Multi-engine oriented mobile application detection task intelligent scheduling method and device

Abstract

The application discloses a kind of multi-engine-oriented mobile application detection task intelligent scheduling method and device, wherein the method comprises: obtaining the current multidimensional attribute data of mobile application to be detected task;According to the current multidimensional attribute data, and the relationship between the pre-configured task priority dimension and priority weight value, determine the current multidimensional task priority weight value;According to the current multidimensional task priority weight value, and the relationship between the pre-configured priority weight value interval and engine state weight value, determine the current engine state weight value;According to the current engine state data of all engines, and the current engine state weight value, determine the allocation value of each engine;The engine with the maximum allocation value is used as the optimal engine of the task to be detected, and the task to be detected is dispatched to the queue of the optimal engine.The application can intelligently and flexibly schedule multiple engines, improve the efficiency, flexibility and reliability of mobile application detection.

Description

Technical Field

[0001] This invention relates to the field of mobile application detection technology, and in particular to an intelligent scheduling method and apparatus for mobile application detection tasks oriented towards multiple engines. Background Technology

[0002] This section is intended to provide background or context for the embodiments of the invention set forth in the claims. The description herein is not an admission that it is prior art simply because it is included in this section.

[0003] With the widespread adoption of mobile internet and the explosive growth of mobile applications, the security and compliance issues of mobile applications have become increasingly prominent, attracting the attention of regulatory agencies, app stores, and users alike. To address the massive detection demands of numerous applications, batch and intelligent analysis technologies based on automated detection engines have become the industry mainstream. Currently, single detection engines often have limitations in detection scope, analysis depth, or specialized capabilities, making it difficult to cope with increasingly complex and covert security threats. Therefore, a multi-engine collaborative detection model that integrates the advantages of multiple technology providers and combines multiple heterogeneous detection engines (such as static analysis engines, dynamic sandbox engines, privacy compliance engines, and malicious behavior analysis engines) has become an inevitable trend for improving detection coverage and accuracy.

[0004] In the practice of multi-engine collaborative detection, existing technical solutions face a series of severe challenges and inherent defects in task scheduling and management, which seriously restrict the overall detection efficiency and the optimal utilization of resources. Existing multi-engine systems mostly use pre-set fixed rules for task allocation, such as simple polling, random allocation, or assignment based on fixed engine labels. This static scheduling mode cannot perceive real-time changes in the system state. It ignores the real-time load of the detection engines, the differences in the specific detection capabilities of different detection engines for specific types of applications (such as games and financial apps) or specific detection items, and the historical success rate fluctuations of the detection engines. When a detection engine is overloaded or temporarily experiences performance degradation due to processing complex samples, static scheduling cannot dynamically divert tasks to other idle or better-performing engines, leading to task queuing congestion, idle detection engines, reduced overall throughput, and potentially affecting the timely completion of high-priority tasks.

[0005] Therefore, there is an urgent need for a solution that can intelligently and flexibly schedule multiple engines to optimize resource allocation and improve the overall efficiency, flexibility and reliability of large-scale automated testing. Summary of the Invention

[0006] This invention provides an intelligent scheduling method for mobile application detection tasks across multiple engines, enabling intelligent and flexible scheduling of multiple engines. The method includes: Obtain the current multi-dimensional attribute data of the mobile application to be tested and the current engine status data of all engines; Based on the current multi-dimensional attribute data and the relationship between the pre-configured task priority dimensions and priority weight values, determine the current multi-dimensional task priority weight value corresponding to the current multi-dimensional attribute data; Based on the current multi-dimensional task priority weight value and the relationship between the pre-configured priority weight value range and the engine status weight value, determine the engine status weight value corresponding to the current multi-dimensional task priority weight value; The allocation value for each engine is determined based on the current engine status data of all engines and the engine status weight value corresponding to the current multi-dimensional task priority weight value. The engine with the largest allocation value is selected as the optimal engine for the task to be detected, and the task to be detected is dispatched to the queue of the optimal engine.

[0007] This invention also provides an intelligent scheduling device for mobile application detection tasks across multiple engines, enabling intelligent and flexible scheduling of multiple engines. The device includes: The acquisition unit is used to acquire the current multi-dimensional attribute data of the mobile application to be detected and the current engine status data of all engines; The multi-dimensional task priority evaluation unit is used to determine the current multi-dimensional task priority weight value corresponding to the current multi-dimensional attribute data based on the current multi-dimensional attribute data and the pre-configured relationship between task priority dimensions and priority weight values. The engine status weight value determination unit is used to determine the engine status weight value corresponding to the current multi-dimensional task priority weight value based on the current multi-dimensional task priority weight value and the relationship between the pre-configured priority weight value range and the engine status weight value. The allocation value determination unit is used to determine the allocation value of each engine based on the current engine status data of all engines and the engine status weight value corresponding to the current multi-dimensional task priority weight value. The optimal engine determination unit is used to select the engine with the largest allocation value as the optimal engine for the task to be detected, and to dispatch the task to be detected to the queue of the optimal engine. This invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the above-described intelligent scheduling method for mobile application detection tasks oriented towards multiple engines.

[0008] This invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described intelligent scheduling method for mobile application detection tasks oriented towards multiple engines.

[0009] This invention also provides a computer program product, which includes a computer program that, when executed by a processor, implements the above-described intelligent scheduling method for mobile application detection tasks oriented towards multiple engines.

[0010] In this embodiment of the invention, the intelligent scheduling scheme for mobile application detection tasks oriented towards multiple engines achieves intelligent and flexible scheduling of multiple engines by: acquiring the current multi-dimensional attribute data of the mobile application to be detected and the current engine status data of all engines; determining the current multi-dimensional task priority weight value corresponding to the current multi-dimensional attribute data based on the current multi-dimensional attribute data and the pre-configured relationship between task priority dimensions and priority weight values; determining the engine status weight value corresponding to the current multi-dimensional task priority weight value based on the current multi-dimensional task priority weight value and the pre-configured relationship between priority weight value range and engine status weight value; determining the allocation value of each engine based on the current engine status data of all engines and the engine status weight value corresponding to the current multi-dimensional task priority weight value; and selecting the engine with the largest allocation value as the optimal engine for the task to be detected, and dispatching the task to be detected to the queue of the optimal engine. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings: Figure 1 This is a flowchart illustrating the intelligent scheduling method for mobile application detection tasks oriented towards multiple engines in an embodiment of the present invention. Figure 2 This is a flowchart illustrating an intelligent scheduling method for mobile application detection tasks oriented towards multiple engines, as shown in another embodiment of the present invention. Figure 3 This is a flowchart illustrating the intelligent scheduling method for mobile application detection tasks oriented towards multiple engines in another embodiment of the present invention. Figure 4 This is a schematic diagram of the intelligent scheduling device for mobile application detection tasks oriented towards multiple engines in an embodiment of the present invention. Figure 5This is a schematic diagram of a computer device structure according to an embodiment of the present invention. Detailed Implementation

[0012] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings. Here, the illustrative embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0013] The acquisition, storage, use, and processing of data in this application all comply with the relevant provisions of national laws and regulations.

[0014] Existing technology 1: The existing technical solution is a detection task distribution system based on static configuration and simple polling. The system has a central scheduler and multiple detection engines. The scheduler maintains a FIFO (First-In-First-Out) global queue of all applications to be inspected, and distributes the tasks in the queue to each engine in sequence according to a preset, fixed engine weight (such as manually set according to the engine hardware performance) or a simple polling order.

[0015] The shortcomings of the existing technology 1: The detection task distribution system based on static configuration and simple polling only considers two single factors, the order of task arrival and static weight, in scheduling decisions. It cannot perceive the real-time load of the engine, the specific requirements of the task (such as the required detection items), the historical success rate, and the dynamic priority of the business. Therefore, it often leads to problems such as uneven task distribution, delays of high-priority tasks, and underutilization of professional engine capabilities. The overall scheduling efficiency is low and lacks flexibility.

[0016] Existing technology 2: A scheduling system based on fixed rules and task type matching. First, each connected detection engine is pre-labeled with the fixed task types it can handle (e.g., engine A only handles static code analysis, engine B only handles dynamic behavior analysis), and a static task type-engine type mapping table is maintained in the scheduling center. When a task arrives, the system only queries this mapping table based on the detection type requested by the task, assigning the task to the corresponding engine that is currently marked as idle. If all engines of that type are busy, the task enters a dedicated waiting queue for that type.

[0017] The disadvantages of the existing technology 2 are: the scheduling strategy of the scheduling system based on fixed rules and hard matching of task types is completely rigid and cannot be flexibly adjusted according to the real-time load of the engine, historical performance (such as a certain engine with a low success rate even if it matches the type), or the urgency of the sample; at the same time, it lacks cross-engine elasticity. Once all the dedicated engines of a certain type fail or become the performance bottleneck, even if other types of engines are idle, the system cannot perform intelligent task switching or cross-type scheduling, which leads to a sharp drop in the overall resource utilization and task throughput of the system and poor robustness.

[0018] To address the aforementioned technical problems in existing technologies, this invention proposes an intelligent scheduling scheme for multi-engine mobile application detection tasks. This scheme constructs an intelligent scheduling center that dynamically coordinates multiple heterogeneous detection engines through configurable policy rules and an intelligent scheduling model. The system comprehensively considers multiple dimensions such as sample attributes, engine status, and business objectives to achieve automatic and optimized allocation and full-process monitoring of detection tasks. Its core lies in breaking the limitations of static scheduling and significantly improving overall throughput efficiency, resource utilization, and task completion timeliness in large-scale automated detection scenarios through flexible strategies and dynamic feedback. The following provides a detailed introduction to this intelligent scheduling scheme for multi-engine mobile application detection tasks.

[0019] Figure 1 This is a flowchart illustrating the intelligent scheduling method for mobile application detection tasks across multiple engines, as described in this embodiment of the invention. Figure 1 As shown, the method includes the following steps: Step 101: Obtain the current multi-dimensional attribute data of the mobile application to be tested and the current engine status data of all engines; Step 102: Based on the current multi-dimensional attribute data and the relationship between the pre-configured task priority dimensions and priority weight values, determine the current multi-dimensional task priority weight value corresponding to the current multi-dimensional attribute data; Step 103: Based on the current multi-dimensional task priority weight value and the relationship between the pre-configured priority weight value range and the engine status weight value, determine the engine status weight value corresponding to the current multi-dimensional task priority weight value; Step 104: Determine the allocation value for each engine based on the current engine status data of all engines and the engine status weight value corresponding to the current multi-dimensional task priority weight value. Step 105: Select the engine with the largest allocation value as the optimal engine for the task to be detected, and dispatch the task to be detected to the queue of the optimal engine.

[0020] The intelligent scheduling method for mobile application detection tasks oriented towards multiple engines provided in this invention operates as follows: It acquires the current multi-dimensional attribute data of the mobile application to be detected and the current engine status data of all engines; based on the current multi-dimensional attribute data and the pre-configured relationship between task priority dimensions and priority weight values, it determines the current multi-dimensional task priority weight value corresponding to the current multi-dimensional attribute data; based on the current multi-dimensional task priority weight value and the pre-configured relationship between priority weight value ranges and engine status weight values, it determines the engine status weight value corresponding to the current multi-dimensional task priority weight value; based on the current engine status data of all engines and the engine status weight value corresponding to the current multi-dimensional task priority weight value, it determines the allocation value for each engine; and it selects the engine with the largest allocation value as the optimal engine for the task to be detected and dispatches the task to be detected to the queue of the optimal engine. This enables intelligent and flexible scheduling of multiple engines, improving the efficiency, flexibility, and reliability of mobile application detection. The following is a detailed description of this intelligent scheduling method for mobile application detection tasks oriented towards multiple engines.

[0021] This invention provides an intelligent scheduling method for mobile application detection tasks across multiple engines. Its core lies in designing an intelligent scheduling model driven by task priority, comprehensively considering the specialized capabilities of each engine, real-time load, and historical performance, ultimately outputting a task allocation scheme that optimizes the overall system efficiency. The main steps include: data preparation and input, multi-dimensional task priority evaluation, multi-dimensional engine scheduling evaluation, and outputting the optimal engine allocation object. Specifically: The data preparation and input phase, namely step 101 above and the steps prior to step 101 such as configuring various relationships, involves obtaining the complete input data required for decision-making, including: 1) Original attribute data of the task: including but not limited to application package name, version, source channel (such as app store, web crawler, manual upload), developer information, SDK list, application type, distribution platform, download volume, update time, etc.

[0022] In one embodiment, the current multi-dimensional attribute data includes: current task source attribute data, current key focus list data, and current dynamically adjusted attribute data.

[0023] In one embodiment, the dynamically adjusted attribute data includes one or any combination of application type data, distribution platform data, download volume data, and update time data.

[0024] 2) Engine Status: Includes engine identifier, real-time health status, current load, list of capability tags, and performance metrics based on historical statistics (such as average task processing time, success rate, etc.).

[0025] In one embodiment, the engine status data includes: engine detection time data, success rate data, pass rate data, and load data; the engine status weight values ​​include: engine detection time weight value, success rate weight value, pass rate weight value, and load weight value.

[0026] 3) Current scheduling strategy configuration: Loaded from the strategy rule base, including source weight, key focus list and weight, distribution platform weight, download volume weight, update time weight, region weight, etc.

[0027] In practice, when pre-configuring various relationships and strategies, the input data mentioned above can use historical data, such as historical multi-dimensional attribute data and historical engine status data. When scheduling mobile application detection tasks using the configured relationships, current data can be used, such as current multi-dimensional attribute data and current engine status data.

[0028] The multi-dimensional task priority evaluation phase, i.e., step 102 above, adopts a model of base values ​​and dynamic additions or subtractions (this model is used to execute step 102 above, and can be as follows). P t The formula transforms the raw attribute data of each task to be detected into a quantifiable and comparable comprehensive priority evaluation value, which serves as an important basis for subsequent engine scheduling evaluation. The calculation process consists of three levels: ① Basic Priority Layer: Determined by the source of the task, reflecting the inherent urgency of the task.

[0029] ② Static risk enhancement layer: Based on a pre-set list of key concerns, it directly reflects known risks.

[0030] ③ Dynamic Attribute Adjustment Layer: The final assessment and adjustment are made based on the application's real-time market attributes (such as influence, activity, and novelty) to reflect potential risks and regulatory value.

[0031] The formula for calculating priority (multi-dimensional task priority weight value) is as follows:

[0032] in To map a basic assessment value based on the source channel of the task, this part reflects the authority and urgency of the channel. To match the application with the key focus list, a risk bonus is awarded for each matched item, and these bonuses can be stacked. Based on the application's multi-dimensional attributes, this part calculates adjustment values ​​through table lookups or functions, enabling popular applications and newly updated applications to receive more attention.

[0033] As can be seen from the above, in one embodiment... Figure 2This is a flowchart illustrating an intelligent scheduling method for mobile application detection tasks oriented towards multiple engines, as shown in another embodiment of the present invention. Figure 2 As shown, based on the current multi-dimensional attribute data and the pre-configured relationship between task priority dimensions and priority weight values, the current multi-dimensional task priority weight value corresponding to the current multi-dimensional attribute data is determined, including step 102': Based on the current task source attribute data, the current key focus list data, the current dynamically adjusted attribute data, and the relationship between the pre-configured task priority dimensions and priority weight values, determine the current multi-dimensional task priority weight value corresponding to the current multi-dimensional attribute data.

[0034] In practice, the relationship between task priority dimension and priority weight value can be the table shown in Table 1 below. Of course, this relationship can also be a neural network recognition model for multi-dimensional task priority weight values. The input of this model can be the current multi-dimensional attribute data, and the output of this model can be the current multi-dimensional task priority weight value.

[0035] In the multi-dimensional engine scheduling and evaluation phase, namely steps 103 and 104 above, a multi-factor dynamic weighted evaluation model under basic capability filtering is adopted (this model is used to execute steps 103 and 104 above, and can be as follows). S i (The formula), based on the clearly defined task priorities. The process involves evaluating and selecting the most suitable engine for the task. After filtering out all engines that are faulty or offline, a quantitative evaluation is performed based on four core dimensions: detection efficiency, reliability, quality, and real-time load. The weights of each dimension are dynamically adjusted according to task priority. Finally, the optimal engine for each task is calculated. The comprehensive scheduling evaluation is as follows:

[0036] in Based on the basic capability matching coefficient (0 or 1), if the engine If the detection capability fully covers the detection requirements of the task, the value is 1; otherwise, it is 0. That is, in one embodiment, Figure 3 This is a flowchart illustrating an intelligent scheduling method for mobile application detection tasks oriented towards multiple engines, as shown in another embodiment of the present invention. Figure 3 As shown, the allocation value of each engine is determined based on the current engine status data of all engines and the engine status weight value corresponding to the current multi-dimensional task priority weight value. This may include step 104': determining the allocation value of each engine based on the basic capability matching coefficient, the current engine status data of all engines, and the engine status weight value corresponding to the current multi-dimensional task priority weight value.

[0037] For the evaluation of detection duration, based on the engine The evaluation value (range [0,1]) is normalized based on the historical average detection time of applications similar to task T. The shorter the processing time, the higher the evaluation.

[0038] To assess the success rate of detection, the engine The normalized value of the historical detection success rate (range [0,1]) is used. The higher the success rate, the higher the evaluation.

[0039] For the evaluation of the pass rate, engine The normalized value of the accuracy of the detection results in subsequent manual review (range [0,1]).

[0040] For current engine load assessment, based on engine The number of real-time concurrent tasks and the priority of tasks in the queue to be detected are greater than The load index is a normalized value (range [0,1]) calculated from the number of tasks, etc. The lighter the load, the higher the evaluation. That is, in one embodiment, the load data is determined based on the engine's real-time concurrent task count and the number of tasks in the engine's detection queue with a priority greater than the current multi-dimensional task priority weight value.

[0041] , , , These are weights for detection time, detection success rate, detection pass rate, and workload, respectively. For high-priority tasks, the detection success rate is given more weight, while for ordinary tasks, detection time and workload are given more weight. The sum of all weight coefficients must satisfy the following condition: + + + =1.

[0042] In the optimal engine allocation phase, i.e., step 105 above, the model outputs the final decision result, which is a clear mapping relationship.<Task_ID, Assigned_Engine_ID> The decision is sent to the task queue manager, which then dispatches the task instance to the target engine's execution queue and updates the relevant queue status.

[0043] To facilitate understanding of how this invention is implemented, the system architecture on which the above-mentioned intelligent scheduling method for mobile application detection tasks oriented towards multiple engines is based is described below.

[0044] The system architecture provided in this embodiment of the invention adopts a layered and decoupled modular design, and the overall architecture is divided into four layers: access layer, intelligent scheduling layer, execution layer and data support layer.

[0045] Access Layer: Responsible for receiving mobile application testing task requests from various channels, including regulatory platforms, app store crawlers, user complaint interfaces, and manual uploads. This layer standardizes the original requests, extracting or supplementing them to form a unified task object to be tested, including application files and metadata (such as package name, version, developer, source channel, version update time, application type, download volume, etc.).

[0046] Intelligent Scheduling Layer: This is the core layer of this invention, containing multiple collaborative functional modules that work together to make scheduling decisions. The main modules include: 1) Policy Management Module: Provides a graphical interface that allows administrators to dynamically configure and maintain all scheduling policies and rules.

[0047] 2) Scheduling configuration module: Provides a graphical interface that allows administrators to configure the task weight classification range and task scheduling strategies with different weights.

[0048] 3) Engine Monitor: Collects status information of each detected engine from the execution layer in real time to form engine status information.

[0049] 4) Intelligent Scheduling Model (This model implements the intelligent scheduling method for mobile application detection tasks oriented towards multiple engines provided in this embodiment of the invention): It performs in-depth analysis of the task objects to be detected. Combined with strategy configuration and priority analysis algorithms, it analyzes task priorities. Based on task priorities, detection engine scheduling strategies, and engine status, it runs a scheduling analysis algorithm and outputs the optimal engine allocation object.

[0050] 5) Task Queue Manager: Establishes task queues for each detection engine, receives analysis results from the intelligent scheduling model, dispatches tasks to the corresponding engine execution queues, and monitors the queue status.

[0051] Execution Layer: Composed of multiple heterogeneous third-party or self-developed automated detection engines, it provides a unified interface to the scheduling layer via API. It receives tasks from the scheduling layer, initiates detection, reports status (idle, faulty, current load), and returns detection results.

[0052] Data Support Layer: This layer includes databases and knowledge bases to support intelligent scheduling. It mainly consists of a policy rule base, an engine knowledge base, and a historical database. The policy rule base stores all policy rules set through the configuration interface. The historical database stores historical detection task information and engine detection records (detection duration, success rate, result summary, etc.). The engine knowledge base stores the engine's capability profile, including the application types it excels at detecting and the detection items it supports.

[0053] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0054] Example of key strategy configuration: For clarity, assume the following initial configuration was performed through the strategy configuration and management module (the relationship between pre-configured task priority dimensions and priority weight values) as shown in Table 1 below: Table 1

[0055] The engine scheduling evaluation weight strategy (the relationship between the pre-configured priority weight value range and the engine status weight value) is shown in Table 2 below: Table 2

[0056] The engine status (current engine status data for all engines) is shown in Table 3 below: Table 3

[0057] In specific implementation, this embodiment of the invention further includes extracting engine status parameter values ​​(e.g., engine detection time parameters 30, 15, 20, or load 2, 8, 10, etc.) from the current engine status data of all engines in Table 3 above. These values ​​are used subsequently in conjunction with the engine detection time weight value, success rate weight value, pass rate weight value, and load weight value, based on the aforementioned... S i The formula determines the allocation value for each engine.

[0058] In step 101 above: task submission and data preparation, assuming a detection task T1 (mobile application to be detected) from the app store arrives, its attributes, after being standardized by the access layer, are currently shown in Table 4 below: Table 4

[0059] In step 102 above: Multi-dimensional task priority evaluation calculates the current multi-dimensional task priority weight value P for task T1 based on the configured algorithm. For the base evaluation value, the source is "App Store," resulting in an evaluation value of 90. Regarding static risk enhancement, developer "Developer A" is on the list of key monitored developers, adding 20 evaluation values. For dynamic attribute adjustments, the application type is "Finance," adding 15 evaluation values; downloads of 50 million (≥10 million) add 15 evaluation values; and the update time is within 5 hours (<24 hours), adding 20 evaluation values. The comprehensive priority (current multi-dimensional task priority weight value) P = 90 + 20 + 15 + 15 + 20 = 160 evaluation values.

[0060] In steps 103 and 104 above: multi-dimensional engine scheduling evaluation, the intelligent scheduling model loads the states of all engines. Assume engines A, B, and C are all healthy and their capabilities cover the mandatory detection items of task T1 (Mc is 1 for all). Based on P=160, a weight configuration is adopted: Wt=0.15, Ws=0.45, Wq=0.3, Wl=0.1. That is, based on the current multi-dimensional task priority weight value and the relationship between the pre-configured priority weight value range and the engine state weight value, the engine state weight value corresponding to the current multi-dimensional task priority weight value is determined. The evaluation of each engine is calculated: Engine A: Sa=1 (0.15 0.6 + 0.45 0.95 + 0.30 0.92 + 0.10 0.8) = 0.8895; Engine B: Sb=1 (0.15 1.0 + 0.45 0.88 + 0.30 0.75 + 0.10 0.2) = 0.761; Engine C: Sc=1 (0.15 0.8 + 0.45 0.90 + 0.30 0.80 + 0.10 0.5) = 0.815.

[0061] In step 105 above: the optimal engine allocation object is output, comparing Sa=0.8895, Sb=0.761, and Sc=0.815. Although engine B is the fastest, due to the high priority of the current task, the algorithm prioritizes success rate and quality, and engine B is overloaded. Engine A, due to its significant advantages in success rate and quality, and its relatively idle state, receives the highest evaluation. Therefore, the intelligent scheduling model outputs the allocation object: assigning task T1 to engine A.

[0062] In practice, the relationship between the priority weight value range and the engine state weight value can be the table shown in Table 2 below. Of course, this relationship can also be a neural network recognition model for the engine state weight value. The input of this model can be the current multi-dimensional task priority weight value, and the output of this model can be the engine state weight value (current engine state weight value) corresponding to the current multi-dimensional task priority weight value.

[0063] In step 105 above: Task execution and feedback, the task queue manager dispatches T1 to the queue of engine A. After performing the detection, engine A reports the results. The system records the status and time of this task execution in the historical database for subsequent updates to engine A's average detection time and success rate profile.

[0064] In summary, the beneficial effects of the technical solution provided by the embodiments of the present invention are as follows: 1. Significantly improved efficiency: By finely matching tasks with the engine, unreasonable task allocation is reduced, the average waiting and processing time of tasks is shortened, and the overall throughput is improved.

[0065] 2. Optimized resource utilization: Based on real-time load scheduling and elastic scaling capabilities, the load of all connected detection engines is balanced, avoiding resource idleness and bottlenecks, and maximizing the return on hardware investment.

[0066] 3. Agile Business Response: Flexible strategy configuration enables the platform to respond quickly to sudden security incidents or regulatory requirements (such as emergency testing of a batch of applications), ensuring that critical tasks are completed with priority.

[0067] 4. Enhanced system robustness: It has automatic fault detection, task migration and retry capabilities, which ensures the high availability and stability of the testing pipeline and reduces the reliance on manual intervention.

[0068] 5. Intelligent Management: The administrator's business experience is transformed into configurable strategies and quantifiable model parameters, reducing the complexity of daily operation and maintenance management and making the scheduling decision-making process more scientific and interpretable.

[0069] In summary, the key technical point of this invention lies in constructing a dynamic and adaptive intelligent scheduling model (executing the intelligent scheduling method for mobile application detection tasks oriented towards multiple engines provided in this invention), based on a strategy-driven collaborative decision-making mechanism. The primary key point is the dynamic multi-dimensional quantification model of task priority (i.e., the model using a base value and dynamic addition or subtraction). This model breaks the limitations of traditional source-based classification, and through configurable strategies, integrates task source, static risk list matching results, and dynamic attributes to generate a comprehensive multi-dimensional task priority weight value, accurately characterizing the business urgency and potential risk value of each task. Secondly, there is the engine comprehensive scheduling evaluation algorithm linked to task priority. While ensuring that the engine meets the task requirements, the algorithm evaluates the engine's detection efficiency (duration), reliability (success rate), result quality (pass rate), and real-time load, and innovatively makes the weights of each dimension (Wt, Ws, Wq) a function of the task priority (P), thereby realizing dynamic adjustment of the scheduling strategy according to task importance: high-priority tasks automatically favor high-reliability engines, while ordinary tasks focus on throughput efficiency and load balancing.

[0070] All parameters, including the aforementioned priority calculation rules and scheduling evaluation weights, can be dynamically adjusted and combined using a graphical management interface. This transforms the entire intelligent scheduling behavior from hard-coded logic into a system that can be directly managed and optimized by business personnel. Furthermore, by continuously collecting task execution result data and automatically updating the engine's historical performance profile, the system provides a dynamically evolving data foundation for the evaluation model or unit, enabling scheduling decisions to continuously self-optimize as the system operates.

[0071] This invention implements: 1) a method for intelligent scheduling of mobile application detection tasks for multiple engines, specifically including a method for dynamic calculation of multi-dimensional task priorities, a method for comprehensive evaluation of engines in conjunction with task priorities, and an optimization decision-making and allocation process based on the evaluation; 2) a system for implementing the above method, particularly an intelligent scheduling architecture including a strategy configuration management module, an engine status monitor, an intelligent scheduling model, and a task queue manager; 3) the specific connotation and implementation method of the strategy configuration, especially the flexible configuration and management methods of strategy items such as task priority calculation rules and priority-engine evaluation weight mapping tables; 4) a complete technical solution implemented by the method and system that can coordinate multiple heterogeneous detection engines for dynamic and optimized task allocation. By protecting these points, the aim is to establish the exclusivity of the overall technical solution of this invention in improving the resource utilization, overall throughput, and response timeliness of key tasks in automated detection systems.

[0072] This invention also provides an intelligent scheduling device for mobile application detection tasks oriented towards multiple engines, as described in the following embodiments. Since the principle by which this device solves the problem is similar to the intelligent scheduling method for mobile application detection tasks oriented towards multiple engines, the implementation of this device can refer to the implementation of the intelligent scheduling method for mobile application detection tasks oriented towards multiple engines; repeated details will not be elaborated further.

[0073] Figure 4 This is a schematic diagram of the intelligent scheduling device for mobile application detection tasks oriented towards multiple engines in an embodiment of the present invention, as shown below. Figure 4 As shown, the device includes: Acquisition Unit 01 is used to acquire the current multi-dimensional attribute data of the mobile application to be detected and the current engine status data of all engines; The multi-dimensional task priority evaluation unit 02 is used to determine the current multi-dimensional task priority weight value corresponding to the current multi-dimensional attribute data based on the current multi-dimensional attribute data and the pre-configured relationship between task priority dimensions and priority weight values. Engine state weight value determination unit 03 is used to determine the engine state weight value corresponding to the current multi-dimensional task priority weight value based on the current multi-dimensional task priority weight value and the relationship between the pre-configured priority weight value range and the engine state weight value. The allocation value determination unit 04 is used to determine the allocation value of each engine based on the current engine status data of all engines and the engine status weight value corresponding to the current multi-dimensional task priority weight value. The optimal engine determination unit 05 is used to select the engine with the largest allocation value as the optimal engine for the task to be detected, and dispatch the task to be detected to the queue of the optimal engine.

[0074] In one embodiment, the current multi-dimensional attribute data includes: current task source attribute data, current key focus list data, and current dynamically adjusted attribute data; the multi-dimensional task priority evaluation unit is specifically used for: Based on the current task source attribute data, the current key focus list data, the current dynamically adjusted attribute data, and the relationship between the pre-configured task priority dimensions and priority weight values, determine the current multi-dimensional task priority weight value corresponding to the current multi-dimensional attribute data.

[0075] In one embodiment, the dynamically adjusted attribute data includes one or any combination of application type data, distribution platform data, download volume data, and update time data.

[0076] In one embodiment, the engine status data includes: engine detection time data, success rate data, pass rate data, and load data; the engine status weight values ​​include: engine detection time weight value, success rate weight value, pass rate weight value, and load weight value.

[0077] In one embodiment, the load data is determined based on the engine's real-time concurrent task count and the number of tasks in the engine's detection queue whose priority is greater than the current multi-dimensional task priority weight value.

[0078] In one embodiment, the allocation value determination unit is specifically used for: Based on the basic capability matching coefficient, the current engine status data of all engines, and the engine status weight value corresponding to the current multi-dimensional task priority weight value, the allocation value of each engine is determined.

[0079] Based on the aforementioned inventive concept, such as Figure 5 As shown, the present invention also proposes a computer device 500, including a memory 510, a processor 520, and a computer program 530 stored in the memory 510 and executable on the processor 520. When the processor 520 executes the computer program 530, it implements the aforementioned intelligent scheduling method for mobile application detection tasks oriented towards multiple engines.

[0080] This invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described intelligent scheduling method for mobile application detection tasks oriented towards multiple engines.

[0081] This invention also provides a computer program product, which includes a computer program that, when executed by a processor, implements the above-described intelligent scheduling method for mobile application detection tasks oriented towards multiple engines.

[0082] In this embodiment of the invention, the intelligent scheduling scheme for mobile application detection tasks oriented towards multiple engines operates as follows: It acquires the current multi-dimensional attribute data of the mobile application to be detected and the current engine status data of all engines; based on the current multi-dimensional attribute data and the pre-configured relationship between task priority dimensions and priority weight values, it determines the current multi-dimensional task priority weight value corresponding to the current multi-dimensional attribute data; based on the current multi-dimensional task priority weight value and the pre-configured relationship between priority weight value ranges and engine status weight values, it determines the engine status weight value corresponding to the current multi-dimensional task priority weight value; based on the current engine status data of all engines and the engine status weight value corresponding to the current multi-dimensional task priority weight value, it determines the allocation value for each engine; and it selects the engine with the largest allocation value as the optimal engine for the task to be detected and dispatches the task to be detected to the queue of the optimal engine. This enables intelligent and flexible scheduling of multiple engines, improving the efficiency, flexibility, and reliability of mobile application detection.

[0083] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0084] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0085] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0086] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0087] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for intelligent scheduling of mobile application detection tasks for multi-engine applications, characterized in that, include: Obtain the current multi-dimensional attribute data of the mobile application to be tested and the current engine status data of all engines; Based on the current multi-dimensional attribute data and the relationship between the pre-configured task priority dimensions and priority weight values, determine the current multi-dimensional task priority weight value corresponding to the current multi-dimensional attribute data; Based on the current multi-dimensional task priority weight value and the relationship between the pre-configured priority weight value range and the engine status weight value, determine the engine status weight value corresponding to the current multi-dimensional task priority weight value; The allocation value for each engine is determined based on the current engine status data of all engines and the engine status weight value corresponding to the current multi-dimensional task priority weight value. The engine with the largest allocation value is selected as the optimal engine for the task to be detected, and the task to be detected is dispatched to the queue of the optimal engine.

2. The method as described in claim 1, characterized in that, The current multi-dimensional attribute data includes: current task source attribute data, current key focus list data, and current dynamically adjusted attribute data; based on the current multi-dimensional attribute data and the pre-configured relationship between task priority dimensions and priority weight values, the current multi-dimensional task priority weight value corresponding to the current multi-dimensional attribute data is determined, including: Based on the current task source attribute data, the current key focus list data, the current dynamically adjusted attribute data, and the relationship between the pre-configured task priority dimensions and priority weight values, determine the current multi-dimensional task priority weight value corresponding to the current multi-dimensional attribute data.

3. The method as described in claim 2, characterized in that, The dynamically adjusted attribute data includes one or any combination of application type data, distribution platform data, download volume data, and update time data.

4. The method as described in claim 1, characterized in that, The engine status data includes: engine detection time data, success rate data, pass rate data, and load data; the engine status weight values ​​include: engine detection time weight value, success rate weight value, pass rate weight value, and load weight value.

5. The method as described in claim 4, characterized in that, The load data is determined based on the engine's real-time concurrent task count and the number of tasks in the engine's detection queue whose priority is greater than the current multi-dimensional task priority weight value.

6. The method as described in claim 1, characterized in that, Based on the current engine status data of all engines and the engine status weight value corresponding to the current multi-dimensional task priority weight value, determine the allocation value for each engine, including: Based on the basic capability matching coefficient, the current engine status data of all engines, and the engine status weight value corresponding to the current multi-dimensional task priority weight value, the allocation value of each engine is determined.

7. A smart scheduling device for mobile application detection tasks oriented towards multiple engines, characterized in that, include: The acquisition unit is used to acquire the current multi-dimensional attribute data of the mobile application to be detected and the current engine status data of all engines; The multi-dimensional task priority evaluation unit is used to determine the current multi-dimensional task priority weight value corresponding to the current multi-dimensional attribute data based on the current multi-dimensional attribute data and the pre-configured relationship between task priority dimensions and priority weight values. The engine status weight value determination unit is used to determine the engine status weight value corresponding to the current multi-dimensional task priority weight value based on the current multi-dimensional task priority weight value and the relationship between the pre-configured priority weight value range and the engine status weight value. The allocation value determination unit is used to determine the allocation value of each engine based on the current engine status data of all engines and the engine status weight value corresponding to the current multi-dimensional task priority weight value. The optimal engine determination unit is used to select the engine with the largest allocation value as the optimal engine for the task to be detected, and to dispatch the task to be detected to the queue of the optimal engine.

8. 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 method of any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method of any one of claims 1 to 6.

10. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the method of any one of claims 1 to 6.

Images