Test method and device of online car-hailing violation penalty system, computer equipment and storage medium

Abstract

This application relates to a testing method, apparatus, computer equipment, and storage medium for a ride-hailing violation penalty system. The method includes: generating test data corresponding to different violation types based on a pre-constructed multi-dimensional testing system, whereby the multi-dimensional testing system includes one or more of a road network simulation database, a driver behavior pattern database, and a policy rule database; combining test data from two or more violation types to generate multiple composite violation test cases; inputting these composite violation test cases into the ride-hailing violation penalty system, and penalizing each composite violation test case through the system; collecting the penalty results output by the ride-hailing violation penalty system for each composite violation test case, analyzing the penalty results, and generating test results for the composite violation detection of the ride-hailing violation penalty system. This method can improve the coverage and testing efficiency of composite violation testing for ride-hailing violation penalty systems.

Description

Technical Field

[0001] This application relates to the field of automated testing technology, and in particular to a testing method, apparatus, computer equipment, and storage medium for a ride-hailing violation judgment system. Background Technology

[0002] With the rapid development of intelligent transportation and the ride-hailing industry, ride-hailing violation judgment systems have become a core support for platform compliance operations and ensuring travel order. Against the backdrop of increasingly stringent regulations and increasingly complex violation scenarios, the accuracy and stability of ride-hailing violation judgment systems not only directly affect the legitimate rights and interests of drivers and passengers, but also have a crucial impact on the efficiency of ride-hailing platform compliance operations and the effectiveness of urban traffic management.

[0003] To ensure the reliable operation of the ride-hailing penalty system, existing testing methods generally adopt a combination of on-site road testing and manual verification. For example, testers simulate violations by driving actual vehicles, and manually review order logs, GPS trajectories, and video data to verify the penalty results. Repeated tests are conducted after rule adjustments to verify whether the system functions properly.

[0004] However, existing testing models have significant drawbacks. For example, real-world violation scenarios require on-road testing, resulting in high reproduction costs and poor repeatability; iterative iterations of penalty rules necessitate repeated full-scale testing, which is time-consuming and inefficient; and multi-source data such as GPS trajectories, audio / video recordings, and order logs are difficult to simulate synchronously, failing to fully cover complex violation scenarios. These issues not only lead to insufficient coverage and low testing efficiency in the composite violation testing of ride-hailing violation penalty systems, but also make the tested ride-hailing penalty systems highly susceptible to missed violations and system malfunctions, severely impacting the accuracy and stability of the ride-hailing penalty system's violation judgments. Summary of the Invention

[0005] Therefore, it is necessary to provide a testing method, device, computer equipment, and storage medium for a ride-hailing violation judgment system to address the aforementioned technical issues. This would not only improve the coverage and efficiency of composite violation testing of the ride-hailing violation judgment system, but also accurately verify the composite violation detection capability of the system, effectively avoid missed violations, and ensure the accuracy and operational stability of the system's judgments.

[0006] According to a first aspect of certain exemplary embodiments of this disclosure, a testing method for a ride-hailing violation judgment system is provided, comprising: generating test data corresponding to the violation type based on a pre-constructed multi-dimensional testing system, wherein the multi-dimensional testing system includes one or more of a road network simulation database, a driver behavior pattern database, and a policy rule database; combining test data of two or more violation types to generate multiple composite violation test cases; inputting the multiple composite violation test cases into the ride-hailing violation judgment system, and judging each composite violation test case through the ride-hailing violation judgment system; collecting the judgment results of each composite violation test case output by the ride-hailing violation judgment system, analyzing the judgment results, and generating test results of composite violation detection of the ride-hailing violation judgment system.

[0007] According to certain exemplary embodiments of this disclosure, the road network simulation database is configured with road network node data and road topology information of each city, for generating test data of vehicle driving trajectory violations; and / or, the driver behavior pattern library is configured with feature maps of various driver violations, for generating test data of driver behavior violations; and / or, the policy rule library is configured with ride-hailing regulatory policy clauses of each city, for generating test data of policy violations.

[0008] According to certain exemplary embodiments of this disclosure, test data corresponding to the violation type is generated based on a pre-constructed multi-dimensional testing system, including: generating test data for vehicle trajectory violations based on road network node data and road topology information from a road network simulation database, wherein the test data for vehicle trajectory violations includes trajectory data of illegal detours and trajectory data that does not conform to the constraints of the real road network; and / or, generating test data for driver behavior violations based on a violation behavior feature map from a driver behavior pattern library, wherein the test data for driver behavior violations includes service quality substandard feature data, order-brushing behavior feature data, and distracted driving feature data; and / or, generating test data for policy-related violations based on urban ride-hailing regulatory policy clauses from a policy rule library, wherein the test data for policy-related violations includes order-taking data of non-local license plate vehicles, vehicle range substandard data, and non-compliant vehicle operation data.

[0009] According to certain exemplary embodiments of this disclosure, test data of two or more violation types are combined to generate multiple composite violation test cases, including: selecting target test data of any two or more violation types from the test data of each violation type; inputting multiple target test data into a large language model to obtain multiple composite violation test cases output by the large language model.

[0010] According to certain exemplary embodiments of this disclosure, a testing method for a ride-hailing violation penalty system further includes: generating normal order test data that conforms to the normal operation specifications of ride-hailing services based on a multi-dimensional testing system; constructing multiple normal order test cases based on the normal order test data; inputting multiple composite violation test cases into the ride-hailing violation penalty system; and penalizing each composite violation test case through the ride-hailing violation penalty system, including: mixing multiple normal order test cases with multiple composite violation test cases, inputting them into the ride-hailing violation penalty system, and penalizing each normal order test case and each composite violation test case through the ride-hailing violation penalty system.

[0011] According to certain exemplary embodiments of this disclosure, multiple normal order test cases are mixed with multiple composite violation test cases and input into a ride-hailing violation judgment system. The ride-hailing violation judgment system then judges each normal order test case and each composite violation test case. This includes: after mixing multiple normal order test cases with multiple composite violation test cases, inputting the mixed test cases into the ride-hailing violation judgment system through a distributed test cluster. The distributed test cluster is used to distribute the mixed test cases to multiple test nodes of the ride-hailing violation judgment system, and the multiple test nodes judge the violations in parallel. Each test node of the ride-hailing violation judgment system outputs the judgment results of each normal order test case and each composite violation test case.

[0012] According to certain exemplary embodiments of this disclosure, the penalty results of each composite violation test case output by the ride-hailing violation penalty system are collected, the penalty results are analyzed, and test results of composite violation detection of the ride-hailing violation penalty system are generated. This includes: collecting the penalty results of each normal order test case and each composite violation test case output by the ride-hailing violation penalty system, analyzing the penalty results of each normal order test case and each composite violation test case, and obtaining the detection rate of composite violation test cases, the average response latency of multi-node parallel penalty, and the false judgment rate of normal order test cases. The test results of composite violation detection of the ride-hailing violation penalty system include the detection rate of composite violation test cases, the average response latency of multi-node parallel penalty, and the false judgment rate of normal order test cases.

[0013] According to a second aspect of certain exemplary embodiments of this disclosure, a testing apparatus for a ride-hailing violation judgment system is provided. The apparatus includes: a first generation module, configured to generate test data corresponding to the violation type based on a pre-constructed multi-dimensional testing system, wherein the multi-dimensional testing system includes one or more of a road network simulation database, a driver behavior pattern database, and a policy rule database; a second generation module, configured to combine test data of two or more violation types to generate multiple composite violation test cases; a judgment module, configured to input the multiple composite violation test cases into the ride-hailing violation judgment system and judge each composite violation test case through the ride-hailing violation judgment system; and an analysis module, configured to collect the judgment results of each composite violation test case output by the ride-hailing violation judgment system, analyze the judgment results, and generate test results of composite violation detection of the ride-hailing violation system.

[0014] According to a third aspect of certain exemplary embodiments of the present disclosure, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of any of the above methods.

[0015] According to a fourth aspect of certain exemplary embodiments of the present disclosure, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of any of the methods described above.

[0016] The aforementioned testing method, apparatus, computer equipment, and storage medium for a ride-hailing violation penalty system generate test data corresponding to the violation types based on a pre-constructed multi-dimensional testing system. This multi-dimensional testing system includes one or more of the following: a road network simulation database, a driver behavior pattern database, and a policy rule database. Test data for two or more violation types are combined to generate multiple composite violation test cases. These composite violation test cases are then input into the ride-hailing violation penalty system, which penalizes each composite violation test case. The penalty results of each composite violation test case output by the ride-hailing violation penalty system are collected, analyzed, and the test results for the composite violation detection of the ride-hailing violation system are generated.

[0017] Therefore, by pre-constructing a multi-dimensional testing system that includes one or more of the following: a road network simulation database, a driver behavior pattern database, and a policy rule database, test data corresponding to a single violation type is generated. This data is then combined to generate composite violation test cases and conduct test analysis. This achieves automated testing of the composite violation detection capability of the ride-hailing violation judgment system, generating accurate composite violation detection test results. This solution not only improves the coverage and efficiency of composite violation testing on the ride-hailing violation judgment system but also accurately verifies the system's composite violation detection capability, effectively avoiding missed violation judgments and ensuring the accuracy and operational stability of the system's judgments. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating a testing method for a ride-hailing violation judgment system, as shown in some exemplary embodiments of this disclosure. Figure 2 This is a flowchart illustrating a method for generating multiple compound violation test cases in some exemplary embodiments of this disclosure; Figure 3 This is a flowchart illustrating a method for judging each normal order test case and each compound violation test case through a ride-hailing violation judgment system in some exemplary embodiments of this disclosure; Figure 4 This is a structural block diagram of a test apparatus for a ride-hailing violation judgment system, as shown in some other exemplary embodiments of this disclosure. Figure 5 This is a diagram illustrating the internal structure of a computer device in some other exemplary embodiments of this disclosure. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0020] The following detailed descriptions are provided to aid the reader in gaining a comprehensive understanding of the methods, apparatus, electronic devices, storage media, and / or computer program products described herein. However, upon understanding the disclosure of this disclosure, various changes, modifications, and equivalents of the methods, apparatus, storage media, and / or computer program products described herein will become apparent. For example, the order of operations described herein is merely illustrative and is not limited to those orders set forth herein, but may be changed as will become clear upon understanding the disclosure of this disclosure, except for operations that must occur in a specific order. Furthermore, for clarity and conciseness, descriptions of features known in the art may be omitted.

[0021] The features described herein may be implemented in different forms and should not be construed as limited to the examples described herein. Rather, the examples described herein are provided only to illustrate some of the many feasible ways of implementing the methods, electronic devices, and / or storage media described herein, many of which will become clear upon understanding this disclosure.

[0022] The terminology used herein is for the purpose of describing various examples only and is not intended to limit disclosure. Unless the context clearly indicates otherwise, the singular form is intended to include the plural form as well. The terms “comprising,” “including,” and “having” indicate the presence of the described features, quantities, operations, components, elements, and / or combinations thereof, but do not exclude the presence or addition of one or more other features, quantities, operations, components, elements, and / or combinations thereof. Unless otherwise stated, “ / ” means “or,” for example, A / B can mean A or B; “and / or” in the text is merely a description of the relationship between related objects, indicating that three relationships can exist, for example, A and / or B can mean: A alone, A and B simultaneously, and B alone. Furthermore, in the description of embodiments of the invention, “multiple” means two or more.

[0023] Unless otherwise defined, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure pertains upon understanding this disclosure. Unless expressly defined herein, terms (such as those defined in a general dictionary) shall be interpreted as having a meaning consistent with their meaning in the context of the relevant field and in this disclosure, and shall not be interpreted in an idealized or overly formalistic manner.

[0024] It should be noted that the terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in some of the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0025] Furthermore, in the description of the examples, detailed descriptions of well-known related structures or functions will be omitted when it is believed that such detailed descriptions would lead to a vague interpretation of this disclosure.

[0026] In the following description, embodiments will be described in detail with reference to the accompanying drawings. However, embodiments may be implemented in various forms and are not limited to the examples described herein.

[0027] The abbreviations and key terms in this disclosure are explained as follows: 1. Multi-dimensional testing system: A standardized data support system built for testing ride-hailing violation judgment system. The core includes three types of databases: road network simulation database, driver behavior pattern database, and policy rule database. These databases can be called individually or in combination according to testing needs. They provide basic data and basis at the road network, behavior, and policy levels for generating various types of violation / normal order test data, and are the core foundation for realizing automated testing.

[0028] 2. Composite violation test cases: These are test cases generated by combining test data of two or more single violation types through feature correlation analysis, logical matching, and multi-source fusion. Each test case contains at least two core violation features and can fully reproduce complex violation scenarios in actual ride-hailing business. They are used to verify the penalty system's ability to detect composite violations.

[0029] 3. Distributed test cluster: A test execution architecture consisting of multiple test nodes. It can evenly distribute mixed normal / compound violation test cases to each node through a load balancing algorithm, realize parallel transmission, input and judgment of multiple nodes, effectively improve the processing efficiency of large-scale test cases, and avoid system lag caused by single-node input.

[0030] 4. Road Network Simulation Database: This database stores road network node data (key node coordinates, road names, road grades, etc.) and road topology information (node ​​connection relationships, traffic directions, road distances, etc.) for each city. It can build simulation models based on real urban traffic road networks and provide a real road network reference for generating vehicle driving trajectory violation test data.

[0031] 5. Driver Behavior Pattern Library: A feature database formed by extracting and modeling the characteristics of various violations by ride-hailing drivers. It includes key feature indicators, manifestations and identification standards of violations such as substandard service quality, order fraud, and distracted driving, and is used to generate test data on driver behavior violations that are close to reality.

[0032] 6. Policy and Rule Database: This database is formed by sorting out and standardizing the regulatory policies and provisions for ride-hailing services in various cities. It covers various compliance requirements such as vehicle qualifications, driver operation, and platform services. It can generate policy-related violation test data such as non-local license plate order taking and insufficient range based on local policies, adapting to the differentiated policy testing needs of different cities.

[0033] 7. Composite Violation Detection Rate: This is a core indicator for measuring the ability of a ride-hailing violation judgment system to identify composite violations. It refers to the proportion of composite violation test cases that the system judges as violations to the total number of composite violation test cases. The higher the value, the stronger the system's ability to identify composite violations.

[0034] 8. Normal Order Misjudgment Rate: This is a core indicator for measuring the accuracy of the ride-hailing violation penalty system in judging normal orders. It refers to the proportion of normal order test cases that are incorrectly judged as violations by the system to the total number of normal order test cases. The lower the value, the higher the accuracy of the system in judging compliant orders.

[0035] 9. Average response latency of multi-node parallel judgment: The core indicator for measuring the processing speed of the ride-hailing violation judgment system in large-scale test scenarios. It refers to the average time taken for all test cases in each test node to go from input to output judgment results. The lower the value, the faster the parallel processing efficiency and response speed of the system.

[0036] First, the specific technical problem this application aims to solve is stated. The closest existing technology to this solution is a testing method for ride-hailing violation penalty systems that combines on-site road testing with manual verification. In this existing technology, testers simulate various violations by driving actual vehicles on the road, simultaneously collecting relevant data such as order logs, GPS trajectories, and audio / video recordings. Then, they manually review and verify the data and the penalty system's output. After the penalty rules are iteratively updated, system regression testing is completed through repeated full-scale road testing and manual verification.

[0037] Specifically, existing technology involves deploying actual vehicles and assigning test personnel to conduct offline road tests, simulating single violations such as detours, fraudulent orders, and vehicle mismatch. Multi-source data generated during the testing process is manually recorded and organized. This data is then imported into the ride-hailing violation penalty system, where the accuracy of the system's penalties is manually verified. When city control policies are updated or platform penalty rules are iterated, a full-scale road test must be reorganized, and the above manual process repeated to complete the functional verification of the entire penalty system.

[0038] In summary, the existing technology for testing ride-hailing violation penalty systems, which combines on-site road testing with manual verification, has the following technical problems: 1. High cost of reproducing real violation scenarios: Existing testing solutions rely heavily on real vehicle road tests to simulate violation scenarios, requiring a large investment of offline resources such as vehicles and personnel. The reproduction of test scenarios is not only costly, but also affected by factors such as environment and road conditions, resulting in poor repeatability of test results and difficulty in stably reproducing the same violation test scenarios.

[0039] 2. Low efficiency in testing iterative rules for penalties: When ride-hailing penalty rules are updated or city-specific control policies are adjusted, existing technologies require repeated full-scale testing of the penalty system, without achieving targeted rule verification. The entire testing process is time-consuming and consumes a lot of manpower and resources, making it unsuitable for the testing needs of high-frequency iteration of control rules.

[0040] 3. Lack of multi-source data synchronization simulation capability: Existing testing methods cannot achieve synchronous simulation and collaborative injection of multi-source data such as GPS trajectory, audio and video, order logs, and vehicle qualifications. They can only perform simple tests on single types of data and cannot fully reproduce complex violation scenarios in real business, resulting in incomplete coverage of test scenarios.

[0041] 4. Manual verification of penalty results is prone to errors: Existing technologies rely on manual processing of test data, verification and analysis of penalty results. This not only results in low overall testing efficiency and difficulty in meeting the needs of large-scale test case verification, but also makes it easy for verification errors to occur due to human operation. As a result, the authenticity and reliability of the test results cannot be guaranteed, making the ride-hailing penalty system after testing prone to missed violations and system failures, which seriously affects the accuracy and stability of the ride-hailing penalty system in making violation judgments.

[0042] In some exemplary embodiments of this disclosure, such as Figure 1 As shown, a testing method for a ride-hailing violation penalty system is provided. Taking the application of this method to a server as an example, the method includes the following steps: Step 101: Generate test data corresponding to the violation type based on the pre-built multi-dimensional testing system.

[0043] The multi-dimensional testing system includes one or more of the following: a road network simulation database, a driver behavior pattern database, and a policy and rule database.

[0044] Specifically, the server can select to call one or more databases from the multi-dimensional testing system based on the required violation type to be tested. By extracting pre-set basic data from the database and combining it with the characteristic attributes of various violations, standardized test data for the corresponding violation type is generated. This test data can cover various information related to violations in the ride-hailing business scenario, such as driver service chat logs, order operation records from the platform backend, vehicle GPS driving trajectories, and vehicle qualification parameter information, providing a data foundation for the subsequent generation of composite violation test cases. When generating test data, the server can match the corresponding violation type's data generation logic according to the attribute characteristics of different databases, ensuring that the test data truly reflects the characteristics of violations in the actual ride-hailing business, while also guaranteeing the standardization and composability of the test data, meeting the need for subsequent combinations of multiple types of violation data to generate composite violation test cases.

[0045] Step 102: Combine test data of two or more violation types to generate multiple composite violation test cases.

[0046] Specifically, the server can select two or more test data from the generated single-violation-type test data according to preset combination rules, and combine them to integrate the core violation feature information in the test data of different violation types to generate multiple composite violation test cases. Each composite violation test case contains core features of at least two violation types. For example, driving trajectory test data containing GPS trajectory data of illegal detours can be combined with driver behavior test data containing chat records of negative service reviews to generate a complete composite violation test case. Each composite violation test case has independent identification information, which facilitates the traceability and analysis of the judgment results after being input into the judgment system. At the same time, the server can generate multiple sets of composite violation test cases with different combinations in batches to ensure the coverage of composite violation scenarios by the test cases.

[0047] Step 103: Input multiple composite violation test cases into the ride-hailing violation judgment system, and judge each composite violation test case through the ride-hailing violation judgment system.

[0048] Specifically, the server can input multiple generated composite violation test cases into the ride-hailing violation judgment system to accurately test the system's composite violation detection capabilities. Each composite violation test case contains core features of two or more violation types, comprehensively covering various composite violation scenarios in actual ride-hailing operations. Upon receiving the composite violation test cases, the ride-hailing violation judgment system invokes its own judgment rules and detection algorithms. The judgment rules are based on the ride-hailing regulatory policies, industry compliance standards, and actual operational logic of each city. The detection algorithm is used to accurately analyze the multi-dimensional violation features in the test cases, independently identifying and comprehensively judging various violation feature information in each composite violation test case. The ride-hailing violation judgment system analyzes each test case through the judgment rules and detection algorithm to determine whether a violation exists, what type of violation exists, and whether it constitutes a composite violation. Finally, it generates a corresponding judgment result for each composite violation test case. This judgment result may include the violation judgment conclusion, violation type identification result, and judgment basis. Simultaneously, the judgment system stores the judgment results of all composite violation test cases and feeds them back to the server. Therefore, it can fully trigger the complex violation judgment logic of the ride-hailing violation judgment system, truly restore the actual detection and judgment process of the system for various complex violation scenarios, and obtain real and effective judgment data for subsequent analysis of the system's complex violation detection capabilities.

[0049] Step 104: Collect the penalty results of each composite violation test case output by the ride-hailing violation penalty system, analyze the penalty results, and generate the test results of composite violation detection of the ride-hailing violation penalty system.

[0050] Specifically, the server can collect the penalty results of all composite violation test cases output by the ride-hailing violation penalty system, and uniquely identify and associate each penalty result with the corresponding composite violation test case, forming a complete dataset of corresponding test cases and penalty results. Then, according to preset test analysis rules, the server performs multi-dimensional analysis of this dataset. For example, it verifies the actual violation characteristics of each composite violation test case against the penalty system's judgment conclusion, analyzes the accuracy of the penalty results, the completeness of violation type identification, and the rationality of composite violation judgment. Simultaneously, it can statistically analyze core test indicators such as composite violation detection rate, missed detection rate, and false detection rate. Finally, based on the above analysis and statistical results, it generates standardized test results that intuitively reflect the composite violation detection capability of the ride-hailing violation penalty system. These test results can cover the system's detection performance for different types of composite violation scenarios, core indicator data, and overall detection capability evaluation. Therefore, it is possible to conduct a comprehensive and objective evaluation of the combined violation detection capability of the ride-hailing violation judgment system based on real judgment data, accurately identify the problems and deficiencies of the system in combined violation detection, provide reliable test basis for the optimization and iteration of the ride-hailing violation judgment system, and enable the combined violation detection capability of the ride-hailing violation judgment system to be improved in a targeted manner, effectively reduce the probability of missed or false judgments of violations, and further ensure the accuracy and operational stability of the ride-hailing violation judgment system in practical applications.

[0051] The aforementioned testing method for a ride-hailing violation penalty system generates test data corresponding to a single violation type by pre-constructing a multi-dimensional testing system comprising one or more of the following: a road network simulation database, a driver behavior pattern database, and a policy rule database. This data is then combined to generate composite violation test cases and analyzed to automate the testing of the ride-hailing violation penalty system's composite violation detection capabilities, resulting in accurate composite violation detection test results. This solution not only improves the coverage and efficiency of composite violation testing for ride-hailing violation penalty systems but also accurately verifies the system's composite violation detection capabilities, effectively avoiding missed violation detections and ensuring the accuracy and operational stability of the system's penalties.

[0052] In some exemplary embodiments of this disclosure, based on the above embodiments, it is further explained that the road network simulation database is configured with road network node data and road topology information of each city, used to generate test data for vehicle driving trajectory violations; and / or, the driver behavior pattern library is configured with feature maps of various driver violations, used to generate test data for driver behavior violations; and / or, the policy rule library is configured with ride-hailing regulatory policy clauses of each city, used to generate test data for policy violations.

[0053] Specifically, the road network node data configured in the road network simulation database includes basic information such as the coordinates of key nodes, road names, and road grades of each city's roads. The configured road topology information includes correlation information such as the connection relationships between road nodes, traffic directions, and road distances. These two types of information together construct a road network model that closely matches actual urban traffic, providing a realistic road network reference for the generation of vehicle trajectory-based violation test data and ensuring that such test data highly matches the actual road network scenario. And / or, the violation behavior feature map configured in the driver behavior pattern library is formed by extracting and modeling the characteristics of various driver violations in the ride-hailing business. It includes key feature indicators, feature manifestations, and data identification standards for various violations, providing feature basis for the generation of driver behavior-based violation test data and ensuring that the test data accurately reflects the actual performance of various violations. And / or, the ride-hailing regulatory policy clauses configured in the policy rule library are formed by sorting out and standardizing various regulations for compliant operation of ride-hailing services in various regions. It includes various policy contents such as vehicle qualification requirements, driver operation requirements, and platform service requirements, providing policy basis for the generation of policy-based violation test data. Based on the configuration of each database, the server can generate various types of violation test data that conform to road network characteristics, behavioral characteristics, and policy requirements, ensuring the authenticity and relevance of the test data.

[0054] In certain exemplary embodiments of this disclosure, based on the above embodiments, the specific implementation method of generating test data corresponding to the violation type based on a pre-built multi-dimensional testing system in step 101 is further described. Step 101, namely the step of generating test data corresponding to the violation type based on a pre-built multi-dimensional testing system, specifically may include the following steps: generating test data for vehicle trajectory violations based on road network node data and road topology information from a road network simulation database; the test data for vehicle trajectory violations includes trajectory data of illegal detours and trajectory data that does not conform to the constraints of the real road network; and / or, generating test data for driver behavior violations based on a violation behavior feature map from a driver behavior pattern library; the test data for driver behavior violations includes service quality substandard feature data, order-brushing behavior feature data, and distracted driving feature data; and / or, generating test data for policy-related violations based on urban ride-hailing regulatory policy clauses from a policy rule library; the test data for policy-related violations includes order-taking data for vehicles with non-local license plates, vehicle range substandard data, and non-compliant vehicle operation data.

[0055] Specifically, when the server generates vehicle trajectory violation test data based on the road network simulation database, it can extract road network node data and road topology information to simulate and generate illegal detour trajectory data that deviates from reasonable driving routes and exceeds normal driving distances. It can also generate trajectory data that does not conform to the actual connection relationship of the urban road network and has contradictions in road traffic logic, thus failing to meet the constraints of the real road network. This type of data may include GPS data of vehicle movement. The server generates driver behavior violation test data based on a driver behavior pattern library. This data can include service quality substandard features such as passenger complaint chat logs and low service ratings; fraudulent order creation and abnormal order cancellation records; and distracted driving features such as frequent mobile phone use and unnecessary long stops. When generating policy-related violation test data based on a policy rule library, the server can generate non-local license plate vehicle order acceptance data and vehicle information that does not meet the city's qualification requirements, as well as vehicle range substandard data (electric vehicles with ranges below local operating requirements) and non-compliant vehicle operation data (vehicles with wheelbase, engine displacement, and other parameters not meeting local compliant vehicle requirements). The server can perform these operations individually or in combination to generate test data for one or more violation types, meeting diverse test data needs and laying the foundation for subsequent combined generation of complex violation test cases and accurate testing of the ride-hailing violation judgment system.

[0056] Based on the above embodiments, the specific implementation method of combining test data of two or more violation types to generate multiple composite violation test cases in step 102 is further explained. For example... Figure 2 As shown, step 102 above, which combines test data of two or more violation types to generate multiple composite violation test cases, may specifically include the following steps: Step 201: Select any two or more target test data types from the test data of each violation type.

[0057] Step 202: Input multiple target test data into the large language model to obtain multiple composite violation test cases output by the large language model.

[0058] Specifically, the server can first select any two or more target test data from the generated test data of different violation types, such as vehicle driving trajectory, driver behavior, and policy, based on the test coverage requirements of the ride-hailing violation judgment system. The selected target test data can cover different dimensions of violation features and adapt to the testing requirements of various complex violation scenarios in actual business. Then, the server inputs the selected multiple types of target test data into the large language model and inputs explicit instructions. The core of these instructions is "Based on the provided multiple types of target test data, extract the core violation features of each data, analyze the correlation and logical matching between features, eliminate logically conflicting combinations, and combine them with the actual business scenarios of ride-hailing to logically match and fuse non-conflicting features to generate multiple composite violation test cases that are feature-complete, logically self-consistent, independent, and diverse." This instruction clarifies the processing objectives, processing flow, and output requirements of the large language model, ensuring that the processing of the large language model conforms to the actual needs of ride-hailing violation testing. For example, the instruction can further specify that "each composite violation test case must completely contain all the core violation features of the corresponding composite violation scenario, and each test case must not be duplicated or logically contradictory."

[0059] Furthermore, after receiving the input instructions and multiple target test data, the large language model will complete the data processing flow step by step according to the instructions: First, it will parse each of the multiple input target test data, extracting the core violation feature information from each type of target test data. For example, it will extract features such as driving route, distance from the normal route, and driving time from the data on illegal detour trajectories; extract features such as service complaint content, passenger ratings, and duration of illegal service from the data on substandard service quality; and extract features such as vehicle qualification mismatch and driver overtime from policy-related violation data. Second, it will analyze the feature correlation and logical matching between target test data of different violation types in conjunction with the actual business logic of ride-hailing services. It will focus on determining whether there are business conflicts in the violation features of various target test data, eliminating logically conflicting combinations that do not conform to the actual scenario, and retaining effective combinations of feature correlation relationships that conform to the actual ride-hailing business scenario, ensuring that the subsequently generated composite violation test cases fit the real business scenario. For example, feature correlation analysis can be performed on illegal detour trajectory data (vehicle driving trajectory) and service quality substandard feature data (driver behavior). These two data points have a natural business connection in actual ride-hailing operations; detours easily lead to passenger complaints and lower service ratings. The large language model can identify such reasonable correlations and retain feature combinations. However, if compliant normal driving trajectory data marked as being within the same order's trip is analyzed with the corresponding fraudulent order behavior feature data, fraudulent order behavior typically lacks genuine compliant trip trajectories, creating a direct logical conflict between the two regarding the authenticity of the order's trip. The large language model will determine this type of combination to be invalid and discard it. Finally, according to the input instruction to "generate multiple composite violation test cases," the large language model combines the feature presentation forms of real composite violations, logically matches and fuses the filtered, logically conflict-free target test data from multiple sources, organically combining core violation features from different dimensions, and ultimately outputting multiple composite violation test cases. Each output composite violation test case can fully reflect all the violation characteristics of the corresponding composite violation scenario, while ensuring the independence and diversity of each test case, thereby meeting the comprehensive testing requirements for the composite violation detection capability of the ride-hailing violation judgment system.

[0060] Based on the above embodiments, a testing method for a ride-hailing violation penalty system further includes the following steps: generating normal order test data that conforms to the normal operation specifications of ride-hailing services based on a multi-dimensional testing system, and constructing multiple normal order test cases based on the normal order test data; step 103, which is to input multiple composite violation test cases into the ride-hailing violation penalty system and penalize each composite violation test case through the ride-hailing violation penalty system, may specifically include the following steps: mixing multiple normal order test cases with multiple composite violation test cases, inputting them into the ride-hailing violation penalty system, and penalizing each normal order test case and each composite violation test case through the ride-hailing violation penalty system.

[0061] Specifically, the server can refer to the generation logic of violation type test data, and extract relevant information such as road network driving, driver service, and vehicle qualifications that comply with the normal operation norms of ride-hailing based on the basic data in the road network simulation database, driver behavior pattern database, and policy rule database. This generates normal order test data without any violation characteristics, and multiple normal order test cases are constructed based on this data to fully reproduce the actual compliant operation business scenario of ride-hailing. After generating composite violation test cases and normal order test cases, the server randomly mixes the two types of test cases to simulate the real business scenario where compliant and violation orders randomly occur in actual ride-hailing operations. All the mixed test cases are then input into the ride-hailing violation judgment system. The judgment system performs indiscriminate judgment on each mixed test case, determining both the existence of violations and distinguishing between compliant and composite violation orders. This comprehensively verifies the judgment system's ability to identify composite violations and distinguish normal orders, avoiding the system's tendency to judge only a single violation scenario and improving the authenticity and reference value of the test results. For example, 1000 sets of normal order test cases are generated as a baseline, and 5% of composite violation test cases are injected (such as multiple composite violation patterns including detours and substandard service quality, order fraud and vehicle mismatch, non-local license plates and illegal detours, etc.). The normal order test cases and composite violation test cases are randomly mixed to form a complete test case set. In this embodiment, the injection ratio of composite violation test cases can be determined according to the actual testing needs of the ride-hailing violation judgment system, and can also be flexibly adjusted according to the violation occurrence rate in the actual operation scenario. This embodiment does not limit this.

[0062] In some exemplary embodiments of this disclosure, based on the above embodiments, a further explanation is provided regarding the specific implementation method of mixing multiple normal order test cases with multiple composite violation test cases and inputting them into the ride-hailing violation judgment system, and then using the ride-hailing violation judgment system to judge each normal order test case and each composite violation test case. For example... Figure 3As shown, the process of mixing multiple normal order test cases with multiple compound violation test cases and inputting them into the ride-hailing violation judgment system, and then judging each normal order test case and each compound violation test case through the ride-hailing violation judgment system, can specifically include the following steps: Step 301: Mix multiple normal order test cases with multiple composite violation test cases, and then input the mixed test cases into the ride-hailing violation judgment system through a distributed test cluster.

[0063] The distributed test cluster is used to distribute the mixed test cases to multiple test nodes of the ride-hailing penalty system, and to make penalties in parallel through multiple test nodes.

[0064] Step 302: Each test node of the ride-hailing violation judgment system outputs the judgment results of each normal order test case and each compound violation test case.

[0065] Specifically, the server first randomly mixes multiple normal order test cases with multiple composite violation test cases to form a test case dataset containing different scenarios. This ensures that the mixed test scenarios closely match the order distribution in actual ride-hailing operations, thereby more comprehensively verifying the accuracy of the ride-hailing violation penalty system. Subsequently, the server distributes this mixed test case dataset to a distributed test cluster. This distributed test cluster, based on the computing power load of its individual test nodes, uses a load balancing algorithm to evenly distribute the test case dataset across multiple test nodes within the cluster. Simultaneously, each test node is configured with an independent penalty request channel, effectively avoiding congestion during test case transmission. This enables parallel transmission and synchronous input of different test cases across multiple test nodes, improving the overall input efficiency of test cases and preventing system lag and response delays caused by inputting too many test cases at once.

[0066] When the ride-hailing violation penalty system receives normal order test cases and compound violation test cases from various test nodes, each test node simultaneously initiates an independent penalty process. For each test case, it calls preset penalty rules and specialized detection algorithms to identify whether any violation characteristics exist and whether a compound violation exists. Normal order test cases are accurately determined to be compliant, while compound violation test cases are precisely identified, their various violation types are precisely identified, and the corresponding specific violation results are determined. Simultaneously, each test node categorizes and stores all penalty results for its processed test cases in a unified format and synchronously feeds them back to the distributed test cluster and server, ensuring the completeness, accuracy, and traceability of the penalty results. This implementation achieves efficient parallel penaltying of large-scale mixed test cases, improving testing efficiency, and comprehensively verifies the penalty accuracy of the ride-hailing violation penalty system under mixed normal and violation scenarios, meeting the core requirements of system testing.

[0067] In certain exemplary embodiments of this disclosure, based on the above embodiments, the specific implementation method of collecting the judgment results of each composite violation test case output by the ride-hailing violation judgment system, analyzing the judgment results, and generating the test results of the composite violation detection of the ride-hailing violation judgment system in step 104 is further described. Step 104, namely, the step of collecting the judgment results of each composite violation test case output by the ride-hailing violation judgment system, analyzing the judgment results, and generating the test results of the composite violation detection of the ride-hailing violation judgment system, can specifically include the following steps: collecting the judgment results of each normal order test case and each composite violation test case output by the ride-hailing violation judgment system, analyzing the judgment results of each normal order test case and each composite violation test case to obtain the detection rate of the composite violation test cases, the average response latency of the multi-node parallel judgment, and the false judgment rate of the normal order test cases. The test results of the composite violation detection of the ride-hailing violation judgment system include the detection rate of the composite violation test cases, the average response latency of the multi-node parallel judgment, and the false judgment rate of the normal order test cases.

[0068] Specifically, the server collects all the penalty results of normal order test cases and compound violation test cases output by the ride-hailing violation penalty system. For compound violation test cases, it counts the number of test cases judged as violations by the system and calculates the detection rate of compound violation test cases, reflecting the system's ability to identify compound violations. For example, the compound violation detection rate is 91.5%. For the entire process of multi-node parallel penalty, the server extracts all the time-consuming data of each test case from input to penalty system to output penalty result in each test node, calculates the average of all time-consuming data, and obtains the average response latency of multi-node parallel penalty, reflecting the system's processing speed in a scenario of large-scale test case parallel input. For example, the average response latency of multi-node parallel penalty is 56ms. For normal order test cases, the server verifies the actual compliance status of each test case with the system penalty result, counts the number of test cases incorrectly judged as violations by the system, and calculates the misjudgment rate of normal order test cases, reflecting the accuracy of the system's judgment of normal operating orders. For example, the misjudgment rate of normal order test cases is 1.8%. The server integrates three data points: the detection rate of composite violation test cases, the average response latency of parallel penalties across multiple test nodes, and the false positive rate of normal order test cases, to generate composite violation detection test results for the ride-hailing violation penalty system. Therefore, it can quantitatively analyze the violation penalty detection capabilities of the ride-hailing violation penalty system from three core dimensions: composite violation identification, large-scale parallel processing, and normal order judgment. This generates comprehensive test results reflecting the system's actual penalty capabilities, providing specific and actionable quantitative data support for the performance evaluation and subsequent optimization of the ride-hailing violation penalty system.

[0069] In summary, this disclosure presents a testing method for a ride-hailing violation penalty system: 1. Significantly reduce the cost of reproducing violation scenarios and improve the repeatability of test scenarios: Relying on a multi-dimensional testing system composed of a pre-built road network simulation database, driver behavior pattern database, and policy rule database, various violation and normal order test data are automatically generated by the server. There is no need to invest in vehicles and personnel to conduct offline real-vehicle road tests, completely eliminating the dependence on offline resources. At the same time, the test data is generated based on standardized database basic information and is not affected by external factors such as environment and road conditions. It can stably reproduce the same violation test scenarios, significantly reducing testing costs while improving the repeatability of test scenarios.

[0070] 2. Improve the testing efficiency of the ride-hailing penalty system and adapt to the testing needs of frequent iterations of penalty rules and city-specific control policies: This solution, based on a multi-dimensional testing system, can generate targeted test data and test cases for corresponding updated rules and adjusted policies. It can complete the targeted verification of updated penalty rules and adjusted control policies without conducting repetitive full-scale testing. At the same time, it can achieve parallel penalty by multiple nodes by combining a distributed testing cluster, which can significantly shorten the testing time and reduce the consumption of manpower and material resources. It can efficiently adapt to the testing needs of frequent iterations of ride-hailing penalty rules and city-specific control policies.

[0071] 3. Achieve synchronous simulation and collaborative injection of multi-source data to fully cover test scenarios: The solution can generate test data covering multiple types such as GPS driving trajectory, driver chat history, order operation log, and vehicle qualification parameters. Through a large language model, it can logically match and fuse multiple types of violation test data, fully reproduce complex and composite violation scenarios involving multi-source data in real business, solve the shortcomings of existing technologies that only test single types of data, realize synchronous simulation and collaborative injection of multi-source heterogeneous data, significantly improve the coverage of test scenarios, and fully verify the penalty system's ability to detect complex violation scenarios.

[0072] 4. Achieve full automation of the testing process and ensure the authenticity and reliability of test results: From test data generation and combination of complex violation test cases to test case input, penalty result collection and analysis, the entire process is completed automatically by the server. There is no need for manual intervention in test data sorting, penalty result verification and other operations. This can not only efficiently meet the verification needs of large-scale test cases and greatly improve the overall testing efficiency, but also avoid verification errors caused by human operation, ensuring the authenticity and reliability of test results. It can effectively avoid the problems of missed violations and system failures in the ride-hailing penalty system after testing, and effectively improve the accuracy and operational stability of the ride-hailing penalty system's violation judgment.

[0073] It should be understood that although the steps in the flowchart are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order constraint on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowchart may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0074] In some exemplary embodiments of this disclosure, such as Figure 4As shown, a testing device for a ride-hailing violation penalty system is provided. The device includes a first generation module 401, a second generation module 402, a penalty module 403, and an analysis module 404. The first generation module 401 generates test data corresponding to the violation type based on a pre-constructed multi-dimensional testing system, which includes one or more of a road network simulation database, a driver behavior pattern database, and a policy rule database. The second generation module 402 combines test data of two or more violation types to generate multiple composite violation test cases. The penalty module 403 inputs the multiple composite violation test cases into the ride-hailing violation penalty system and penalizes each composite violation test case through the system. The analysis module 404 collects the penalty results of each composite violation test case output by the ride-hailing violation penalty system, analyzes the penalty results, and generates test results for the composite violation detection of the ride-hailing violation penalty system.

[0075] In one embodiment of this disclosure, the road network simulation database is configured with road network node data and road topology information of each city, used to generate test data for vehicle driving trajectory violations; and / or, the driver behavior pattern database is configured with feature maps of various driver violations, used to generate test data for driver behavior violations; and / or, the policy rule database is configured with ride-hailing regulatory policy clauses of each city, used to generate test data for policy violations.

[0076] In one embodiment of this disclosure, the first generation module 401 is specifically used to: generate test data for vehicle trajectory violations based on road network node data and road topology information from a road network simulation database, wherein the test data for vehicle trajectory violations includes trajectory data of illegal detours and trajectory data that does not conform to the constraints of the real road network; and / or, generate test data for driver behavior violations based on a violation behavior feature map from a driver behavior pattern library, wherein the test data for driver behavior violations includes service quality substandard feature data, order-brushing behavior feature data, and distracted driving feature data; and / or, generate test data for policy-related violations based on urban ride-hailing regulatory policy clauses from a policy rule library, wherein the test data for policy-related violations includes order-taking data of non-local license plate vehicles, vehicle range substandard data, and non-compliant vehicle operation data.

[0077] In one embodiment of this disclosure, the second generation module 402 is specifically used to: select target test data of any two or more violation types from the test data of each violation type; input multiple target test data into a large language model to obtain multiple composite violation test cases output by the large language model.

[0078] In one embodiment of this disclosure, a testing device for a ride-hailing violation judgment system further includes: a third generation module, used to generate normal order test data that conforms to the normal operation specifications of ride-hailing based on a multi-dimensional testing system, and to construct multiple normal order test cases based on the normal order test data; and a judgment module 403, specifically used to: mix the multiple normal order test cases with multiple composite violation test cases, input them into the ride-hailing violation judgment system, and judge each normal order test case and each composite violation test case through the ride-hailing violation judgment system.

[0079] In one embodiment of this disclosure, multiple normal order test cases and multiple composite violation test cases are mixed and input into a ride-hailing violation judgment system. The ride-hailing violation judgment system then judges each normal order test case and each composite violation test case. This includes: after mixing multiple normal order test cases and multiple composite violation test cases, the mixed test cases are input into the ride-hailing violation judgment system through a distributed test cluster. The distributed test cluster is used to distribute the mixed test cases to multiple test nodes of the ride-hailing violation judgment system, and the multiple test nodes judge in parallel. Each test node of the ride-hailing violation judgment system outputs the judgment results of each normal order test case and each composite violation test case.

[0080] In one embodiment of this disclosure, the analysis module 404 is specifically used to: collect the judgment results of each normal order test case and each composite violation test case output by the ride-hailing violation judgment system, analyze the judgment results of each normal order test case and each composite violation test case, and obtain the detection rate of composite violation test cases, the average response latency of multi-node parallel judgment, and the false judgment rate of normal order test cases. The test results of composite violation detection of the ride-hailing violation judgment system include the detection rate of composite violation test cases, the average response latency of multi-node parallel judgment, and the false judgment rate of normal order test cases.

[0081] Specific limitations regarding the testing device for a ride-hailing violation judgment system can be found in the above-described limitations of the testing method for such a system, and will not be repeated here. Each module in the aforementioned testing device for a ride-hailing violation judgment system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in the electronic device, or stored in the memory of the electronic device as software, so that the processor can call and execute the corresponding operations of each module.

[0082] In some exemplary embodiments of this disclosure, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 5As 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 in the non-volatile storage media. The database stores relevant data such as road network simulation data, driver behavior pattern data, policy and rule data, and composite violation test case data. The network interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a testing method for a ride-hailing violation penalty system.

[0083] Those skilled in the art will understand that Figure 5 The 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 electronic device to which the present application is applied. The specific electronic device may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.

[0084] In some exemplary embodiments of this disclosure, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of a test method for a ride-hailing violation judgment system as described in any of the exemplary embodiments above.

[0085] 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 the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may 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 data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0086] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0087] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method for testing a system for judging penalties for violations of a ride-hailing service, the method comprising: The method includes: Test data corresponding to the violation type is generated based on a pre-built multi-dimensional testing system, which includes one or more of a road network simulation database, a driver behavior pattern database, and a policy rule database. Combine test data of two or more violation types to generate multiple composite violation test cases; The multiple composite violation test cases are input into the ride-hailing violation judgment system, and the ride-hailing violation judgment system judges each composite violation test case. The penalty results of each composite violation test case output by the ride-hailing violation penalty system are collected, the penalty results are analyzed, and the test results of the composite violation detection of the ride-hailing violation penalty system are generated.

2. The method of claim 1, wherein, The road network simulation database is configured with road network node data and road topology information for each city, used to generate test data for vehicle driving trajectory violations; and / or, the driver behavior pattern database is configured with feature maps of various driver violations, used to generate test data for driver behavior violations; and / or, the policy rule database is configured with ride-hailing regulatory policy clauses for each city, used to generate test data for policy violations.

3. The method of claim 2, wherein, The test data generated based on the pre-built multi-dimensional testing system for corresponding violation types includes: Based on the road network node data and road topology information of the road network simulation database, test data for vehicle driving trajectory violations are generated. The test data for vehicle driving trajectory violations includes trajectory data of illegal detours and trajectory data that does not conform to the constraints of the real road network. And / or, test data for driver behavior violations are generated based on the violation behavior feature map of the driver behavior pattern library. The test data for driver behavior violations includes service quality substandard feature data, order-brushing behavior feature data, and distracted driving feature data. And / or, test data on policy-related violations are generated based on the urban ride-hailing regulatory policy provisions in the policy rule base. The test data on policy-related violations includes order-taking data of vehicles with non-local license plates, data on vehicles with insufficient range, and operational data of non-compliant vehicle models.

4. The method of claim 1, wherein, The method of combining test data of two or more violation types to generate multiple composite violation test cases includes: Select any two or more target test data types from the test data of each violation type; Multiple target test data are input into a large language model to obtain multiple composite violation test cases output by the large language model.

5. The method of claim 1, wherein, The method further includes: Based on the multi-dimensional testing system, normal order test data that conforms to the normal operation standards of ride-hailing services is generated, and multiple normal order test cases are constructed based on the normal order test data; The step of inputting the multiple composite violation test cases into the ride-hailing violation judgment system, and then judging each composite violation test case through the ride-hailing violation judgment system, includes: The multiple normal order test cases are mixed with the multiple composite violation test cases and input into the ride-hailing violation judgment system. The ride-hailing violation judgment system then judges and penalizes each normal order test case and each composite violation test case.

6. The method according to claim 5, characterized in that, The process of mixing the multiple normal order test cases with the multiple composite violation test cases and inputting them into the ride-hailing violation judgment system, and then using the ride-hailing violation judgment system to judge and penalize each normal order test case and each composite violation test case, includes: After mixing the multiple normal order test cases with the multiple composite violation test cases, the mixed test cases are input into the ride-hailing violation judgment system through a distributed test cluster. The distributed test cluster is used to distribute the mixed test cases to multiple test nodes of the ride-hailing violation judgment system, and the multiple test nodes make judgments in parallel. The test nodes of the ride-hailing violation judgment system output the judgment results of each normal order test case and each compound violation test case.

7. The method according to claim 6, characterized in that, The process involves collecting the penalty results of each composite violation test case output by the ride-hailing violation penalty system, analyzing the penalty results, and generating test results for the composite violation detection of the ride-hailing violation penalty system, including: The system collects the penalty results of each normal order test case and each compound violation test case output by the ride-hailing violation penalty system. It then analyzes these results to obtain the detection rate of the compound violation test cases, the average response latency of the multi-node parallel penalty process, and the false positive rate of the normal order test cases. The test results of the ride-hailing violation penalty system for compound violation detection include the detection rate of the compound violation test cases, the average response latency of the multi-node parallel penalty process, and the false positive rate of the normal order test cases.

8. A testing device for a ride-hailing violation judgment system, characterized in that, The device includes: The first generation module is used to generate test data for the corresponding violation type based on a pre-built multi-dimensional testing system, wherein the multi-dimensional testing system includes one or more of a road network simulation database, a driver behavior pattern database, and a policy rule database. The second generation module is used to combine test data of two or more violation types to generate multiple composite violation test cases. The penalty module is used to input the multiple composite violation test cases into the ride-hailing violation penalty system, and to penalize each composite violation test case through the ride-hailing violation penalty system. The analysis module is used to collect the penalty results of each composite violation test case output by the ride-hailing violation penalty system, analyze the penalty results, and generate the test results of the composite violation detection of the ride-hailing violation penalty system.

9. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, 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 a computer program stored thereon, 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