Online car-hailing cross-platform transport capacity cooperative scheduling method and system

Abstract

The invention relates to the technical field of vehicle scheduling, in particular to an online car-hailing cross-platform transport capacity cooperative scheduling method and system. According to the method, semantic ablation and protocol escape of underlying data can be realized by acquiring parameters such as self-developed interface protocols, data formats and positioning precision of all platforms and building decentralized middleware, and abnormal data can be found and isolated in time by adopting double-link cross check and a multi-stage abnormal data judgment model; the dynamic weight calibration scheduling model is combined, more flexible and accurate transport capacity scheduling is provided based on real-time order data and factors such as positioning accuracy and data transmission stability of each platform, optimal matching between user requirements and optimal platform drivers is realized, and the scheduling efficiency and user experience of the whole system are improved.

Description

Technical Field

[0001] This invention relates to the field of vehicle dispatching technology, and in particular to a method and system for cross-platform collaborative dispatching of ride-hailing capacity. Background Technology

[0002] Cross-platform capacity collaborative scheduling for ride-hailing services breaks down the barriers between the capacity pools and order data of various ride-hailing platforms. Through a unified scheduling hub or collaborative mechanism, drivers, vehicles, and orders from multiple platforms are treated as a global resource pool, enabling real-time supply and demand matching, capacity allocation, and order distribution. This systematic scheduling model aims to improve overall capacity utilization, reduce passenger waiting time, and decrease driver empty-running rates.

[0003] Currently, the most widely applicable technology is cross-platform aggregation scheduling. This involves using a leading aggregation platform as the core, connecting to the transportation resources of various ride-hailing platforms through standardized API interfaces, uniformly accepting user orders, and allocating them to the optimal platform drivers as needed. However, because the interface protocols, data formats, positioning accuracy, and order status definitions of each transportation platform are all self-developed standards without industry-wide uniformity, the aggregation platform's adaptation can only perform surface-level data conversion and cannot solve the underlying data heterogeneity problem. This easily leads to technical issues such as driver positioning deviations, order status synchronization delays, and errors in pick-up distance calculations. Summary of the Invention

[0004] The main objective of this invention is to provide a cross-platform collaborative scheduling method for ride-hailing services, aiming to solve the technical problems in the prior art.

[0005] This invention proposes a cross-platform collaborative scheduling method for ride-hailing services, comprising: The self-developed interface protocols, heterogeneous data formats, positioning accuracy parameters, and order status definition rules of each ride-hailing platform are obtained to build a decentralized heterogeneous data semantic ablation middleware, and the protocol escaping capability and dynamic mapping capability of the middleware are obtained to build a primary and backup dual-link data acquisition channel. The original capacity data and mapped standardized data of each capacity platform are obtained to perform dual-link cross-validation, the validation results are obtained, and a set of suspected abnormal data is obtained based on the validation results. The suspected abnormal data set is input into a preset multi-level abnormal data judgment model to obtain the model output results, and real-time abnormal data, precision abnormal data and platform-level systemic abnormal data are obtained based on the model output results. Based on the real-time abnormal data, accuracy abnormal data, and platform-level systemic abnormal data, perform layered isolation and circuit breaker operations to obtain the isolation and circuit breaker results, and obtain effective standardized capacity data based on the isolation and circuit breaker results; The positioning accuracy level and data transmission stability parameters of each transportation platform are obtained and combined with effective standardized transportation data to construct a dynamic weight calibration scheduling model; The system obtains order feature data based on user-initiated travel orders and inputs the order feature data into a dynamic weight calibration scheduling model to obtain calibration scheduling results and achieve optimal matching between orders and transportation capacity.

[0006] This application also provides a cross-platform ride-hailing capacity collaborative scheduling system, including: The module is used to acquire the self-developed interface protocols, heterogeneous data formats, positioning accuracy parameters and order status definition rules of each ride-hailing capacity platform to build a decentralized heterogeneous data semantic ablation middleware, and to acquire the protocol escaping capability and dynamic mapping capability of the middleware to build a primary and backup dual-link data acquisition channel. The verification module is used to obtain the original capacity data and the mapped standardized data of each capacity platform to perform dual-link cross-verification, obtain the verification results, and obtain a set of suspected abnormal data based on the verification results. The judgment module is used to input the suspected abnormal data set into a preset multi-level abnormal data judgment model, obtain the model output results, and obtain real-time abnormal data, precision abnormal data and platform-level systemic abnormal data based on the model output results. The execution module is used to perform layered isolation and circuit breaker operations based on the real-time abnormal data, accuracy abnormal data and platform-level systemic abnormal data, obtain the isolation and circuit breaker results, and obtain effective standardized transportation capacity data based on the isolation and circuit breaker results; The module is used to obtain the positioning accuracy level and data transmission stability parameters of each transportation platform and combine them with effective standardized transportation data to build a dynamic weight calibration scheduling model; The input module is used to obtain order feature data based on the user-initiated travel order, and input the order feature data into the dynamic weight calibration scheduling model to obtain the calibration scheduling result to complete the optimal matching of order and capacity.

[0007] The present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described cross-platform collaborative scheduling method for ride-hailing services.

[0008] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-described cross-platform capacity collaborative scheduling method for ride-hailing services.

[0009] The beneficial effects of this invention are as follows: By acquiring parameters such as the self-developed interface protocols, data formats, and positioning accuracy of various platforms and building decentralized middleware, this invention can achieve semantic ablation and protocol escaping of underlying data. By adopting a dual-link cross-validation and multi-level abnormal data judgment model, it can promptly detect and isolate abnormal data. Combined with a dynamic weight calibration scheduling model, based on real-time order data and factors such as the positioning accuracy and data transmission stability of various platforms, it provides more flexible and accurate capacity scheduling, realizes the optimal matching of user needs with the best platform drivers, and improves the overall system scheduling efficiency and user experience. Attached Figure Description

[0010] Figure 1 This is a schematic diagram of a method flow according to an embodiment of the present invention.

[0011] Figure 2 This is a schematic diagram of the system structure according to an embodiment of the present invention.

[0012] Figure 3 This is a schematic diagram of the internal structure of a computer device according to an embodiment of this application.

[0013] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0014] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0015] like Figure 1 As shown, this application provides a method for cross-platform collaborative scheduling of ride-hailing capacity, including: S1. Obtain the self-developed interface protocols, heterogeneous data formats, positioning accuracy parameters and order status definition rules of each ride-hailing capacity platform to build a decentralized heterogeneous data semantic ablation middleware, and obtain the protocol escaping capability and dynamic mapping capability of the middleware to build a primary and backup dual-link data acquisition channel. S2. Obtain the original capacity data and mapped standardized data of each capacity platform to perform dual-link cross-validation, obtain the validation results, and obtain a set of suspected abnormal data based on the validation results; S3. Input the suspected abnormal data set into a preset multi-level abnormal data judgment model to obtain the model output results, and obtain real-time abnormal data, precision abnormal data and platform-level systemic abnormal data based on the model output results; S4. Perform a layered isolation circuit breaker operation based on the real-time abnormal data, precision abnormal data and platform-level systemic abnormal data to obtain the isolation circuit breaker result, and obtain effective standardized transportation capacity data based on the isolation circuit breaker result; S5. Obtain the positioning accuracy level and data transmission stability parameters of each transportation platform and construct a dynamic weight calibration scheduling model by combining effective standardized transportation data; S6. Obtain order feature data based on the user-initiated travel order, and input the order feature data into the dynamic weight calibration scheduling model to obtain the calibration scheduling result to complete the optimal matching of order and transportation capacity.

[0016] As described in steps S1-S6 above, the core principle of the existing cross-platform aggregation scheduling technology is that users initiate ride-hailing requests through the aggregation platform. The platform simultaneously sends order requests to each connected ride-hailing platform, collects the responses and location information of drivers from each platform, and then calculates the optimal matching degree of drivers from each platform based on multi-dimensional indicators such as real-time driver location, order-taking distance, road conditions, historical fulfillment rate, and platform commission. The platform prioritizes selecting the driver with the highest matching degree to accept the order. When the original platform driver's response timeout or there is no available capacity, the order is automatically assigned to an idle driver on another platform. At the same time, the platform is linked to complete closed-loop processes such as trip settlement and service evaluation. The supply and demand status of each platform is monitored in real time through a grid heat map, and cross-platform drivers are dynamically guided to flow to high-demand areas to alleviate local capacity shortages. After achieving the optimal matching of orders and transportation capacity, it is necessary to conduct source tracing analysis on the abnormal data of the isolation circuit breaker to obtain the type of abnormal data, the transportation platform to which it belongs, and the specific triggering reason. Based on the abnormal data type and triggering reason, targeted protocol escaping optimization instructions, field mapping optimization instructions, and positioning calibration optimization instructions are generated. Various optimization instructions are encoded and integrated to generate standardized optimization feedback instructions. The association between optimization feedback instructions and the corresponding middleware sub-nodes of the transportation platform is established. The optimization feedback instructions are sent to the corresponding middleware sub-nodes through an encrypted transmission channel, and the instruction sending time and content are recorded to form a feedback instruction log. The middleware receives optimization feedback instructions through its sub-nodes and parses out protocol escaping adjustment requirements, field mapping adjustment requirements, and positioning calibration adjustment requirements based on the instructions. It updates the parsing rules of the protocol escaping engine based on the protocol escaping adjustment requirements, improves the matching relationship of the dynamic field mapping library based on the field mapping adjustment requirements, optimizes the calibration algorithm of the positioning data calibration module based on the positioning calibration adjustment requirements, and updates the state matching rules of the global state semantic mapping module. The updated middleware is then applied to the subsequent heterogeneous data processing process to achieve adaptive and accurate ablation of heterogeneous data from various capacity platforms. It continuously collects subsequent abnormal data and repeatedly executes the feedback optimization process, forming a cross-platform capacity collaborative scheduling closed loop of data processing, anomaly judgment, isolation circuit breaking, scheduling matching, and feedback optimization, thereby achieving continuous iterative optimization of the scheduling method. This invention establishes a decentralized heterogeneous data semantic ablation middleware by standardizing and converting self-developed protocols, data formats, and order status definitions across different platforms. It solves the problems of inconsistent data formats and standards between different platforms in existing technologies. By providing protocol escaping and dynamic mapping capabilities, it eliminates semantic differences between platforms, avoiding problems such as positioning deviations and order status asynchronization caused by inconsistent interface adaptations. This improves the efficiency and accuracy of multi-platform data interaction. Through dual-link cross-validation technology, it can effectively compare and verify the consistency and accuracy of data from different platforms, avoiding the reliability issues of a single data source. During the verification process, it can not only identify data-level anomalies (such as errors in pick-up distance calculation) but also... By comparing standardized data, potential data anomalies are discovered, providing a precise basis for subsequent anomaly data judgment, improving data quality, reducing decision-making errors caused by single platform errors, and enhancing the accuracy and reliability of the entire system. The multi-level anomaly data judgment model can perform in-depth analysis of suspected anomaly data from different dimensions and levels, distinguishing different types of anomalies, such as immediate anomalies, precision anomalies, and platform-level systemic anomalies. This avoids the limitations of traditional methods that only perform simple data screening. By identifying different types of anomalies, the nature and source of data anomalies can be determined more accurately, thereby reducing misjudgments and omissions, improving the accuracy of anomaly data processing, enhancing the system's ability to handle complex anomalies, and helping to ensure the accuracy and reliability of the final data. By implementing layered isolation and circuit breaker operations to manage different types of abnormal data, it is possible to quickly isolate abnormal data that affects system performance, preventing it from interfering with subsequent scheduling decisions. This avoids the crude approach of treating all abnormal data the same way in traditional technologies, helping to precisely control the scope of the impact of abnormal data, ensuring the stability of system operation, and retaining effective standardized data to ensure that data quality is not affected by abnormal data. This effectively improves the accuracy and efficiency of capacity scheduling. By combining factors such as positioning accuracy and data transmission stability, and adopting a dynamic weight calibration scheduling model, it can dynamically adjust scheduling priorities according to the real-time status of the capacity platform, and optimize resource allocation based on the real-time status of different platforms. This system improves the intelligence of cross-platform scheduling by avoiding scheduling errors caused by data inaccuracy or transmission delays. It enables the system to automatically adapt and optimize across different platforms, enhancing the robustness and accuracy of the scheduling process. By optimally matching users' travel needs with transportation resources, it can provide users with accurate travel services based on real-time calibrated scheduling results. By combining the actual situation of transportation capacity on each platform with a dynamic weight calibration scheduling model, it ensures the accuracy and timeliness of order matching, avoiding service delays or inaccuracies caused by uneven resource allocation or insufficiently intelligent matching algorithms in existing technologies. This improves user experience and system scheduling efficiency, ensuring the best match between user needs and transportation resources.

[0017] In one embodiment, step S1, which involves obtaining the self-developed interface protocols, heterogeneous data formats, positioning accuracy parameters, and order status definition rules of each ride-hailing platform to build a decentralized heterogeneous data semantic ablation middleware, includes: S11. Obtain a set of transmission protocol types according to each self-developed interface protocol to build a protocol escaping engine, and obtain field naming rules and numerical unit standards according to each heterogeneous data format to build a dynamic field mapping library. S12. Obtain the positioning method type and coordinate system type according to each of the positioning accuracy parameters to build a positioning data calibration module; obtain the status description text and status transition logic according to each of the order status definition rules to build a global status semantic mapping module. S13. Modularly integrate the protocol escaping engine, dynamic field mapping library, positioning data calibration module and global state semantic mapping module to obtain the integration result; S14. Generate a decentralized heterogeneous data semantic ablation middleware based on the integration results, and configure dedicated middleware sub-nodes for each transportation platform.

[0018] As described in steps S11-S14 above, the above integration mode is achieved by adopting a globally unified architecture combined with dedicated adaptation nodes for each capacity platform. The core is to build a globally unified decentralized heterogeneous data semantic ablation middleware main architecture. This main architecture has built-in basic capabilities such as a general protocol escaping engine, a dynamic field mapping library, a positioning data calibration module, and a global state semantic mapping module. At the same time, each capacity platform is configured with a dedicated middleware sub-node. Based on the basic capabilities of the main architecture, the sub-node loads the dedicated adaptation rules corresponding to the self-developed standard of the corresponding capacity platform (such as dedicated protocol escaping rules, dedicated field mapping relationships, dedicated positioning calibration parameters, and dedicated state semantic mapping logic), so as to achieve the integration effect of one main architecture adapting to the whole platform and one sub-node corresponding to one platform. This invention achieves unified processing of different transmission protocols by acquiring the self-developed interface protocols of each platform and establishing a protocol escaping engine, avoiding transmission failures or data packet format errors during cross-platform data access. Through a dynamic field mapping library, synonymous fields from different platforms can be mapped to unified fields, while standardizing units to avoid distance calculation or positioning errors caused by naming or unit differences. By acquiring the positioning method and coordinate system type, the calibration module can unify the coordinates of different platforms to the same standard coordinate system, directly solving the driver positioning deviation problem. Different platforms have heterogeneous descriptions of order status. By mapping the status of each platform to a unified semantic, the aggregated scheduling can accurately identify the order flow status in real time, thereby reducing synchronization delay problems. The calibrated positioning data and unified status data can be directly used for underlying algorithms such as pick-up distance calculation, route planning, and driver matching, improving scheduling accuracy and solving the technical limitations that traditional surface conversion cannot provide. Modular integration ensures collaborative work among modules, forming an end-to-end heterogeneous data processing chain. This avoids interface misalignment or data semantic mismatch caused by isolated module operation. Through modular design, adding or modifying platform interfaces does not require redeveloping all functions; only specific modules need to be upgraded, enabling flexible technical expansion and directly solving the problem of unifying heterogeneous data. The integrated data undergoes protocol escaping, field standardization, coordinate calibration, and status unification processing, ensuring semantic consistency of the entire data flow from input to output. This significantly reduces positioning errors and order status asynchronization issues. By configuring independent sub-nodes for each platform, protocol escaping, field mapping, coordinate calibration, and status semantic unification can be completed locally, avoiding latency and conflict issues caused by centralized data processing. Middleware sub-nodes directly process platform-developed data, achieving underlying data ablation rather than the surface mapping of traditional aggregation platforms. This significantly reduces positioning deviation, status latency, and pick-up distance calculation errors. Adding or changing platform interfaces only requires updating the configuration of the corresponding sub-nodes, without affecting other platforms. This achieves technical scalability for cross-platform real-time scheduling while avoiding single points of failure and improving system robustness.

[0019] In one embodiment, step S2, which involves obtaining the original capacity data and mapped standardized data of each capacity platform to perform dual-link cross-validation, obtaining the validation results, and obtaining a set of suspected abnormal data based on the validation results, includes: S21. Based on the original transportation capacity data and the mapped standardized data, obtain the corresponding driver location coordinates, order status information and pick-up distance data to establish a one-to-one correspondence between the two-link data. S22. Based on the one-to-one correspondence of the dual-link data, compare the positioning coordinate deviation, order status synchronization time difference, and pick-up distance difference of the original capacity data and the mapped standardized data to generate data consistency feature information. S23. Determine the verification pass status of a single transport capacity data based on the data consistency feature information, and summarize multiple verification pass statuses to obtain the dual-link cross-verification result; S24. Compare the dual-link cross-validation results with a preset threshold, mark the capacity data that exceeds the preset threshold as suspected abnormal data, and summarize all suspected abnormal data to generate a suspected abnormal data set.

[0020] As described in steps S21-24 above, by setting positioning deviation threshold, status delay threshold and distance error threshold, and comparing the positioning coordinate difference, order status synchronization time difference and pick-up distance calculation difference with the corresponding threshold, the capacity data exceeding the threshold range is marked as suspected abnormal data, and a suspected abnormal data set is generated based on the marking results. This invention matches raw capacity data with standardized data to ensure effective mapping of driver location, order status, and pick-up distance data from different platforms, providing a reliable foundation for subsequent data comparison and verification. By constructing a one-to-one correspondence between two links, it overcomes differences in data formats and protocols between platforms, reducing error handling and transmission problems caused by inconsistent data formats. By matching driver location coordinates, order status information, and pick-up distance data, it achieves data fusion between different platforms, providing a unified input format for subsequent verification steps and further reducing errors caused by data differences between platforms. By comparing the positioning deviation, order status synchronization time difference, and pick-up distance difference between raw capacity data and standardized data, it can directly reveal positioning errors, status synchronization delays, and distance calculation deviations caused by differences in platform protocols. Through comparative analysis, it can accurately identify temporal or spatial inconsistencies between platforms during data transmission, thus providing clear data consistency characteristic information. The generation of data consistency characteristic information helps to quickly identify data deviations, enabling the aggregation platform to promptly detect potential problems caused by data heterogeneity, such as inaccurate positioning, status lag, and distance errors, thereby providing a scientific basis for subsequent data verification. By combining data consistency feature information to determine the verification status of individual data, automated cross-platform data consistency verification is achieved, reducing manual intervention and improving system reliability and efficiency. By aggregating multiple data verification statuses, batch verification of large amounts of capacity data can be performed to ensure the accuracy of the overall system. By setting preset thresholds, abnormal data exceeding the thresholds is automatically identified, avoiding omissions or errors in human identification. Through threshold-based automated anomaly detection, data anomalies can be detected and marked in real time. By filtering suspected abnormal data sets, data errors caused by platform protocol differences can be quickly identified, preventing these abnormal data from flowing into subsequent scheduling and allocation stages. By timely marking and processing suspected abnormal data, the system can effectively prevent abnormal data from causing more instability in subsequent processes, ensuring the reasonable scheduling and allocation of overall capacity resources and improving the stability and reliability of the aggregation platform.

[0021] In one embodiment, step S3, which involves inputting the suspected abnormal data set into a preset multi-level abnormal data judgment model to obtain the model output results, and acquiring real-time abnormal data, precision abnormal data, and platform-level systemic abnormal data based on the model output results, includes: S31. Build a first-level real-time lightweight judgment module, a second-level precision calibration judgment module and a third-level global correlation judgment module, and connect the first-level real-time lightweight judgment module, the second-level precision calibration judgment module and the third-level global correlation judgment module in a hierarchical manner to generate a preset multi-level abnormal data judgment model. S32. Input the suspected abnormal data set into the first-level real-time lightweight judgment module to obtain positioning instantaneous jump data, status timeout delay data and distance deviation data, and summarize them to obtain real-time abnormal data. S33. Input the remaining suspected abnormal data after being judged by the first-level real-time lightweight judgment module into the second-level precision calibration judgment module to obtain low-precision no-update data, road network deviation data and state contradiction data, and summarize them to obtain precision abnormal data. S34. Input the remaining suspected abnormal data and the batch data characteristics of each transportation platform after the second-level precision calibration judgment module to the third-level global association judgment module to obtain the platform batch abnormal data and middleware mapping error data, and summarize them to obtain the platform-level systemic abnormal data.

[0022] As described in steps S31-S34 above, this invention establishes multi-level judgment modules and connects them hierarchically, enabling customized processing for different types of abnormal data. This effectively avoids the impact of data heterogeneity in traditional cross-platform aggregation scheduling, such as positioning accuracy deviations and order status delays caused by differences in platform protocols and data formats. The first-level module is responsible for real-time rapid anomaly judgment, the second-level module is responsible for accuracy calibration, and the third-level module focuses on global correlation judgment, thereby improving the accuracy and efficiency of data judgment, enhancing the classification and processing capabilities of abnormal data, reducing error accumulation in the system, and optimizing data compatibility between platforms. The first-level real-time lightweight judgment module focuses on rapid detection of real-time data, which can immediately detect positioning deviations (such as instantaneous positioning jumps) and status timeouts caused by network latency or different data formats during order flow between platforms. By capturing these abnormal data in real time, the module can detect problems and take measures in the early stages, avoiding subsequent data processing errors caused by delays and errors, improving the system's response speed and processing capabilities to abnormal situations, and effectively improving the real-time issues in cross-platform data interaction. The secondary precision calibration judgment module uses precision calibration technology to further screen and process abnormal data that is not clear after the primary judgment. By calibrating data without updates and road network deviation data, it can significantly solve the precision error problem caused by different positioning accuracy and data update frequency differences among platforms in cross-platform scheduling. The accurate judgment of precision abnormal data can effectively reduce the system error caused by data mismatch between platforms, providing more accurate basic data for subsequent global correlation judgment and improving the overall system accuracy. The tertiary global correlation judgment module performs global correlation analysis on the batch data characteristics of each platform, which can identify systemic problems caused by data platform inconsistency, such as batch data anomalies or middleware mapping errors. By handling the correlation between platforms from a global perspective, it can effectively solve large-scale system errors caused by platform interface differences, reduce the impact of erroneous data on the operation of the entire system, and improve the uniformity and accuracy of data.

[0023] In one embodiment, step S4, which involves performing a tiered isolation circuit breaker operation based on the real-time abnormal data, accuracy abnormal data, and platform-level systemic abnormal data to obtain the isolation circuit breaker result, and then obtaining valid standardized capacity data based on the isolation circuit breaker result, includes: S41. Create an independent abnormal data isolation pool based on the real-time abnormal data, and establish scheduling permission blocking rules based on the independent abnormal data isolation pool. Import the real-time abnormal data into the independent abnormal data isolation pool and block its scheduling participation permission. S42. Based on the accuracy anomaly data, obtain positioning deviation characteristics and data validity information to perform road network fitting calibration operation, and filter out the calibration-compliant and non-compliant accuracy anomaly data; S43. Calculate the percentage of abnormal data on a single platform based on the platform-level systemic abnormal data, set the circuit breaker trigger threshold based on the percentage of abnormal data, determine whether a single platform has reached the circuit breaker condition, and execute the corresponding temporary circuit breaker operation. S44. Based on the above-mentioned isolation, calibration, and circuit breaker actions, generate a hierarchical isolation and circuit breaker result, and filter out the capacity data that were not isolated, calibrated to standard, and not circuit breaker based on the hierarchical isolation and circuit breaker result; S45. Perform a second verification based on the selected capacity data and standardized data verification rules, and obtain valid standardized capacity data based on the results of the second verification.

[0024] As described in steps S41-S45 above, this invention effectively prevents unreliable data from entering the main dispatch system by independently isolating abnormal data, thereby avoiding erroneous results during order allocation and driver dispatch. After isolating immediate abnormal data, other normal data can be dispatched according to predetermined rules, avoiding interference from invalid or erroneous data to the dispatch algorithm and improving the system's dispatch accuracy. By acquiring positioning deviation characteristics and performing road network fitting calibration, deviations caused by positioning errors can be accurately calibrated, ensuring accurate geographic data input into the system, thereby reducing pick-up distance errors caused by insufficient positioning accuracy. By filtering calibrated data that meets or does not meet the standards, it ensures that only data that meets the standards enters subsequent capacity dispatch and... In the order allocation process, to avoid low-precision data affecting system efficiency and accuracy, precision calibration can improve data consistency across the capacity platform, ensuring that data from different platforms are processed using the same standards. This effectively solves the underlying data heterogeneity problem. Through statistical analysis of platform-level abnormal data, the source platform of data problems can be identified, providing a basis for subsequent temporary circuit breakers. This effectively addresses data inconsistencies across different platforms, preventing system-wide failures caused by problems on a specific platform. After setting a circuit breaker threshold, when the proportion of abnormal data from a single platform reaches a predetermined value, a temporary circuit breaker operation is automatically performed, preventing the problem on that platform from spreading globally. This effectively reduces the probability of system-wide failures and ensures a better user experience. By aggregating the results of layered isolation and circuit breaker, valid data that is not isolated, meets calibration standards, and has not been circuit-breaker-broken can be accurately screened out. This effectively eliminates abnormal data, improves overall data quality and scheduling efficiency, and ensures that only high-quality, accurate data is used for final scheduling. This reduces scheduling errors caused by data anomalies and ensures the best service experience for users. The aggregated result is the product of multi-platform data merging, which can effectively solve the problem of inconsistent data standards across different platforms. It provides more accurate and effective capacity data for cross-platform aggregated scheduling. The secondary verification is a final check on the screened valid capacity data to ensure that it meets standardization requirements. This can effectively avoid omissions or deviations in the previous calibration and screening steps and ensure the high quality of the final data. The use of standardized data verification rules can unify the data format of different platforms, solve the problem of inconsistent underlying data caused by differences in self-developed standards between platforms, and ensure that cross-platform scheduling can proceed smoothly.

[0025] In one embodiment, step S5, which involves obtaining the positioning accuracy level and data transmission stability parameters of each transportation platform and constructing a dynamic weight calibration scheduling model based on effective standardized transportation data, includes: S51. Obtain the valid identifier and accuracy classification feature of the positioning data according to the positioning accuracy level, and establish the positioning dimension weight assignment rule according to the valid identifier and accuracy classification feature. S52. Obtain data packet loss records and delay fluctuation characteristics based on the data transmission stability parameters, and establish transmission dimension weight assignment rules based on the packet loss records and delay fluctuation characteristics; S53. Obtain driver pick-up capability characteristics and historical performance service characteristics based on the effective standardized transportation capacity data, and establish transportation capacity dimension weight assignment rules based on the driver pick-up capability characteristics and performance service characteristics. S54. Based on the positioning dimension weight assignment rules, transmission dimension weight assignment rules and capacity dimension weight assignment rules, a multi-dimensional weight dynamic adaptation framework is built, and the effective standardized capacity data is input into the multi-dimensional weight dynamic adaptation framework to generate a dynamic weight calibration scheduling model.

[0026] As described in steps S51-S54 above, this invention, through precise positioning accuracy level classification and feature identification, can effectively distinguish the positioning accuracy of each platform, ensuring that only positioning data meeting certain accuracy requirements is used for scheduling, thereby reducing the negative impact of positioning deviation. Through the extraction and rule formulation of accuracy grading features, dynamic weighting can be assigned to the positioning accuracy of different platforms, ensuring that the system can flexibly respond to differences between platforms in multi-platform scheduling and optimizing overall scheduling performance. By establishing weighting rules for the transmission dimension, the system can identify and dynamically adjust for data packet loss and latency fluctuations, ensuring stable data transmission quality and improving the response speed and stability of the scheduling system. By formulating adaptation rules based on data transmission characteristics, dynamic calibration can be performed for fluctuations in different platforms and network environments, avoiding scheduling errors caused by unstable data transmission. Through weighting rules based on pick-up capability and fulfillment service characteristics, the system can accurately assess driver performance. By leveraging the capabilities and fulfillment records of drivers, the system ensures that each order is assigned to the most suitable driver, thereby improving overall service quality and fulfillment rate. By extracting and standardizing these features, the system achieves dynamic cross-platform adaptation, enhancing the dispatch system's intelligent decision-making capabilities when processing driver data from different platforms and reducing operational losses due to improper driver matching. Through a multi-dimensional weighted dynamic adaptation framework, the system can simultaneously consider multiple dimensions such as positioning accuracy, transmission stability, and capacity, and dynamically adapt and calibrate according to the characteristics of different platforms. This breaks down the limitations of cross-platform data heterogeneity, achieving efficient data fusion. By combining multi-dimensional features, dynamic adjustment and weight optimization are achieved, enabling more accurate cross-platform dispatch decisions. When data or conditions change between platforms, the framework can adapt and adjust weights in real time, ensuring the accuracy and flexibility of dispatch decisions. This avoids the problem of traditional methods only performing superficial data transformation, improving the intelligence and accuracy of cross-platform dispatch.

[0027] In one embodiment, step S6, which involves obtaining order feature data based on user-initiated travel orders and inputting the order feature data into a dynamic weight calibration scheduling model to obtain calibration scheduling results and achieve optimal matching between orders and capacity, includes: S61. Obtain the departure coordinates, destination coordinates, order initiation time, and vehicle usage type based on the user's travel order, and summarize the departure coordinates, destination coordinates, order initiation time, and vehicle usage type to generate order feature data; S62. Input the order feature data into the dynamic weight calibration scheduling model, output the comprehensive scheduling score of each driver, and sort the drivers' capacity according to the comprehensive scheduling score from high to low to obtain the capacity ranking table. S63. Select the driver with the highest rating as the optimal matching capacity, and simultaneously push the order matching information to the driver's terminal and user terminal corresponding to the optimal matching capacity. S64. Obtain the driver's order acceptance response. If the driver successfully accepts the order, the order allocation is completed. If the driver rejects the order, the capacity is removed and the next best matching capacity is selected from the capacity sorting table until the order matching is completed.

[0028] As described in steps S61-S64 above, this invention constructs a unified and structured order data representation by simultaneously acquiring multi-dimensional order features. This helps eliminate the heterogeneity problem of raw order data from different platforms, aggregating scattered order information to generate order feature data, providing complete input for subsequent dynamic weight calculation. This avoids the problem of missing information or incorrect mapping caused by differences in interface protocols and data formats in traditional aggregation scheduling. Precise geographic coordinates and time information can be directly used for high-precision distance calculation and estimated arrival time assessment, thereby reducing driver positioning deviation and pick-up distance error. This achieves a technical avoidance of the underlying data heterogeneity problem. The dynamic weight calibration model can automatically adjust the scoring rules according to different order features, achieving accurate quantification of the differences in driver capacity across different platforms. This solves the problem that traditional aggregation platforms cannot accurately compare the capacity of different platforms due to differences in interfaces and standards. By ranking through comprehensive scheduling scores, scheduling decisions no longer rely solely on surface data transformation. Intelligent computing based on underlying features directly mitigates scheduling errors caused by positioning deviations and status synchronization delays. The capacity ranking table provides a direct basis for automatically selecting the optimal driver, avoiding suboptimal matching caused by platform priority or simple rule allocation in traditional models, thus improving matching accuracy and efficiency. Optimal selection based on comprehensive scheduling scores achieves accurate matching calculated from underlying data, directly reducing pick-up distance errors and scheduling delays. The dual-end synchronous push mechanism ensures consistency of order status between drivers and users, helping to solve the status synchronization delay problem caused by platform heterogeneity in traditional aggregated scheduling. By introducing an order acceptance feedback loop, the scheduling system can dynamically correct matching results in real-time response, thereby reducing scheduling failures caused by driver order rejections or unavailable capacity. Through a cyclical iterative selection mechanism, optimal matching of orders and capacity can still be achieved in heterogeneous platform environments, overcoming the technical shortcomings of traditional aggregated platforms in failing to achieve accurate scheduling under uncontrollable factors on the driver side.

[0029] like Figure 2 As shown, this application also provides a cross-platform ride-hailing capacity collaborative scheduling system, including: The module is used to acquire the self-developed interface protocols, heterogeneous data formats, positioning accuracy parameters and order status definition rules of each ride-hailing capacity platform to build a decentralized heterogeneous data semantic ablation middleware, and to acquire the protocol escaping capability and dynamic mapping capability of the middleware to build a primary and backup dual-link data acquisition channel. The verification module is used to obtain the original capacity data and the mapped standardized data of each capacity platform to perform dual-link cross-verification, obtain the verification results, and obtain a set of suspected abnormal data based on the verification results. The judgment module is used to input the suspected abnormal data set into a preset multi-level abnormal data judgment model, obtain the model output results, and obtain real-time abnormal data, precision abnormal data and platform-level systemic abnormal data based on the model output results. The execution module is used to perform layered isolation and circuit breaker operations based on the real-time abnormal data, accuracy abnormal data and platform-level systemic abnormal data, obtain the isolation and circuit breaker results, and obtain effective standardized transportation capacity data based on the isolation and circuit breaker results; The module is used to obtain the positioning accuracy level and data transmission stability parameters of each transportation platform and combine them with effective standardized transportation data to build a dynamic weight calibration scheduling model; The input module is used to obtain order feature data based on the user-initiated travel order, and input the order feature data into the dynamic weight calibration scheduling model to obtain the calibration scheduling result to complete the optimal matching of order and capacity.

[0030] It should be noted that each module and unit in the cross-platform ride-hailing capacity collaborative scheduling system corresponds one-to-one with the steps in the cross-platform ride-hailing capacity collaborative scheduling method.

[0031] like Figure 3 As shown, this application also provides a computer device, which can be a server, and its internal structure can be as follows: Figure 3 As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores all data required for the cross-platform ride-hailing capacity collaborative scheduling method. The network interface is used for communication with external terminals via network connection. The computer program is executed by the processor to implement the cross-platform ride-hailing capacity collaborative scheduling method.

[0032] Those skilled in the art will understand that Figure 3The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer equipment on which the present application is applied.

[0033] An embodiment of this application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements any of the above-described methods for cross-platform collaborative scheduling of ride-hailing capacity.

[0034] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in this application and in the embodiments can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual-speed SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0035] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, apparatus, article, or method. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, apparatus, article, or method that includes that element.

[0036] The above description is merely a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for cross-platform capacity coordination and scheduling of online car-hailing, characterized in that, The method comprises the following steps: obtaining self-developed interface protocols, heterogeneous data formats, positioning accuracy parameters and order state definition rules of each online car-hailing platform to build a decentralized heterogeneous data semantic fusion middleware, and obtaining the protocol escape capability and dynamic mapping capability of the middleware to build a primary and backup dual-link data collection channel; obtaining the original transport capacity data and the mapped standardized data of each transport platform to perform dual-link cross verification, obtain the verification result, and obtain a suspected abnormal data set according to the verification result; inputting the suspected abnormal data set into a preset multi-level abnormal data judgment model to obtain a model output result, and obtaining instant abnormal data, precision abnormal data and platform-level systematic abnormal data according to the model output result; performing hierarchical isolation and fuse operation according to the instant abnormal data, precision abnormal data and platform-level systematic abnormal data to obtain an isolation and fuse result, and obtaining effective standardized transport capacity data according to the isolation and fuse result; obtaining the positioning accuracy level and data transmission stability parameters of each transport platform and combining the effective standardized transport capacity data to construct a dynamic weight calibration scheduling model; obtaining order feature data according to a user-initiated travel order, and inputting the order feature data into the dynamic weight calibration scheduling model to obtain a calibration scheduling result to complete the optimal matching of the order and the transport capacity. 2.The method of claim 1, wherein, The step of obtaining the self-developed interface protocols, heterogeneous data formats, positioning accuracy parameters and order state definition rules of each online car-hailing platform to build a decentralized heterogeneous data semantic fusion middleware comprises: obtaining a transmission protocol type set according to each self-developed interface protocol to build a protocol escape engine, and obtaining a field naming rule and a numerical unit standard according to each heterogeneous data format to build a dynamic field mapping library; obtaining a positioning method type and a coordinate system type according to each positioning accuracy parameter to build a positioning data calibration module, and obtaining a state description text and a state transition logic according to each order state definition rule to build a global state semantic mapping module; modularly integrating the protocol escape engine, the dynamic field mapping library, the positioning data calibration module and the global state semantic mapping module to obtain an integrated result; generating a decentralized heterogeneous data semantic fusion middleware according to the integrated result, and configuring a dedicated middleware sub-node for each transport platform. 3.The method of claim 1, wherein, The step of obtaining the original transport capacity data and the mapped standardized data of each transport platform to perform dual-link cross verification, obtain the verification result, and obtain a suspected abnormal data set according to the verification result comprises: obtaining corresponding driver positioning coordinates, order state information and pick-up distance data according to the original transport capacity data and the mapped standardized data to establish a one-to-one correspondence relationship between the dual-link data; comparing the positioning coordinate deviation, order state synchronization time difference and pick-up distance difference of the original transport capacity data and the mapped standardized data according to the one-to-one correspondence relationship between the dual-link data to generate data consistency feature information; determining the verification qualified state of single transport data according to the data consistency feature information, and aggregating a plurality of verification qualified states to obtain a dual-link cross verification result; The results of the dual-link cross-validation are compared with a preset threshold. Capacity data exceeding the preset threshold is marked as suspected abnormal data, and all suspected abnormal data are aggregated to generate a suspected abnormal data set. 4.The method of claim 1, wherein, The steps of inputting the suspected abnormal data set into a preset multi-level abnormal data judgment model, obtaining the model output results, and acquiring real-time abnormal data, precision abnormal data, and platform-level systemic abnormal data based on the model output results include: A first-level real-time lightweight judgment module, a second-level precision calibration judgment module, and a third-level global correlation judgment module are constructed, and the first-level real-time lightweight judgment module, the second-level precision calibration judgment module, and the third-level global correlation judgment module are connected in a hierarchical manner to generate a preset multi-level abnormal data judgment model. The suspected abnormal data set is input into the first-level real-time lightweight judgment module to obtain positioning instantaneous jump data, status timeout delay data and distance deviation data, and the sum is used to obtain real-time abnormal data. The remaining suspected abnormal data after being judged by the first-level real-time lightweight judgment module is input into the second-level precision calibration judgment module to obtain low-precision data without updates, road network deviation data and state contradiction data, and the accuracy abnormal data is obtained by summarizing them. The remaining suspected abnormal data after the secondary precision calibration judgment module and the batch data characteristics of each transportation platform are input into the tertiary global correlation judgment module to obtain the platform batch abnormal data and middleware mapping error data, and then the platform-level systemic abnormal data are obtained by summarizing them. 5.The method of claim 1, wherein, The steps of performing layered isolation and circuit breaker operations based on the real-time abnormal data, accuracy abnormal data, and platform-level systemic abnormal data, obtaining the isolation and circuit breaker results, and acquiring effective standardized capacity data based on the isolation and circuit breaker results include: An independent abnormal data isolation pool is created based on the real-time abnormal data, and a scheduling permission blocking rule is established based on the independent abnormal data isolation pool. The real-time abnormal data is imported into the independent abnormal data isolation pool and its scheduling participation permission is blocked. Based on the accuracy anomaly data, positioning deviation characteristics and data validity information are obtained to perform road network fitting calibration operations, and calibration-compliant and non-compliant accuracy anomaly data are screened out; The platform-level systemic anomaly data is used to calculate the percentage of anomaly data on a single platform, and the circuit breaker trigger threshold is set according to the percentage of anomaly data. The system then determines whether a single platform has reached the circuit breaker condition and executes the corresponding temporary circuit breaker operation. Based on the above-mentioned isolation, calibration, and circuit breaker actions, a tiered isolation and circuit breaker result is generated, and based on the tiered isolation and circuit breaker result, capacity data that has not been isolated, has met the calibration standards, and has not been circuit breaker is selected. The selected capacity data is verified a second time using standardized data verification rules, and valid standardized capacity data is obtained based on the results of the second verification. 6.The method of claim 1, wherein, The steps of obtaining the positioning accuracy level and data transmission stability parameters of each transportation platform and constructing a dynamic weight calibration scheduling model by combining effective standardized transportation data include: Based on the positioning accuracy level, obtain the valid identifier and accuracy classification feature of the positioning data, and establish positioning dimension weight assignment rules based on the valid identifier and accuracy classification feature; Data packet loss records and latency fluctuation characteristics are obtained based on the data transmission stability parameters, and transmission dimension weight assignment rules are established based on the packet loss records and latency fluctuation characteristics. Based on the effective standardized transportation capacity data, driver pick-up capability characteristics and historical performance service characteristics are obtained, and transportation capacity dimension weight assignment rules are established based on the driver pick-up capability characteristics and performance service characteristics. Based on the weight assignment rules for the positioning dimension, transmission dimension, and capacity dimension, a multi-dimensional weight dynamic adaptation framework is built. Effective standardized capacity data is then input into the multi-dimensional weight dynamic adaptation framework to generate a dynamic weight calibration scheduling model.

7. The method of claim 1, wherein, The steps of obtaining order feature data based on user-initiated travel orders and inputting the order feature data into a dynamic weight calibration scheduling model to obtain calibration scheduling results to achieve optimal matching between orders and transportation capacity include: Based on the user's travel order, obtain the departure coordinates, destination coordinates, order initiation time, and vehicle use type, and summarize the departure coordinates, destination coordinates, order initiation time, and vehicle use type to generate order feature data; The order feature data is input into the dynamic weight calibration scheduling model, and the comprehensive scheduling score of each driver is output. The drivers' capacity is then sorted from high to low according to the comprehensive scheduling score to obtain a capacity ranking table. The driver with the highest rating is selected as the optimal matching capacity, and the order matching information is simultaneously pushed to the driver's and user's terminals corresponding to the optimal matching capacity. Obtain the driver's order acceptance response. If the driver successfully accepts the order, the order allocation is completed. If the driver rejects the order, the capacity is removed and the next best matching capacity is selected from the capacity ranking table until the order matching is completed.

8. A cross-platform capacity coordination and scheduling system for online car-hailing, characterized in that, include: The module is used to acquire the self-developed interface protocols, heterogeneous data formats, positioning accuracy parameters and order status definition rules of each ride-hailing capacity platform to build a decentralized heterogeneous data semantic ablation middleware, and to acquire the protocol escaping capability and dynamic mapping capability of the middleware to build a primary and backup dual-link data acquisition channel. The verification module is used to obtain the original capacity data and the mapped standardized data of each capacity platform to perform dual-link cross-verification, obtain the verification results, and obtain a set of suspected abnormal data based on the verification results. The judgment module is used to input the suspected abnormal data set into a preset multi-level abnormal data judgment model, obtain the model output results, and obtain real-time abnormal data, precision abnormal data and platform-level systemic abnormal data based on the model output results. The execution module is used to perform layered isolation and circuit breaker operations based on the real-time abnormal data, accuracy abnormal data and platform-level systemic abnormal data, obtain the isolation and circuit breaker results, and obtain effective standardized transportation capacity data based on the isolation and circuit breaker results; The module is used to obtain the positioning accuracy level and data transmission stability parameters of each transportation platform and combine them with effective standardized transportation data to build a dynamic weight calibration scheduling model; The input module is used to obtain order feature data based on the user-initiated travel order, and input the order feature data into the dynamic weight calibration scheduling model to obtain the calibration scheduling result to complete the optimal matching of order and capacity. 9.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-8 when the computer program is executed by the processor. When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images