Shared electric bicycle platform vehicle capacity evaluation method based on multi-component perception

Abstract

The application discloses a shared electric bicycle platform vehicle capacity evaluation method based on multi-component perception and relates to the technical field of data processing and analysis. The method comprises the following steps: S1, a multi-component test environment for shared electric bicycle business is constructed to simulate the whole-link process of vehicle data reporting, processing and storage; S2, vehicle load modeling is performed based on shared electric bicycle business characteristics to generate dynamic test load conforming to the space-time law of vehicle use; S3, multi-component perception index fusion is implemented to establish a health degree tensor evaluation model of the vehicle data processing link; S4, based on the health degree tensor, multi-component collaborative analysis vehicle capacity evaluation calculation is performed; and S5, bottleneck positioning analysis of the vehicle data processing link is performed to identify the key components limiting the system capacity. The application improves the evaluation precision of the platform vehicle capacity by constructing an intelligent load model, a multi-dimensional index and a capacity evaluation system.

Description

Technical Field

[0001] This invention relates to the field of data processing and analysis technology, specifically to a method for assessing vehicle capacity on a shared electric bicycle platform based on multi-component perception. Background Technology

[0002] With the popularization of shared mobility, shared e-bikes have become an important part of urban short-distance transportation. The operating platform needs to process massive amounts of data reported by vehicles in real time, including GPS location, battery status, and riding status, and then store, calculate, and provide feedback. This places extremely high demands on the processing power and stability of the backend system. Vehicle usage exhibits a significant tidal effect, with huge concurrent data flows during morning and evening rush hours causing drastic periodic fluctuations in system load. Against this backdrop, accurately assessing system capacity becomes crucial for ensuring service quality and optimizing resource allocation.

[0003] Traditional system capacity assessment methods often rely on static threshold judgments or stress tests based on fixed patterns. Static testing models struggle to simulate the complex load fluctuations that change over time and location in real-world business operations, leading to a disconnect between test results and actual operational conditions. Secondly, traditional monitoring methods frequently observe components as isolated islands, neglecting the close performance interrelationships between components such as gateways, message queues, caches, and databases in a microservice architecture, thus failing to assess capacity bottlenecks from a holistic system perspective. Therefore, there is an urgent need for a capacity assessment method that can accurately reflect business characteristics and perceive the overall system status.

[0004] For example, Chinese Patent Publication No. CN113010211A discloses a method for porting an application system, as well as a method, device, and storage medium for estimating capacity specifications. The porting method includes: obtaining initial capacity assessment elements and original system capacity parameters based on the evaluation objectives of the original host platform; obtaining multi-level capacity parameters; calculating the target capacity specification of the new system based on the original system capacity parameters and the multi-level capacity parameters; and determining the capacity specifications corresponding to each physical subsystem of the target system according to the target capacity specification of the new system, thereby porting the application system to each physical subsystem. This method calculates the new system capacity specification and the corresponding capacity specifications of each physical subsystem of the target system based on the evaluation objectives of the original host platform and the obtained multi-level capacity parameters. Therefore, it can port the application system from the original host platform to each physical subsystem of the target system according to production needs, not only maintaining the service capabilities of the application system but also effectively saving hardware resource investment.

[0005] For example, Chinese patent CN103631662A discloses a system and method for capacity assessment of an event analysis engine in a complex event processing cloud platform. The system includes an event analysis engine for receiving external events and loading various event processing rules; an event source management module for managing event metadata in the complex event processing cloud platform and providing event structures corresponding to various event types; and an analysis engine capacity assessment module for assessing the capacity of the event analysis engine based on the number and types of events arriving in the engine and the event structures corresponding to each event type in the event source management module. This system and method for capacity assessment of the event analysis engine in a complex event processing cloud platform, based on a single node (i.e., a single JVM process) for event throughput, significantly improves the overall operational level of the complex event processing platform. The method is stable, reliable, and widely applicable. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to address the shortcomings of the prior art by providing a method for assessing the vehicle capacity of a shared electric bicycle platform based on multi-component perception.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0008] A method for assessing vehicle capacity on a shared electric bicycle platform based on multi-component perception includes the following steps:

[0009] Step S1: Construct a multi-component test environment for shared electric bicycle business to simulate the entire process of vehicle data reporting, processing and storage;

[0010] Step S2: Based on the characteristics of shared electric bicycle business, perform vehicle load modeling to generate dynamic test load that conforms to the spatiotemporal patterns of vehicle use;

[0011] Step S3: Implement multi-component perception index fusion and establish a health tensor assessment model for the vehicle data processing link;

[0012] Step S4: Based on the health tensor, perform vehicle capacity assessment calculation through multi-component collaborative analysis;

[0013] Step S5: Perform bottleneck location analysis on the vehicle data processing link to identify key components that limit system capacity.

[0014] Furthermore, step S1 specifically includes the following steps:

[0015] Step S1.1: Configure the containerized cluster environment at the infrastructure layer;

[0016] Step S1.2: At the component layer, deploy a core component cluster directly related to vehicle data processing. The core component cluster includes: an access gateway cluster for receiving massive amounts of vehicle-reported data, a message queue cluster for buffering real-time vehicle messages, a cache and database cluster for storing vehicle status information, and a search cluster for vehicle trajectory retrieval.

[0017] Step S1.3: At the monitoring layer, deploy a dedicated monitoring system for the vehicle data processing link to collect performance indicators of each component in processing vehicle data at a fixed granularity.

[0018] Furthermore, in step S2, the vehicle load modeling is performed using a spatiotemporal load distribution function, the specific formula of which is:

[0019]

[0020] in, The spatiotemporal load distribution function represents the modeling of vehicle load. Indicates time, Indicates the daily fluctuation amplitude. Represents angular frequency. Indicates the base load factor. Indicates phase shift, This represents the random disturbance coefficient. This represents a Gaussian white noise function with a mean of 0 and a variance of 1. This indicates the basic vehicle scale, that is, the total number of vehicles operated by the platform. This represents the sine function.

[0021] Furthermore, step S3 specifically includes the following steps:

[0022] Step S3.1: Define the three evaluation dimensions of the health tensor, including the time dimension, component dimension, and indicator dimension;

[0023] Step S3.2: Assign dynamic weights to each component's indicators based on the importance of each component in the vehicle data processing link;

[0024] Step S3.3: Integrate the real-time indicator values ​​and indicator change trends of each component to calculate the health tensor; wherein, the value of the health tensor is the sum of the tensor products of the weighted real-time indicator values ​​and the indicator change trend values.

[0025] Furthermore, in step S3.1, the components included in the component dimension are: access gateway, message queue, cache, search engine, and database.

[0026] Furthermore, step S4 specifically includes the following steps:

[0027] Step S4.1: Calculate the real-time capacity adequacy index, which is calculated by the ratio of actual throughput to theoretical maximum throughput, the average delay effect smoothed by a logarithmic function, and the Euclidean norm of the comprehensive resource utilization rate.

[0028] Step S4.2: Calculate the trend prediction capacity coefficient. The trend prediction capacity coefficient is based on the real-time capacity adequacy index, and the first and second derivatives of capacity change are introduced for prediction.

[0029] Step S4.3: Calculate the component bottleneck impact factor, which is obtained by taking the partial derivative of the real-time capacity adequacy index with respect to the component performance index.

[0030] Step S4.4: Calculate the system resilience score, which is based on the capacity integral, the ratio of peak to trough values, and the capacity volatility over a specified time period.

[0031] Furthermore, step S5 specifically includes the following steps:

[0032] Step S5.1: Based on the difference between the current capacity and the target capacity, and the rate of change of the error rate with the load, adaptively adjust the load step size applied to the system.

[0033] Step S5.2: Construct a directed graph of performance impact, wherein the vertices of the directed graph of performance impact represent each component, and each edge corresponds to an impact strength;

[0034] Step S5.3: Analyze the directed graph of performance impact to locate the bottleneck component that has the greatest impact on the overall system performance.

[0035] Furthermore, in step S5.2, the method for calculating the influence intensity in the directed graph of performance influence is as follows: calculate the correlation coefficient of the performance index between related components and multiply it by the ratio of the coefficient of variation of the index on the corresponding component.

[0036] A storage medium storing instructions that, when read by a computer, cause the computer to execute a vehicle capacity assessment method for a shared electric bicycle platform based on multi-component perception.

[0037] An electronic device includes a processor and the storage medium, the processor executing instructions in the storage medium.

[0038] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0039] 1. This invention comprehensively and realistically simulates business scenarios by introducing a dynamic load model that integrates spatiotemporal characteristics and a health tensor that integrates multiple dimensions of indicators, thereby improving the reliability of the evaluation results.

[0040] 2. By calculating the bottleneck impact factor of components and constructing a directed graph of performance impact, this invention can accurately quantify the marginal impact of each component on the overall capacity and locate the root cause of the bottleneck, making the expansion and optimization schemes more targeted.

[0041] 3. This invention uses a system resilience scoring formula to comprehensively evaluate the stability and recovery capability of a system under fluctuating loads, enabling operators to clearly understand the system's robustness and providing data support for capacity planning in terms of stability.

[0042] 4. This invention uses a trend prediction capacity coefficient formula to introduce the first and second derivatives of capacity changes for prediction, achieving a leap from static assessment to dynamic prediction. It can identify capacity risks in advance and reserve time for operation and maintenance intervention. Attached Figure Description

[0043] Other features, objects, and advantages of the invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0044] Figure 1 This is a flowchart illustrating an embodiment of the present invention;

[0045] Figure 2 This is a directed example diagram illustrating the performance impact of an embodiment of the present invention;

[0046] Figure 3 This is a trend chart of capacity assessment indicators for the shared electric bicycle platform according to an embodiment of the present invention. Detailed Implementation

[0047] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0048] like Figure 1 As shown, the vehicle capacity assessment method for shared electric bicycle platforms based on multi-component perception includes the following steps:

[0049] Step S1: Construct a multi-component test environment for shared electric bicycle business to simulate the entire process of vehicle data reporting, processing and storage;

[0050] Step S2: Based on the characteristics of shared electric bicycle business, perform vehicle load modeling to generate dynamic test load that conforms to the spatiotemporal patterns of vehicle use;

[0051] Step S3: Implement multi-component perception index fusion and establish a health tensor assessment model for the vehicle data processing link;

[0052] Step S4: Based on the health tensor, perform vehicle capacity assessment calculation through multi-component collaborative analysis;

[0053] Step S5: Perform bottleneck location analysis on the vehicle data processing link to identify key components that limit system capacity.

[0054] Step S1 specifically includes the following steps:

[0055] Step S1.1: Configure the containerized cluster environment at the infrastructure layer;

[0056] Step S1.2: At the component layer, deploy a core component cluster directly related to vehicle data processing. The core component cluster includes: an access gateway cluster for receiving massive amounts of vehicle-reported data, a message queue cluster for buffering real-time vehicle messages, a cache and database cluster for storing vehicle status information, and a search cluster for vehicle trajectory retrieval.

[0057] Step S1.3: At the monitoring layer, deploy a dedicated monitoring system for the vehicle data processing link to collect performance indicators of each component in processing vehicle data at a fixed granularity.

[0058] Infrastructure layer: Employs a Docker container cluster with over 50 nodes to ensure scalability and isolation of the environment. Each node is configured with standard compute and memory resources to simulate a production environment.

[0059] The component layer includes the following components:

[0060] Access gateway: Five instances of Nginx and Spring Cloud Gateway are deployed to handle high concurrency requests and achieve load balancing.

[0061] Message queue: A Kafka cluster with 3 brokers is used, and the replication factor is set to 2 to ensure high availability and reliability of messages.

[0062] Caching layer: Uses Redis Cluster, which contains 8 nodes and a total memory of 32GB, supporting distributed caching and data persistence.

[0063] Search layer: Uses an Elasticsearch cluster with 10 nodes and 20 shards to optimize indexing and query performance.

[0064] Database: A MySQL master-slave cluster is used, consisting of 1 master node and 3 slave nodes, to achieve read-write separation and data backup.

[0065] Monitoring layer: Prometheus is integrated for metric collection, Grafana for visualization and monitoring, and a customized collection agent is developed to collect component-level metrics. Monitoring data is stored in 1-minute increments for subsequent analysis.

[0066] In step S2, the vehicle load modeling is performed using a spatiotemporal load distribution function, the specific formula of which is:

[0067]

[0068] in, The spatiotemporal load distribution function represents the modeling of vehicle load. Indicates time, Indicates the daily fluctuation amplitude. Represents angular frequency. Indicates the base load factor. Indicates phase shift, This represents the random disturbance coefficient. This represents a Gaussian white noise function with a mean of 0 and a variance of 1. This indicates the basic vehicle scale, that is, the total number of vehicles operated by the platform. This represents the sine function.

[0069] Daily fluctuation amplitude The default value is 0.3, representing the daily fluctuation range of the load; it adjusts according to the time zone of the business to match the peak load time with the actual business peak hours; base load factor. The default value is 1.0; random disturbance coefficient. The value ranges from 0.1 to 0.3, simulating random fluctuations caused by sudden events (such as weather changes).

[0070] Step S3 specifically includes the following steps:

[0071] Step S3.1: Define the three evaluation dimensions of the health tensor, including the time dimension, component dimension, and indicator dimension;

[0072] Step S3.2: Assign dynamic weights to each component's indicators based on the importance of each component in the vehicle data processing link;

[0073] Step S3.3: Integrate the real-time indicator values ​​and indicator change trends of each component to calculate the health tensor; wherein, the value of the health tensor is the sum of the tensor products of the weighted real-time indicator values ​​and the indicator change trend values.

[0074] The formula for calculating the health tensor is:

[0075]

[0076] in, This represents the health tensor of the three assessment dimensions. The time dimension is indicated, with a granularity of 1 minute, representing the time point at which the indicator was collected. Indicates component dimension, The metrics include throughput, latency, error rate, and resource utilization. This represents dynamic weights, allocated based on component importance. Indicates the component index. Indicates the total number of components. This represents the real-time metric value of the m-th component. This represents the trend value of the indicator, i.e., the first derivative of the real-time indicator value. This represents the tensor product operator.

[0077] Dynamic weights The specific allocation includes:

[0078] The database component has the highest weight, set to 0.30, because it is the final storage point for data, and any delay or error will directly affect the persistence and querying of vehicle data.

[0079] The access gateway component has a weight of 0.25. As the entry point for vehicle data reporting, its throughput and latency directly affect the user experience.

[0080] The message queue component has a weight of 0.20 and is responsible for buffering peak traffic. If it accumulates, it will cause data processing delays.

[0081] The cache component has a weight of 0.15 and is used to store the real-time status of the vehicle. Although a failure may cause a performance degradation, it will not completely interrupt the service.

[0082] The search component has a weight of 0.10 and is mainly used for vehicle trajectory queries. Its performance has a relatively small impact on the core business flow.

[0083] In step S3.1, the components included in the component dimension are: access gateway, message queue, cache, search engine, and database.

[0084] Step S4 specifically includes the following steps:

[0085] Step S4.1: Calculate the real-time capacity adequacy index. The real-time capacity adequacy index is calculated by the ratio of actual throughput to theoretical maximum throughput, the average latency effect smoothed by a logarithmic function, and the Euclidean norm of the overall resource utilization rate. The specific formula is as follows:

[0086]

[0087] in, This represents the real-time capacity adequacy index. This represents the actual throughput, indicating the number of requests the system is currently processing per second. This represents the theoretical maximum throughput, calculated through stress testing or the upper limit of component performance. Indicates average latency, representing the average response time for a request. Indicates CPU utilization. Indicates memory usage. Indicates I / O utilization. This represents a logarithmic function with a constant e as the base; the exponent ranges from 0 to 1, and the closer it is to 1, the more abundant the capacity.

[0088] Step S4.2: Calculate the trend prediction capacity coefficient. This trend prediction capacity coefficient, based on the real-time capacity adequacy index, incorporates the first and second derivatives of capacity change for prediction. The specific formula is as follows:

[0089]

[0090] in, Indicates the trend forecasting capacity coefficient. This represents the empirical coefficient for trend analysis; the default value is 0.3. This represents the empirical coefficient of acceleration, with a default value of 0.1. The first derivative, representing the change in capacity, is calculated using historical data, such as the rate of change over the past 5 minutes, and indicates the trend of capacity change. The second derivative represents the capacity, and the acceleration represents the change. This indicates the capacity adequacy index within the time window;

[0091] Step S4.3: Calculate the component bottleneck impact factor. The component bottleneck impact factor is obtained by calculating the partial derivative of the real-time capacity adequacy index with respect to the component performance index. The specific formula is as follows:

[0092]

[0093] in, Indicates the bottleneck factor of the component. This represents the performance metrics of component m;

[0094] Step S4.4: Calculate the system resilience score. The system resilience score is calculated based on the capacity integral, the ratio of peak to trough values, and the capacity volatility over a specified time period. The specific formula is as follows:

[0095]

[0096] in, This indicates the system's resilience score. This represents the integral of the content sufficiency index over the time interval [t1, t2]. This represents the peak capacity, i.e., the maximum value within the interval. This represents the valley capacity, i.e., the minimum value within the interval. This represents capacity volatility.

[0097] Among them, the capacity volatility rate is the standard deviation of the real-time capacity adequacy index within a specified assessment period (e.g., the past 24 hours).

[0098] Step S5 specifically includes the following steps:

[0099] Step S5.1: Based on the difference between the current capacity and the target capacity, and the rate of change of the error rate with load, adaptively adjust the load step size applied to the system. The specific formula is as follows:

[0100]

[0101] in, Indicates the load adjustment amount. This represents the scaling factor, with a default value of 0.5, which controls the adjustment step size. Indicates the target capacity adequacy index. This represents the weighting coefficient, with a default value of 0.2, used to avoid excessively high error rates. This indicates the current error rate of the system. Indicates load, This represents the rate of change of error rate with load.

[0102] Step S5.2: Construct a directed graph of performance impact G=(V, E), where the vertices V of the directed graph of performance impact represent each component, and each edge corresponds to an impact strength.

[0103] Step S5.3: Analyze the directed graph of performance impact to locate the bottleneck component that has the greatest impact on the overall system performance.

[0104] In step S5.2, the method for calculating the influence intensity in the directed graph of performance influence is as follows: calculate the correlation coefficient of the performance indicators between related components and multiply it by the ratio of the coefficient of variation of the indicator on the corresponding component. The specific formula is:

[0105]

[0106] in, This indicates the strength of the influence from component x to component y. The Pearson correlation coefficient represents the relationship between the performance metrics of component x and component y. and These represent the sum of the coefficients of variation for the corresponding indicators of components x and y, respectively. The coefficient of variation is the ratio of the standard deviation to the absolute value, used to measure the relative dispersion of the data.

[0107] The higher the value, the greater the likelihood that component x is the key bottleneck causing performance issues in component y. This is achieved by comparing all edges. By analyzing the value, we can pinpoint the bottleneck that has the greatest impact on the overall system performance.

[0108] like Figure 2 As shown, the directed graph contains 5 vertices and 7 directed edges. The vertices are arranged in a ring, and the weight values ​​on the edges reflect the strength of the influence.

[0109] Vertex V represents the various system components, including:

[0110] V1: Access Gateway, V2: Message Queue, V3: Cache, V4: Search Engine, V5: Database;

[0111] Specific values ​​include:

[0112] V1→V2: 0.85 (A decrease in gateway throughput will significantly increase message queue backlog);

[0113] V2→V3: 0.72 (Message backlog can lead to increased cache access latency);

[0114] V3→V4: 0.58 (Cache performance affects search response speed);

[0115] V3→V5: 0.65 (Cache invalidation will increase database load);

[0116] V4→V5: 0.41 (Complex search queries will increase database pressure);

[0117] This diagram identifies two key impact paths:

[0118] Main bottleneck path: V1→V2→V3→V5 (maximum cumulative impact);

[0119] Secondary impact path: V1→V2→V3→V4→V5.

[0120] There is a high-impact path from V1 to V5, indicating that gateway performance fluctuations will ultimately significantly affect database performance through message queues and caching layers.

[0121] like Figure 3 As shown, the left-hand Y-axis displays the capacity assessment index, ranging from 0 to 1.0, and includes two key curves:

[0122] Real-time capacity index curve: calculated based on the real-time capacity adequacy formula, reflecting the current actual capacity status of the system;

[0123] Predicted capacity index curve: Generated using a trend-based capacity coefficient formula, providing capacity forecasts for the next 2-4 hours.

[0124] The Y-axis on the right displays the business load metric, with a value range of 0-100%, including:

[0125] Vehicle load curve: The simulated load generated based on the spatiotemporal load distribution function L(t) reflects the tidal effect of vehicle use;

[0126] System error rate curve: The actual error rate data collected by the monitoring layer reflects the system stability.

[0127] Key time periods are marked: The chart uses red shaded areas to clearly mark the morning and evening peak hours (07:00-09:00 and 17:00-19:00). These periods correspond to the peak periods of vehicle use and are the key focus areas for capacity assessment.

[0128] The following aspects can be observed through this trend chart:

[0129] Negative correlation between capacity and load: When vehicle load reaches its peak during morning and evening rush hours, the real-time capacity index drops to its trough accordingly, verifying the effectiveness of the load model.

[0130] Predictive and early warning function: The predicted capacity curve has a clear leading indication effect compared to the real-time curve. For example, before the morning peak at 08:00, the predicted curve has already indicated the downward trend of capacity.

[0131] Correlation between error rate and capacity: The system error rate increases significantly when the capacity index is below 0.7, proving that there is a direct causal relationship between capacity adequacy and system stability.

[0132] Any combination of one or more computer-readable media may be used. A computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0133] The examples described herein are merely preferred embodiments of the invention and are not intended to limit the concept and scope of the invention. Any modifications and improvements made by those skilled in the art to the technical solutions of the invention without departing from the design concept of the invention should fall within the protection scope of the invention.

Claims

1. A method for evaluating vehicle capacity of a shared electric bicycle platform based on multi-component perception, characterized in that, Comprise the following steps: Step S1, construct a multi-component test environment for shared electric bicycle business, simulate the whole link process of vehicle data reporting, processing and storage; Step S2, based on the characteristics of shared electric bicycle business, vehicle load modeling is performed to generate dynamic test load conforming to the space-time law of vehicle use; Step S3, implement multi-component aware index fusion, establish a health degree tensor evaluation model of the vehicle data processing link, wherein the components include: access gateway, message queue, cache, search engine and database; Step S4, based on the obtained health degree tensor, perform multi-component collaborative analysis of vehicle capacity evaluation calculation, specifically including the following steps: Step S4.1, calculating the real-time capacity adequacy index in the following manner: wherein, represents the real-time capacity adequacy index, represents the actual throughput, represents the theoretical maximum throughput, represents the average delay, represents the CPU usage rate, represents the memory usage rate, represents the I / O usage rate; Step S4.2, calculate the trend prediction capacity coefficient, which introduces the first and second derivatives of capacity change based on the real-time capacity adequacy index for prediction; Step S4.3, calculate the component bottleneck influence factor, which is obtained by taking the partial derivative of the real-time capacity adequacy index with respect to the component performance index; Step S4.4, calculate the system elasticity score, which is calculated based on the capacity integral, the ratio of peak value to valley value and the capacity fluctuation rate in a specified time period; Step S5, perform bottleneck positioning analysis for the vehicle data processing link, identify the key components that limit the system capacity, specifically including the following steps: Step S5.1, according to the difference between the current capacity and the target capacity and the change rate of the system current error rate with the load, adaptively adjust the load step applied to the system; Step S5.2, constructing a performance influence directed graph, vertices of the performance influence directed graph represent components, each edge corresponds to an influence strength, the influence strength is calculated as follows: wherein, represents the influence strength from component x to component y, the higher the value, the greater the possibility that component x is a key bottleneck leading to performance problems of component y; represents the Pearson correlation coefficient between component x and component y performance indicators, and respectively represent the sum of the coefficients of variation of the corresponding indicators of components x and y, wherein the coefficient of variation is the ratio of the standard deviation to the absolute value, used to measure the relative dispersion degree of data; Step S5.3, analyze the performance impact directed graph to locate the bottleneck component that has the greatest impact on the overall system performance.

2. The method of claim 1, wherein, The step S1 specifically includes the following steps: Step S1.1, configure a containerized cluster environment at the infrastructure layer; Step S1.2, deploy a core component cluster directly related to vehicle data processing at the component layer, wherein the core component cluster includes: an access gateway cluster for receiving massive vehicle reported data, a message queue cluster for buffering vehicle real-time messages, a cache and database cluster for storing vehicle state information, and a search cluster for vehicle trajectory retrieval; Step S1.3, deploy a special monitoring system for the vehicle data processing link at the monitoring layer to collect performance indicators of each component processing vehicle data at a fixed granularity.

3. The method of claim 2, wherein, In step S2, the vehicle load modeling uses a space-time load distribution function for modeling, and the specific formula is: ； wherein, denotes a spatiotemporal load distribution function modeling the vehicle load, denotes time, denotes the daily fluctuation amplitude, denotes the angular frequency, denotes the base load coefficient, denotes the phase shift, denotes the random perturbation coefficient, denotes a Gaussian white noise function with mean 0 and variance 1, denotes the base vehicle scale, i.e. the total number of vehicles operated by the platform, denotes the sinusoidal function.

4. The method of claim 3, wherein, The step S3 specifically includes the following steps: Step S3.1, define three evaluation dimensions of the health degree tensor, including time dimension, component dimension and index dimension; Step S3.2, according to the importance of each component in the vehicle data processing link, assign a dynamic weight to each component index; Step S3.3, fuse the real-time index value and index change trend value of each component to calculate the health degree tensor; wherein the value of the health degree tensor is the tensor product sum of the weighted real-time index value and the index change trend value.

5. A storage medium, characterized by The storage medium has stored instructions, when a computer reads the instructions, the computer executes the multi-component perception-based shared electric bicycle platform vehicle capacity evaluation method in any one of claims 1-4.

6. An electronic device, comprising: Include a processor and the storage medium of claim 5, the processor executes the instructions in the storage medium.

Images