maximum number of iterations is reached.

[0161] Following the steps described above, the population is continuously updated and optimized, with each iteration generating better individuals. When the number of iterations reaches the preset maximum, the algorithm stops running, and the population at this point contains multiple candidate solutions that are relatively good under the current conditions.

[0162] Step S1468: Extract all non-dominated solutions from the final population as multiple sets of candidate optimization solutions. Each set of candidate optimization solutions includes the corresponding train timetable adjustment parameters, line resource allocation adjustment parameters, and station passenger flow management adjustment parameters.

[0163] The non-dominated solutions in the final population are those solutions that cannot be dominated by any other solution on any objective function; these constitute the Pareto optimal solution set. Extracting these non-dominated solutions from the final population, each set corresponds to a specific set of train timetable adjustment parameters, line resource allocation adjustment parameters, and station passenger flow management adjustment parameters. These parameters together constitute a complete set of scheduling adjustment schemes, capable of optimizing the operation of the rail transit system from different perspectives.

[0164] The parameters for adjusting train timetables involve details such as train departure times, arrival times, and speeds within each section. For example, to address congestion on a line during the morning rush hour, the departure times of some trains may be adjusted to create more reasonable intervals between trains, preventing overcrowding or excessively large gaps. Simultaneously, based on the actual conditions of the line, the speeds of trains in certain sections may be fine-tuned to ensure that trains can pass through these sections more efficiently while maintaining safety.

[0165] The parameters for adjusting line resource allocation mainly include the allocation of the number of trains on the line and the allocation of track usage rights. For example, on a line with high passenger volume, the number of trains during peak hours may be increased to improve the line's transport capacity; for some branch lines or lines with lower passenger volume, the number of trains may be appropriately reduced to avoid wasting resources. At the same time, track usage rights are rationally allocated according to the different train operation plans to ensure that there are no track usage conflicts between trains.

[0166] The parameters for adjusting passenger flow at stations include the installation of directional signs within the station, staff deployment, and the number of turnstiles opened. For example, at a station experiencing a sudden surge in passenger traffic, temporary directional signs may be added to guide passengers to enter and exit the station quickly; at the same time, additional staff may be deployed to key locations such as platforms and stairwells to manage passenger flow and prevent congestion; and the number of turnstiles opened may be adjusted based on real-time passenger flow to speed up passenger entry and exit.

[0167] Step S150: Perform dynamic constraint verification on the initial set of scheduling adjustment schemes. After verification, generate the final rail transit scheduling instruction, which is used to trigger the rail transit scheduling control system to perform scheduling parameter update operations.

[0168] While the initial set of scheduling adjustment schemes performs well in the optimization algorithm, they still need to be verified under actual constraints to ensure their feasibility in real-world operation. Through dynamic constraint verification, schemes that meet all constraints are selected, and the optimal scheme is chosen to generate the final rail transit scheduling instructions, guiding the scheduling control system to update its parameters.

[0169] Step S151: Select an initial scheduling adjustment scheme from the initial scheduling adjustment scheme set, and extract the train timetable adjustment parameters, line resource allocation adjustment parameters, and station passenger flow management adjustment parameters contained in the initial scheduling adjustment scheme.

[0170] From the initial set of scheduling adjustment schemes, the first initial scheduling adjustment scheme is selected in a predetermined order. Then, this scheme is parsed to extract the train timetable adjustment parameters, line resource allocation adjustment parameters, and station passenger flow management adjustment parameters. These parameters will serve as inputs for subsequent verification processes to check whether the scheme meets the various constraints.

[0171] Step S152: Input the train timetable adjustment parameters into the line simulation module to simulate the adjusted train running path and stopping time, and generate line simulation results.

[0172] The track simulation module is a computer-based simulation system that can simulate the operation of trains on tracks. After the extracted train timetable adjustment parameters are input into the module, the module will construct a virtual train operation scenario based on these parameters, simulating the entire process of a train departing from the starting station and traveling through various sections and stopping at various stations according to the adjusted departure time, speed, and other parameters.

[0173] During the simulation, the system records the train's route in real time, including the track sections the train passes through and the travel time in each section; it also records the train's stopping time at each station, including detailed information such as arrival and departure times. This information collectively constitutes the line simulation results, intuitively demonstrating the adjusted train operation.

[0174] Step S153: Compare the line resource allocation adjustment parameters with the real-time line occupancy status data to verify whether the adjusted resource allocation meets the line capacity constraint rules and the minimum train interval rules.

[0175] Real-time track occupancy data reflects the current distribution of trains on the track and the usage status of the tracks. By comparing the track resource allocation adjustment parameters with this real-time data, the first step is to check whether the adjusted number of trains is within the capacity constraints of the track, ensuring that the track will not be overloaded due to an excessive number of trains, and guaranteeing the safe operation of the track.

[0176] Secondly, it is verified whether the adjusted train intervals comply with the minimum train interval rules. By comparing the operation plans of preceding and following trains, the time interval between adjacent trains is calculated to ensure that this interval is not less than the prescribed minimum interval, thus preventing safety accidents such as train collisions.

[0177] Step S154: Input the station passenger flow management adjustment parameters into the station passenger flow simulation module to simulate the passenger flow change trend after the adjustment and verify whether the passenger flow management effect meets the station passenger flow carrying capacity rules.

[0178] The station passenger flow simulation module can simulate passenger flow within a station based on input parameters. After inputting the station passenger flow management and adjustment parameters into the module, it will construct a virtual scene of the station based on these parameters, simulating a series of passenger behaviors such as entering the station, purchasing tickets, waiting for the train, checking tickets, and exiting the station.

[0179] During the simulation, the system tracks passenger flow trends in real time, including the number of passengers in different areas of the station, passenger flow speed, and passenger dwell time within the station. Using this data, the system analyzes whether the adjusted passenger flow management measures can effectively alleviate station passenger pressure, ensure that the number of passengers within the station does not exceed the station's passenger capacity limit, and verify whether the passenger flow management effect meets the station's passenger capacity rules.

[0180] Step S155: Based on the integrated line simulation results, resource allocation verification results, and passenger flow management verification results, calculate the comprehensive constraint satisfaction parameter of the initial scheduling adjustment scheme.

[0181] The comprehensive constraint satisfaction parameter is an indicator that comprehensively reflects the degree to which the initial scheduling adjustment scheme satisfies various constraints. When calculating this parameter, the simulation results of the line, the verification results of resource allocation, and the verification results of passenger flow management are first quantitatively scored.

[0182] For the line simulation results, scores are given based on whether the trains can run smoothly according to the adjusted plan and whether there are delays. For the resource allocation verification results, scores are given based on whether the number of trains meets the capacity constraints and whether the train intervals meet the minimum interval rules. For the passenger flow management verification results, scores are given based on whether the number of passengers in the station is within the carrying capacity and whether the passenger flow is smooth.

[0183] Then, based on the importance of each constraint, a corresponding weight is assigned to each score. For example, line capacity constraints and train minimum interval rules are related to operational safety, so their weights may be set higher; while the weight of passenger flow management effectiveness is relatively lower. Finally, each score is multiplied by its corresponding weight and then summed to obtain the comprehensive constraint satisfaction parameter of the initial scheduling adjustment plan.

[0184] Step S156: If the comprehensive constraint satisfaction parameter reaches the preset threshold, then mark the initial scheduling adjustment scheme as a valid scheme.

[0185] The preset threshold is determined based on the operational requirements and safety standards of the rail transit system, representing the minimum acceptable level of satisfaction for the proposed solution. When the calculated comprehensive constraint satisfaction parameter reaches this threshold, it indicates that the initial scheduling adjustment scheme has met the specified requirements in all constraints, ensuring the safe and efficient operation of the rail transit system, and is therefore marked as a valid scheme.

[0186] Step S157: If the comprehensive constraint satisfaction parameter does not reach the preset threshold, select the next initial scheduling adjustment scheme from the initial scheduling adjustment scheme set and re-verify it.

[0187] If the overall constraint satisfaction parameter does not reach the preset threshold, it indicates that the initial scheduling adjustment scheme does not meet the constraint conditions in some aspects, which may pose a risk to the operation of the rail transit system or affect its operational efficiency. In this case, the scheme needs to be abandoned, and the next scheme should be selected from the initial scheduling adjustment scheme set. The scheme should be re-verified according to the process of steps S152 to S155 until a valid scheme that meets the conditions is found.

[0188] Step S158: Sort all valid solutions from high to low according to the comprehensive constraint satisfaction parameter, and select the valid solution ranked first as the final scheduling adjustment solution.

[0189] After verifying all schemes in the initial scheduling adjustment scheme set, all schemes marked as valid are collected. Then, these schemes are sorted from highest to lowest according to their comprehensive constraint satisfaction parameter. The higher the comprehensive constraint satisfaction parameter, the better the scheme performs in satisfying the constraints and the more adaptable it is to the actual operation requirements of the rail transit system.

[0190] The top-ranked effective scheme is selected as the final scheduling adjustment scheme. This final scheduling adjustment scheme is the best overall performance among all effective schemes and can optimize the operation of the rail transit system to the greatest extent.

[0191] Step S159: Convert the final scheduling adjustment plan into the final rail transit scheduling instruction.

[0192] The final scheduling adjustment plan includes a series of parameters and measures that need to be converted into an instruction format that the rail transit scheduling control system can recognize and execute. This conversion process requires refining and standardizing the various parameters and measures in the plan. For example, train departure times are converted into specific time instructions, line resource allocation plans are converted into clear train scheduling instructions, and station passenger flow management measures are converted into specific staff deployment instructions and equipment operation instructions.

[0193] The converted final rail transit dispatch instructions will be sent to the rail transit dispatch control system. Based on these instructions, the dispatch control system will make real-time adjustments to train operation, allocation of line resources, and passenger flow management at stations to ensure that the rail transit system can operate efficiently and safely according to the optimized plan.

[0194] Figure 2The diagram illustrates exemplary hardware and software components of an AI-based rail transit scheduling and analysis system 100 that can implement the ideas of this application, according to some embodiments of this application. For example, a processor 120 can be used in the AI-based rail transit scheduling and analysis system 100 and to perform the functions described in this application.

[0195] For example, the AI-based rail transit scheduling and analysis system 100 may include a network port 110 connected to a network, one or more processors 120 for executing program instructions, a communication bus 130, and various forms of storage media 140, such as a disk, ROM, or RAM, or any combination thereof. Exemplarily, the AI-based rail transit scheduling and analysis system 100 may also include program instructions stored in ROM, RAM, or other types of non-transitory storage media, or any combination thereof. The methods of this application can be implemented according to these program instructions. The AI-based rail transit scheduling and analysis system 100 also includes an I / O interface 150 between the computer and other input / output devices.

[0196] Furthermore, this embodiment of the invention also provides a readable storage medium, wherein computer-executable instructions are preset in the readable storage medium, and when the processor executes the computer-executable instructions, the above-mentioned artificial intelligence-based rail transit scheduling and analysis method is implemented.

[0197] It should be noted that, in order to simplify the description of the present invention and thus help to understand one or more embodiments of the invention, multiple features may sometimes be grouped into one embodiment, drawing or description thereof in the foregoing description of the embodiments of the present invention.

Claims

1. A rail transit scheduling and analysis method based on artificial intelligence, characterized in that, The method includes: A dynamic scheduling baseline for the rail transit system is established, which includes baseline operation characteristics constructed from historical normal scheduling data, rule constraint characteristics of a preset scheduling rule base, and scenario association characteristics of typical operating scenarios. Real-time scheduling data of the rail transit system is collected, and the real-time scheduling data is compared and analyzed with the dynamic scheduling baseline in multiple dimensions to generate a baseline deviation feature set, which includes time dimension deviation features, spatial dimension deviation features and resource dimension deviation features. The pre-trained scheduling coordination evaluation model is invoked to perform dynamic correlation analysis on the baseline deviation feature set, and a scheduling coordination evaluation result is generated. The scheduling coordination evaluation result includes deviation propagation path parameters, coordination conflict probability parameters, and resource adaptability parameters. Based on the scheduling coordination evaluation results, a multi-objective scheduling optimization model is constructed. The deviation propagation path parameters, coordination conflict probability parameters, and resource adaptability parameters are coordinated and optimized through the multi-objective scheduling optimization model to generate an initial set of scheduling adjustment schemes. The initial set of scheduling adjustment schemes is subjected to dynamic constraint verification. After verification, a final rail transit scheduling instruction is generated. The final rail transit scheduling instruction is used to trigger the rail transit scheduling control system to perform scheduling parameter update operations. The multi-objective scheduling optimization model is constructed based on the scheduling coordination evaluation results. This model is used to perform coordinated optimization on deviation propagation path parameters, coordination conflict probability parameters, and resource adaptability parameters, generating an initial set of scheduling adjustment schemes, including: Using the deviation propagation path parameters, coordination conflict probability parameters, and resource adaptability parameters from the scheduling coordination evaluation results as optimization objectives, a multi-objective scheduling optimization model objective function is constructed. The objective function includes the objectives of shortening the deviation propagation path, reducing the coordination conflict probability, and improving resource adaptability. The line capacity constraint rules, train minimum interval rules, and station passenger flow carrying capacity rules are extracted from the preset scheduling rule base as constraints for the multi-objective scheduling optimization model. The path length parameter and node influence intensity parameter in the deviation propagation path parameters are used as input variables for the deviation propagation path shortening target, and the path length shortening weight coefficient is set. The cooperative conflict probability parameter is used as the input variable for the cooperative conflict probability reduction target, and a conflict probability reduction weighting coefficient is set. Use the resource adaptability parameter as the input variable for the resource adaptability improvement target, and set the adaptability improvement weight coefficient; The objective function is solved under constraints using a multi-objective optimization algorithm, generating multiple sets of candidate optimization solutions that include train timetable adjustment parameters, line resource allocation adjustment parameters, and station passenger flow management adjustment parameters. Multiple candidate optimization solutions are sorted using a non-dominated algorithm, and the candidate optimization solutions that satisfy all objective function optimization requirements are selected as the initial scheduling adjustment scheme set.

2. The rail transit scheduling and analysis method based on artificial intelligence according to claim 1, characterized in that, The establishment of a dynamic scheduling baseline for the rail transit system includes: Collect historical normal dispatch data of the rail transit system, which includes historical train operation data, historical line status data, historical passenger flow data, and historical dispatch instruction data; The historical normal scheduling data is divided into multiple typical operational scenario data units based on operational time period characteristics, weather condition characteristics, and holiday attribute characteristics. From each typical operational scenario data unit, train punctuality rate features, interval running time features, station dwell time features, and resource occupancy rate features are extracted, and the extracted features are combined to generate baseline operational features. Obtain a preset scheduling rule base, extract line capacity constraint rules, train minimum interval rules, station passenger flow carrying capacity rules, and emergency scheduling priority rules from the preset scheduling rule base, and convert the extracted rules into rule constraint features; Analyze the correlation between baseline operating features and rule constraint features in each typical operating scenario data unit, calculate the matching degree parameter of the correlation, and based on the matching degree parameter, fuse the baseline operating features, rule constraint features, and scenario identification features of typical operating scenarios to generate a dynamic scheduling baseline.

3. The rail transit scheduling and analysis method based on artificial intelligence according to claim 2, characterized in that, The analysis examines the correlation between baseline operational features and rule constraint features in each typical operational scenario data unit, calculates a matching degree parameter for the correlation, and then fuses the baseline operational features, rule constraint features, and scenario identification features of typical operational scenarios based on the matching degree parameter to generate a dynamic scheduling baseline, including: The feature vectors of baseline operating features and rule constraint features are extracted from data units of typical operating scenarios. After being converted into vector representations of the same dimension, the cosine similarity between the baseline operating feature vector and the rule constraint feature vector is calculated, and the cosine similarity value is used as the initial matching degree parameter. The correlation strength between the train punctuality rate feature in the benchmark operation characteristics and the minimum train interval rule in the rule constraint characteristics is analyzed, and the compliance parameter of punctuality rate with interval rule is calculated. The correlation strength between the resource occupancy rate feature in the baseline operating characteristics and the line capacity constraint rule in the rule constraint characteristics is analyzed, and the compliance parameter of the resource occupancy rate with the capacity rule is calculated. The preliminary matching degree parameter, the on-time rate compliance parameter with the interval rule, and the resource utilization rate compliance parameter with the capacity rule are weighted and summed to generate the comprehensive matching degree parameter. If the overall matching degree parameter is higher than the preset matching threshold, the baseline operation feature, rule constraint feature and the corresponding typical operation scenario identifier feature will be spliced ​​together to generate the scenario baseline unit of the typical operation scenario. All typical operational scenario baseline units are classified and stored according to scenario identification characteristics to build a baseline library for dynamic scheduling baselines; Set the dynamic update cycle of the baseline library, recalculate the comprehensive matching degree parameter of each scenario baseline unit in each update cycle, and replace the scenario baseline units whose comprehensive matching degree parameter is lower than the preset threshold.

4. The rail transit scheduling and analysis method based on artificial intelligence according to claim 1, characterized in that, The real-time scheduling data of the collected rail transit system is compared and analyzed with the dynamic scheduling baseline in multiple dimensions to generate a baseline deviation feature set, including: Real-time dispatch data is collected through the rail transit dispatch monitoring system. The real-time dispatch data includes real-time train location data, real-time line occupancy status data, real-time station passenger flow data, and real-time dispatch instruction execution data. Based on the timestamp information of real-time scheduling data, the corresponding typical operation scenario data unit is matched from the dynamic scheduling baseline to obtain the benchmark operation characteristics and rule constraint characteristics corresponding to the typical operation scenario data unit. The real-time train location data is aligned with the historical train location data in the benchmark operation features on the time axis, and the difference between the real-time arrival time and the benchmark arrival time in the same line section is calculated as the time dimension deviation feature. The real-time line occupancy status data is compared with the line capacity constraint rules in the rule constraint features, and the deviation value between the current line interval occupancy rate and the maximum occupancy rate allowed by the rule is identified as the spatial dimension deviation feature. Real-time station passenger flow data is matched with historical passenger flow data in the baseline operation characteristics by time period. Combined with resource allocation information in real-time dispatch instruction execution data, the adaptation deviation value between the current resource configuration and passenger flow demand is calculated as the resource dimension deviation feature. The time dimension deviation features, spatial dimension deviation features, and resource dimension deviation features are correlated and integrated according to the preset dimension weights to generate a baseline deviation feature set.

5. The rail transit scheduling and analysis method based on artificial intelligence according to claim 4, characterized in that, The process of aligning real-time train location data with historical train location data in the reference operating characteristics along the time axis, and calculating the difference between the real-time arrival time and the reference arrival time within the same line section as the time dimension deviation feature, includes: Extract real-time train location data from real-time dispatch data, and parse the train identification information, line section identification information, and arrival timestamp information from the real-time train location data. Based on train identification information and line section identification information, the corresponding historical train location data is matched from the baseline operating features to extract historical arrival timestamp information. Convert real-time arrival timestamps and historical arrival timestamps into time values ​​in the same time coordinate system; Calculate the difference between the real-time arrival time and the historical arrival time to obtain the time deviation value for a single line section; A sequence analysis of the time deviation values ​​of the same train in continuous track sections is performed to identify the changing trend of the time deviation values ​​and calculate the trend change rate parameter. The time deviation value and trend change rate parameter of a single line interval are combined to generate a time dimension deviation feature, which includes the interval time deviation sequence and the deviation trend change sequence.

6. The rail transit scheduling and analysis method based on artificial intelligence according to claim 1, characterized in that, The pre-trained scheduling collaborative evaluation model is invoked to perform dynamic correlation analysis on the baseline deviation feature set, generating scheduling collaborative evaluation results, including: The baseline deviation feature set is input into the feature preprocessing layer of the scheduling collaborative evaluation model to standardize the time dimension deviation features, spatial dimension deviation features and resource dimension deviation features to generate a standard deviation feature vector. By using the graph structure construction module of the scheduling collaborative evaluation model, a deviation correlation graph structure is constructed with standard deviation feature vectors as nodes and the correlation strength between deviation features of each dimension as edge weights. The path mining algorithm of the scheduling collaborative evaluation model is used to perform path search processing on the deviation correlation graph structure, identify the key transmission path from the initial deviation node to the affected node, and extract the path length parameter, node influence intensity parameter and path transmission efficiency parameter as deviation transmission path parameters. The cumulative impact value of deviation characteristics on each key transmission path is calculated based on the deviation transmission path parameters. The cumulative impact value is then input into the collaborative conflict probability calculation module, and combined with the rule constraint features of the preset scheduling rule base, collaborative conflict probability parameters are generated. The resource adaptability analysis module of the scheduling collaboration evaluation model is invoked to match and analyze the resource dimension deviation characteristics with the resource configuration information in the real-time scheduling data, and calculate the matching ratio parameter between the resource supply and the deviation repair demand as the resource adaptability parameter. The scheduling coordination evaluation results are generated by integrating deviation propagation path parameters, coordination conflict probability parameters, and resource adaptability parameters.

7. The rail transit scheduling and analysis method based on artificial intelligence according to claim 6, characterized in that, The path mining algorithm using the scheduling collaborative evaluation model performs path search processing on the deviation correlation graph structure, identifies the key transmission path from the initial deviation node to the influencing node, and extracts path length parameters, node influence strength parameters, and path transmission efficiency parameters as deviation transmission path parameters, including: Identify all initial deviation nodes from the deviation correlation graph structure. The initial deviation nodes are nodes whose deviation values ​​exceed a preset deviation threshold in the time dimension deviation feature, spatial dimension deviation feature, or resource dimension deviation feature. Starting from each initial deviation node, a depth-first search algorithm is used to search for paths in the deviation association graph structure and record the propagation path from the initial deviation node to all other nodes. Calculate the sum of edge weights for each transmission path and use the sum of edge weights as the path influence strength parameter. The larger the path influence strength parameter, the more significant the transmission influence of the transmission path. Count the number of nodes in each transmission path and use the number of nodes as the path length parameter; The ratio of the path influence intensity parameter to the path length parameter is calculated, and this ratio is used as the path transmission efficiency parameter. The higher the path transmission efficiency parameter, the higher the influence transmission efficiency per unit path length. All transmission paths are sorted from high to low according to the path influence intensity parameter, and a preset number of transmission paths are selected as key transmission paths before sorting. Extract the path length parameter, node influence intensity parameter, and path transmission efficiency parameter for each path from the critical transmission path, and combine them to generate the deviation transmission path parameter.

8. The rail transit scheduling and analysis method based on artificial intelligence according to claim 1, characterized in that, The step of performing dynamic constraint verification on the initial scheduling adjustment scheme set, and generating the final rail transit scheduling instruction after passing the verification, includes: Select an initial scheduling adjustment scheme from the initial scheduling adjustment scheme set, and extract the train timetable adjustment parameters, line resource allocation adjustment parameters, and station passenger flow management adjustment parameters contained in the initial scheduling adjustment scheme; Input the train timetable adjustment parameters into the line simulation module to simulate the adjusted train running path and stopping time, and generate line simulation results. The adjusted parameters for line resource allocation are compared with real-time line occupancy status data to verify whether the adjusted resource allocation meets the line capacity constraint rules and the minimum train interval rules. Input the station passenger flow management adjustment parameters into the station passenger flow simulation module to simulate the passenger flow change trend after the adjustment and verify whether the passenger flow management effect meets the station passenger flow carrying capacity rules. Based on the combined results of line simulation, resource allocation verification, and passenger flow management verification, the comprehensive constraint satisfaction parameter of the initial scheduling adjustment scheme is calculated. If the comprehensive constraint satisfaction parameter reaches the preset threshold, the initial scheduling adjustment scheme is marked as a valid scheme; If the comprehensive constraint satisfaction parameter does not reach the preset threshold, the next initial scheduling adjustment scheme will be selected from the initial scheduling adjustment scheme set and re-verified; All effective solutions are sorted from high to low according to the comprehensive constraint satisfaction parameter, and the effective solution ranked first is selected as the final scheduling and adjustment solution. The final scheduling adjustment plan is converted into the final rail transit scheduling instruction.

9. A rail transit scheduling and analysis system based on artificial intelligence, characterized in that, The system includes a processor and a memory, the memory being connected to the processor. The memory is used to store programs, instructions, or code, and the processor is used to execute the programs, instructions, or code in the memory to implement the artificial intelligence-based rail transit scheduling and analysis method according to any one of claims 1-8.

Images