Multi-task cooperative processing method and system of industrial vehicle-mounted computer

Abstract

The invention provides a multi-task cooperative processing method and system for an industrial vehicle-mounted computer, and relates to the field of data processing, and the method comprises the steps: extracting an original time sequence, and carrying out the preprocessing of the original time sequence, and obtaining a smooth driving behavior sequence; extracting original features and performing component analysis to obtain a driver habit feature vector; a plurality of driving behavior modes are obtained through classification, and a priority distribution rule is constructed; extracting real-time driving characteristics, matching the real-time driving characteristics with the driving behavior mode, and calling a priority allocation rule to obtain a first task execution sequence; adjusting the first task execution sequence to obtain a second task execution sequence, allocating resources according to the second task execution sequence and a priority allocation rule, and performing resource optimization to obtain an optimized resource allocation scheme; and adjusting the second task execution sequence to obtain a third task execution sequence. The real-time performance of the key task is ensured, the resource utilization rate of the system is improved, and the real-time performance and the resource efficiency of the vehicle-mounted system are effectively balanced.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and in particular to a multi-task collaborative processing method and system for an industrial vehicle-mounted computer. Background Art

[0002] As a core component in the field of intelligent transportation and autonomous driving, industrial vehicle computers play a key role in improving vehicle operating efficiency and safety. Their multi-tasking collaborative processing capabilities directly determine the real-time response and resource optimization effects of the vehicle system in complex driving environments.

[0003] Driver behavior data can reflect the driver's personalized driving habits. Therefore, integrating historical driving data can be used to perform personalized optimization for the driver, forming a driving plan that is more in line with the driver's habits. However, current industrial vehicle-mounted computers often face problems such as uneven resource capacity allocation, insufficient dynamic adjustment of task priorities, and low data processing efficiency when performing multi-task collaborative processing. These problems make it difficult for vehicle systems to simultaneously meet the requirements of task timeliness and processing accuracy in high-load scenarios, especially when integrating historical driving data for personalized optimization. These problems together constitute the core challenge faced by industrial vehicle-mounted computers in multi-task collaborative processing: how to achieve dynamic optimization of multi-task collaborative execution based on driver habit characteristics within limited resource capacity, while balancing the contradiction between functional degradation and high-speed processing.

[0004] Therefore, how to achieve multi-task collaborative processing based on the driver's driving habits under limited resource capacity is a key issue that needs to be urgently solved in the field of industrial vehicle computers. Summary of the Invention

[0005] In view of this, the present invention proposes a multi-task collaborative processing method and system for an industrial vehicle-mounted computer to solve the problem that the industrial vehicle-mounted computer cannot simultaneously meet the requirements of task timeliness and processing accuracy when performing multi-task collaborative processing.

[0006] The technical solution of the present invention is achieved as follows: According to a first aspect, an embodiment of the present invention provides a multi-task collaborative processing method for an industrial vehicle-mounted computer, the method comprising: Obtain historical driving data, extract the original time series from the historical driving data, preprocess the original time series data, and obtain a smooth driving behavior sequence; Extracting original features from the smooth driving behavior sequence, and determining a driver habit feature vector carrying the driver's driving habits from the original features; The driver's habit feature vector is classified into several driving behavior patterns, and a corresponding priority allocation rule is constructed for each driving behavior pattern. The priority allocation rule includes the tasks involved, the order of each task, and the priority weight. acquiring real-time driving data, extracting real-time driving features from the real-time driving data, matching the real-time driving features with driving behavior patterns in a behavior pattern library to obtain a matching driving behavior pattern, retrieving a corresponding priority allocation rule based on the matching driving behavior pattern, and obtaining a first task execution order based on the priority allocation rule; Determine whether there is a circular dependency between tasks according to the priority allocation rule, adjust the execution order of the first tasks according to the circular dependency and the priority weights to obtain the execution order of the second tasks, allocate resources according to the execution order of the second tasks and the priority allocation rule, and obtain an initial resource allocation plan; Obtain the resource occupation threshold of the task according to the initial resource allocation plan, adjust the initial resource allocation plan according to the resource occupation threshold, and obtain the optimized resource allocation plan; The second task execution sequence is adjusted according to the optimized resource allocation scheme to obtain a third task execution sequence, and the tasks are executed according to the third task execution sequence.

[0007] In conjunction with the first aspect, in a first embodiment of the first aspect, the driver's habit feature vector is behaviorally classified to obtain a plurality of driving behavior patterns, and a corresponding priority allocation rule is constructed for each driving behavior pattern, specifically including: The driver habit feature vectors are clustered to obtain several clusters. When the variance within the cluster is determined to be lower than a preset threshold, the clusters with variances lower than the preset threshold are merged to obtain the initial behavior pattern. Each cluster represents a pattern in the initial behavior pattern, and each pattern contains several driver habit feature vectors. The behavior weight of each driver's habit feature vector in the initial behavior pattern is calculated by the entropy method, and the initial behavior pattern with behavior weight is divided into patterns by the hierarchical clustering method to obtain the driving behavior pattern; determining the distance between each driving behavior pattern according to the behavior weight, and merging the driving behavior patterns whose mutual distance is less than the first preset value when the distance is determined to be less than a first preset value; A corresponding priority allocation rule is constructed for each driving behavior pattern, and the driving behavior pattern is stored in a pre-set behavior pattern library and the priority allocation rule is stored in a pre-set rule library.

[0008] In combination with the first embodiment of the first aspect, in the second embodiment of the first aspect, acquiring real-time driving data, extracting real-time driving features from the real-time driving data, matching the real-time driving features with driving behavior patterns in a behavior pattern library to obtain matching driving behavior patterns, retrieving corresponding priority allocation rules based on the matching driving behavior patterns, and obtaining the first task execution order based on the priority allocation rules specifically includes: Acquire real-time driving data by a time series sampling method to obtain a first data stream carrying timestamp information; Performing moving average filtering on the first data stream to obtain a second data stream; Performing principal component analysis on the second data stream to obtain a third data stream containing real-time feature vectors; determining a Euclidean distance of real-time feature vectors in the third data stream, and if it is determined that the Euclidean distance exceeds a preset environment complexity threshold, performing vector normalization processing on the real-time feature vectors that exceed the preset environment complexity threshold to generate a real-time feature vector; Determine the Euclidean distance between the real-time driving feature and each driving behavior pattern in the behavior pattern library, and determine the driving behavior pattern that matches the real-time driving feature based on the Euclidean distance; According to the matching driving behavior pattern, the corresponding priority allocation rule is retrieved from the rule library, a task priority table is generated according to the priority allocation rule, and the first task execution order is obtained according to the task priority table.

[0009] In combination with the first aspect, in a third implementation of the first aspect, determining whether there is a circular dependency among tasks according to a priority allocation rule, adjusting the execution order of the first tasks according to the circular dependency and the priority weights to obtain the execution order of the second tasks, and allocating resources according to the second task execution order and the priority allocation rule to obtain the initial resource allocation plan specifically includes: Determine the dependencies between tasks according to the priority allocation rules, and determine the estimated duration of each task based on the dependencies and the completion time of the tasks; Based on the estimated duration and dependency order of the tasks, a topological sorting sequence of the tasks is constructed. Based on the topological sorting sequence, whether there is a cyclic dependency between the tasks is determined. If a dependency cycle is determined to exist, the execution order of the first tasks is adjusted based on the dependency cycle and the priority weights to obtain the execution order of the second tasks. Obtain resource capacity, allocate resources using a forward recursive strategy based on the second task execution order and resource capacity, and if it is determined that resource capacity is insufficient, reallocate resources based on the estimated task duration and priority weight to obtain an initial resource allocation plan.

[0010] In combination with the first aspect, in a fourth implementation of the first aspect, obtaining a resource occupancy threshold of a task according to the initial resource allocation plan, adjusting the initial resource allocation plan according to the resource occupancy threshold, and obtaining an optimized resource allocation plan specifically include: According to the initial resource allocation plan, the resource occupancy ratio of all tasks whose response time does not exceed the preset time is obtained to obtain the resource occupancy threshold; When it is determined that the resource occupancy threshold is lower than the preset standard, a genetic algorithm is used to reallocate resource capacity to obtain an optimized resource allocation plan.

[0011] In combination with the first aspect, in a fifth implementation of the first aspect, adjusting the execution order of the second tasks according to the optimized resource allocation scheme to obtain the execution order of the third tasks specifically includes: According to the optimized resource allocation plan and the conflict detection algorithm, it is determined whether there is resource competition between tasks. If resource competition or conflict exists, a conflict-free task execution path is determined. Through the conflict-free task execution path and the estimated duration of the task, a linear programming algorithm is used to adjust the allocation of resource capacity to obtain the allocation efficiency with the highest resource utilization. According to the allocation efficiency and the conflict-free task execution path, the second task execution order is adjusted to obtain the third task execution order.

[0012] In combination with the first aspect, in a sixth implementation of the first aspect, the method further includes the following steps: According to the third task execution order, the delay of the task whose priority weight is lower than the preset priority is obtained. When it is determined that the delay exceeds the preset degradation threshold, the resource capacity of the task whose priority weight is lower than the preset priority is reduced to obtain the fourth task execution order, and the tasks are executed according to the fourth task execution order.

[0013] In combination with the first aspect, in a seventh implementation of the first aspect, extracting original features from the smooth driving behavior sequence and determining a driver habit feature vector carrying the driver's driving habits from the original features specifically includes: Align the data in the smooth driving behavior sequence to a unified time axis, and extract original features from the aligned smooth driving behavior sequence; the original features include acceleration change rate, steering angular velocity, and braking frequency; Normalize the acceleration rate of change and steering angular velocity, and discretize and normalize the braking frequency to obtain standardized features; Determine the variance between the standardized features. If it is determined that the variance between the standardized features exceeds a preset variance, calculate the feature weight corresponding to each standardized feature by using an entropy method. If it is determined that the variance between the standardized features does not exceed the preset variance, equally divide the feature weight corresponding to each standardized feature. The standardized features with feature weights are subjected to dimensionality reduction processing to obtain dimensionality reduction features, the information retention rate of the dimensionality reduction features is determined, and the dimensionality reduction features whose information retention rate exceeds the preset retention rate are used as the driver habit feature vectors.

[0014] In combination with the first aspect, in an eighth implementation of the first aspect, acquiring historical driving data, extracting an original time series sequence from the historical driving data, and preprocessing the original time series data to obtain a smoothed driving behavior sequence specifically includes: Obtain historical driving data and extract original time series from the historical driving data; Perform data cleaning on the original time series, delete duplicate values and outliers in the original time series, and mark missing points in the original time series to obtain the first time series; Performing a sliding average filtering process on the first time series to remove noise data in the first time series to obtain a second time series; Performing linear interpolation on the second time series to fill in missing points in the second time series to obtain a third time series; The third time series is subjected to time series data analysis to extract driving behavior characteristics, which are then processed by fast Fourier transform to determine the frequency distribution of the driving behavior characteristics. Smoothed driving behavior characteristics are obtained based on the frequency distribution. The driving behavior characteristics include the distribution of acceleration change rate, steering angle change rate, and braking frequency.

[0015] According to a second aspect, an embodiment of the present invention provides a multi-task collaborative processing system for an industrial vehicle-mounted computer, the system comprising: The sequence acquisition module is used to obtain historical driving data, extract the original time series sequence from the historical driving data, and preprocess the original time series data to obtain a smooth driving behavior sequence; A habit acquisition module is used to extract original features from the smooth driving behavior sequence and determine a driver habit feature vector carrying the driver's driving habits from the original features; The behavior classification module is used to classify the driver's habit feature vectors into several driving behavior patterns and construct corresponding priority allocation rules for each driving behavior pattern; the priority allocation rules include the included tasks, the order of each task, and the priority weight; a first sequence module, configured to obtain real-time driving data, extract real-time driving features from the real-time driving data, match the real-time driving features with driving behavior patterns in a behavior pattern library to obtain matching driving behavior patterns, retrieve corresponding priority allocation rules based on the matching driving behavior patterns, and obtain a first task execution order based on the priority allocation rules; a second order module, configured to determine whether there is a circular dependency among tasks according to a priority allocation rule, adjust the execution order of the first tasks according to the circular dependency and the priority weights to obtain a second task execution order, and allocate resources according to the second task execution order and the priority allocation rule to obtain an initial resource allocation plan; The resource optimization module is used to obtain the resource occupation threshold of the task according to the initial resource allocation plan, adjust the initial resource allocation plan according to the resource occupation threshold, and obtain the optimized resource allocation plan; The sequence determination module is used to adjust the second task execution sequence according to the optimized resource allocation plan to obtain a third task execution sequence, and execute the tasks according to the third task execution sequence.

[0016] The multi-task collaborative processing method and system of the industrial vehicle-mounted computer of the present invention has the following advantages over the prior art: By extracting the original time series from the historical driving data and preprocessing the original time series data to obtain a smooth driving behavior sequence, the original features are extracted from the smooth driving behavior sequence, and the driver habit feature vector containing the driver's driving habits is accurately extracted from the original features through principal component analysis. Then, by performing two behavioral classifications on the driver habit feature vector, a number of driving behavior patterns are obtained. Such driving behavior patterns can provide personalized driving suggestions for drivers and help optimize the classification and management of driving behaviors. Next, corresponding priority allocation rules are constructed for each driving behavior pattern, and real-time driving features are extracted from real-time driving data. The real-time driving features are compared with the behavior pattern library. The driving behavior pattern is matched to obtain a matching driving behavior pattern and, based on the matching driving behavior pattern, the corresponding priority allocation rule is retrieved to obtain the first task execution order. The third task execution order is then optimized in combination with the task timeliness requirements and the optimal task execution time to obtain the final third task execution order. The third task execution order can reasonably allocate resource capacity, effectively avoid resource waste and ensure the completion of high-priority tasks, realize high-speed collaborative processing of multiple tasks, and can adaptively adjust the vehicle system task scheduling strategy according to the driver's behavior characteristics and the current driving environment. While ensuring the real-time performance of key tasks, it improves the utilization rate of system resources and effectively balances the real-time performance and resource efficiency of the vehicle system. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1This is a flow chart of a multi-task collaborative processing method for an industrial vehicle-mounted computer according to the present invention; Figure 2 This is a second flow chart of the multi-task collaborative processing method of the industrial vehicle-mounted computer of the present invention; Figure 3 This is a third flow chart of the multi-task collaborative processing method of the industrial vehicle-mounted computer of the present invention; Figure 4 This is a fourth flow chart of the multi-task collaborative processing method of the industrial vehicle-mounted computer of the present invention; Figure 5 This is a fifth flow chart of the multi-task collaborative processing method of the industrial vehicle-mounted computer of the present invention; Figure 6 This is a sixth flow chart of the multi-task collaborative processing method of the industrial vehicle-mounted computer of the present invention; Figure 7 It is a structural diagram of the multi-task collaborative processing system of the industrial vehicle-mounted computer of the present invention. DETAILED DESCRIPTION

[0019] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0020] As a core component in the field of intelligent transportation and autonomous driving, industrial vehicle computers play a key role in improving vehicle operating efficiency and safety. Their multi-tasking collaborative processing capabilities directly determine the real-time response and resource optimization effects of the vehicle system in complex driving environments.

[0021] Driver behavior data can reflect the driver's personalized driving habits. Therefore, integrating historical driving data can be used to perform personalized optimization for the driver, forming a driving plan that is more in line with the driver's habits. However, current industrial vehicle-mounted computers often face problems such as uneven resource capacity allocation, insufficient dynamic adjustment of task priorities, and low data processing efficiency when performing multi-task collaborative processing. These problems make it difficult for vehicle systems to simultaneously meet the requirements of task timeliness and processing accuracy in high-load scenarios, especially when integrating historical driving data for personalized optimization. These problems together constitute the core challenge faced by industrial vehicle-mounted computers in multi-task collaborative processing: how to achieve dynamic optimization of multi-task collaborative execution based on driver habit characteristics within limited resource capacity, while balancing the contradiction between functional degradation and high-speed processing.

[0022] Specifically, current industrial onboard computers generally lack efficient dynamic modeling mechanisms when extracting driver habit feature vectors from historical driving data, making it difficult to reflect subtle changes in driving behavior in real time. Furthermore, there is a significant conflict between task timeliness and resource capacity allocation. For example, high-priority tasks may be delayed due to resource competition, while low-priority tasks may consume excessive resources, resulting in a decrease in the overall data processing performance of the vehicle system.

[0023] Therefore, how to achieve multi-task collaborative processing based on the driver's driving habits under limited resource capacity is a key issue that needs to be urgently solved in the field of industrial vehicle computers.

[0024] The multi-task collaborative processing method for industrial vehicle computers provided in this manual is designed to accurately extract driving behavior patterns from historical driving data, dynamically calculate the optimal task execution time based on task timeliness requirements, and achieve high-speed collaborative processing of multiple tasks through resource capacity reallocation and function degradation.

[0025] See also Figure 1 , Figure 1 A flowchart of a multi-task collaborative processing method for an industrial vehicle-mounted computer according to an embodiment of the present invention is shown. The method may include the following steps: S101: Acquire historical driving data, extract original time series from the historical driving data, and preprocess the original time series data to obtain a smoothed driving behavior sequence.

[0026] Historical driving data can be pre-stored in electronic devices or acquired externally. For example, electronic devices can acquire data from various types of onboard sensors. For example, when a vehicle is driving on urban roads, various onboard sensors can record historical driving data including acceleration, steering angle, braking frequency, and other information.

[0027] On-board sensors typically include accelerometers, gyroscopes, and brake pedal position sensors. Accelerometers record the vehicle's acceleration values on the X, Y, and Z axes, gyroscopes capture dynamic changes in steering angles, and brake sensors determine braking action based on pedal pressure or position. In acquiring acceleration, steering angle, and braking data from on-board sensors, these sensors can collect data at a preset acquisition frequency (e.g., 10Hz), recording acceleration, steering angle, and braking status every 0.1 seconds. This high-frequency acquisition ensures the temporal continuity of the data, providing rich basic information for subsequent analysis.

[0028] There is no restriction on the specific form of obtaining the historical driving data. It only needs to ensure that the electronic device can obtain the historical driving data.

[0029] In this embodiment, the original time series contains information such as historical acceleration, historical steering angles, and historical braking frequency. When extracting the original time series containing information such as historical acceleration, historical steering angles, and historical braking frequency from historical driving data, the original time series often contains noisy data. This is caused by abnormal spikes caused by jitter of the on-board sensors, omissions caused by interruptions in data collection, and so on. In the case that the original time series also contains noisy data, in this embodiment, the original time series is preprocessed to remove the noise data.

[0030] Preferably, preprocessing methods include, but are not limited to, deleting duplicate values and outliers, sliding average filtering, interpolation, and fast Fourier transforms. For example, if a segment of historical acceleration data is 0, 2, 2, 10, 2, 10, it can be clearly seen that the historical acceleration data deviates from the normal range and can be considered an outlier and deleted. If a segment of historical acceleration data contains duplicate values, such as consecutive 2s, 2s, these can be merged into a single valid point, and the duplicate values deleted. If data is missing at a certain moment, such as a null value in a data sequence containing historical braking frequencies, this point is marked as a missing data point.

[0031] S102: Extracting original features from the smooth driving behavior sequence, and determining a driver habit feature vector carrying the driver's driving habits from the original features. In this embodiment, the original features include acceleration change rate, steering angular velocity, and braking frequency.

[0032] In this embodiment, the degree of feature differentiation between each of the aforementioned raw features is determined, and a corresponding feature weight is assigned to each raw feature based on the degree of differentiation. Subsequently, a principal component analysis algorithm is used to reduce the dimensionality of the weighted raw features, retaining principal components with information retention rates exceeding a preset retention rate. Therefore, the driver habit feature vector generated in this manner preserves the driver's driving habits and can be used to analyze driving style. For example, a high acceleration rate and high steering angle velocity in the vector may indicate aggressive driving, while a low braking frequency indicates stable driving. This multi-dimensional analysis of the feature vector provides a reliable basis for subsequent driving behavior assessment.

[0033] S103: Classify the driver's habit feature vector to obtain a number of driving behavior patterns. A corresponding priority allocation rule is constructed for each driving behavior pattern. It is understood that a mapping relationship exists between the constructed driving behavior patterns and the corresponding priority allocation rules. The driving behavior patterns may include aggressive, stable, and conservative driving patterns.

[0034] Preferably, the driving behavior pattern can be stored in a pre-set behavior pattern library and the priority allocation rule can be stored in a pre-set rule library to facilitate the subsequent matching of the driving behavior pattern and the retrieval of the priority allocation rule.

[0035] In this embodiment, behavior classification is performed twice. The first behavior classification is to divide the driver's habit feature vector into initial behavior patterns based on similarity. The second behavior classification is to perform a second division based on the classification results obtained in the first behavior classification to obtain the driving behavior pattern, thereby refining the granularity of the behavior classification.

[0036] The priority allocation rules include the tasks included in the driving behavior mode, the order of each task, and the priority weight. For example, for a smooth driving mode, the corresponding priority allocation rule is "cruise control > navigation prompts > entertainment system"; for an aggressive driving mode, the corresponding priority allocation rule is "brake assist > warning prompts > navigation prompts", so as to respond to potential risks in a timely manner. In other words, each task has a corresponding priority weight.

[0037] S104: Real-time driving data is acquired, real-time driving features are extracted from the real-time driving data, and the real-time driving features are matched with driving behavior patterns to obtain a matching driving behavior pattern. A corresponding priority allocation rule is retrieved based on the matching driving behavior pattern, and a first task execution order is determined based on the priority allocation rule. Because the first task execution order is based on the priority allocation rule, it can meet task timeliness requirements.

[0038] In this embodiment, step S104 specifically includes: S1041. Real-time driving data is acquired using a time series sampling method to obtain a first data stream carrying timestamp information, where the first data stream is a structured data stream. Specifically, the real-time driving data can be acquired using the time series sampling method. The time series sampling method collects the data at a fixed frequency, such as 10 Hz, and each piece of data is timestamped.

[0039] S1042. Perform moving average filtering on the first data stream to remove noise data in the first data stream and smooth the data to obtain a second data stream. The second data stream obtained in this way not only retains the real-time driving behavior trend but also removes the sudden noise. For example, if the difference between the real-time acceleration, real-time steering angle, real-time braking frequency, etc. and the previous and next data exceeds a set threshold, it can be replaced with the average value of the previous and next data through moving average filtering. In a sharp turn scenario, the original steering angle data may fluctuate due to vibration. The moving average filtering process can filter out high-frequency noise and retain only the low-frequency signal reflecting the actual movement of the vehicle. After such processing, the second data stream can better reflect the actual driving intention and improve the accuracy of subsequent analysis.

[0040] S1043. Perform principal component analysis on the second data stream to obtain a third data stream containing real-time feature vectors. A principal component analysis algorithm is used to extract real-time feature vectors of key features such as real-time acceleration, real-time steering angle, and real-time braking frequency from the second data stream, generating a high-dimensional feature vector containing at least three dimensions to obtain the third data stream.

[0041] Preferably, a time-domain analysis method can be used to perform the principal component analysis described above. That is, the covariance matrix of the data can be calculated to extract the first three principal components to form a high-dimensional feature vector. For example, in urban driving scenarios, the principal components may reflect the strong correlation between acceleration and braking frequency. The three-dimensional vector of the third data stream may be [0.65, 0.32, 0.15]. This reduces data redundancy and highlights key behavioral characteristics.

[0042] S1044: If it is determined that the Euclidean distance of the real-time feature vectors in the third data stream exceeds a preset environmental complexity threshold, perform vector normalization on the real-time feature vectors that exceed the preset environmental complexity threshold to generate a standardized real-time feature vector, thereby obtaining the real-time driving features. It is understood that in complex driving environments, such as congested roads or sharp turns, the feature vectors of the third data set may fluctuate significantly due to environmental changes.

[0043] In this embodiment, a preset environmental complexity threshold, such as 0.5, can be pre-set. By calculating the Euclidean distance between vectors, if it exceeds the preset environmental complexity threshold, vector normalization is performed, mapping the vector values to the interval [0, 1]. For example, a vector [0.8, 0.4, 0.2] is normalized to [0.62, 0.31, 0.15]. The normalized real-time driving characteristics are more suitable for subsequent behavioral pattern analysis and can adapt to different driving scenarios.

[0044] Among them, time series sampling ensures the temporal integrity of the data, moving average filtering improves data quality, principal component analysis focuses on core features, and vector normalization enhances environmental adaptability. The processing from step S1041 to step S1044 jointly supports the accurate construction and acquisition of real-time driving features, providing a reliable foundation for subsequent applications.

[0045] S1045: Determine the Euclidean distance between the real-time driving feature and each driving behavior pattern in the behavior pattern library, and determine the driving behavior pattern that matches the real-time driving feature based on the Euclidean distance, that is, the current driving behavior pattern.

[0046] After completing the above steps, we can obtain real-time driving features for Euclidean distance calculation. This feature vector retains the core information of the data while reducing dimensionality. When calculating Euclidean distance, it is necessary to compare the real-time feature vector with the pre-stored driving behavior patterns in the behavior pattern library.

[0047] If the calculated Euclidean distance to a specific driving behavior pattern is less than a preset distance, the pattern matching rule is used to determine the current driving behavior pattern. This matching rule is intuitive and efficient, enabling rapid identification of behavior types. The behavior pattern library may contain multiple driving behavior patterns. For example, the vector corresponding to a stable driving pattern is (0.2, 0.03, 0.5), while the vector corresponding to an aggressive driving pattern is (0.8, 0.1, 1.2).

[0048] Specifically, assuming the real-time driving feature vector is (0.3, 0.05, 0.8), and the Euclidean distance from the stable driving mode is small and less than the preset distance, the current driving mode is determined to be stable. It should be noted that the threshold setting needs to be adjusted according to the actual scenario. For example, urban roads may be more stringent, while highways can be more relaxed. The rule base plays a central role in generating the task priority table based on the current driving behavior pattern.

[0049] S1046. According to the matching driving behavior pattern, the corresponding priority allocation rule is retrieved from the rule library, and a task priority table is generated according to the tasks, task sequence, and task weights contained in the priority allocation rule. In this way, the first task execution order can be obtained according to the task priority table.

[0050] After determining the priority assignment rules, the system generates a first-order execution order based on the sorting algorithm and priority weights, such as cruise control with a weight of 90, navigation with 70, and entertainment with a weight of 50. This dynamic adjustment ensures that critical tasks are prioritized, improving driving efficiency and significantly enhancing the system's adaptability, ensuring that task assignments are closely aligned with the driving environment. In complex driving environments, this complete process, from raw signals to task sequencing, effectively supports real-time decision-making. Matching driving behavior patterns and prioritizing tasks further optimizes the logic and timeliness of task execution.

[0051] S105: Determine whether there are circular dependencies between tasks based on the priority allocation rules. Adjust the execution order of the first tasks based on the circular dependencies and the priority weights to obtain a second task execution order. Allocate resources based on the second task execution order and the priority allocation rules to obtain an initial resource allocation plan. The second task execution order quantifies the task duration and dependency order, ensuring a rigorous scheduling logic, thereby achieving the optimal task execution time.

[0052] S106: Obtain the resource occupation threshold of the task according to the initial resource allocation plan, adjust the initial resource allocation plan according to the resource occupation threshold, and obtain an optimized resource allocation plan. The resource occupation threshold reflects the degree of system resource shortage and helps identify potential resource bottlenecks.

[0053] In this embodiment, the resource occupancy thresholds for time-sensitive tasks are specifically obtained. This is determined by taking the resource capacity allocations for these time-sensitive tasks from the initial allocation plan, and then calculating the resource capacity occupancy ratios for each assigned task. If the resource occupancy threshold falls below a preset threshold, a genetic algorithm is used to initialize the population and iterate the optimization to generate an optimized allocation plan.

[0054] In this embodiment, high-timeliness tasks generally refer to tasks with strict response time requirements, such as real-time data processing or critical business operations, that is, tasks with a response time that does not exceed a preset time.

[0055] S107. Adjust the second task execution order according to the optimized resource allocation plan to obtain a third task execution order, i.e., a final task execution schedule. Finally, execute the tasks according to the third task execution order to achieve layer-by-layer optimization from dependency to resource allocation.

[0056] In this embodiment, step S107 specifically includes: S1071. Determine whether there is resource competition between tasks based on the optimized resource allocation plan and a conflict detection algorithm. If resource competition, i.e., conflict, is determined, determine a conflict-free task execution path to adjust the execution order of the second task.

[0057] S1072. Using a linear programming algorithm, the allocation of resource capacity is adjusted based on the conflict-free task execution path and the estimated duration of the task to obtain the allocation efficiency with the highest resource utilization. Then, the second task execution order is adjusted based on the allocation efficiency and the conflict-free task execution path to obtain the third task execution order.

[0058] For example, a project includes Task A (requirements analysis, estimated duration 2 days), Task B (design, estimated duration 3 days, dependent on Task A), Task C (coding, estimated duration 5 days, dependent on Task B), and Task D (testing, estimated duration 2 days, dependent on Task C). Task A is determined to have the highest priority because it has no dependencies; Task B is next in priority because it depends on Task A. Similarly, the second-order task execution order is: Task A → Task B → Task C → Task D. This second-order task execution order ensures that tasks are arranged in order of dependency, preventing scheduling failures due to missing dependencies.

[0059] Based on the optimized resource allocation plan, a conflict detection algorithm is used to identify resource contention. Suppose the project has two servers: Tasks A and B each require one server, and Task C requires two servers. In the second task execution sequence, Task C requires two servers. However, if Task D is scheduled in parallel with Task C, a conflict may arise due to insufficient servers. The conflict detection algorithm identifies this issue and adjusts the scheduling order, for example, deferring Task D until Task C is completed, to ensure conflict-free resource allocation. A linear programming algorithm is used to optimize resource allocation based on conflict-free task execution paths and estimated durations. Continuing with this project, assuming the total construction period must be within 12 days, the linear programming algorithm analyzes the resource utilization and time constraints of each task. For example, since Task C is longer in duration, two servers may be prioritized to ensure progress, while Tasks A and B can share a single server to reduce resource utilization. This optimization ensures maximum resource utilization while meeting the construction period.

[0060] The third task execution order generated based on the allocation efficiency and conflict-free task execution path can be: Task A is scheduled for days 1-2, Task B for days 3-5, Task C for days 6-10, and Task D for days 11-12. The final execution order for the third task is verified to meet time constraints, such as a total duration of 12 days and no task dependencies being violated. If so, the final task execution schedule is generated. If not, the scheduling order or resource allocation can be retroactively adjusted.

[0061] Priority calculation ensures clear task logic, conflict detection avoids resource bottlenecks, linear programming improves resource efficiency, and the final schedule ensures controllable project duration. Each link supports each other, forming a complete task scheduling system suitable for scenarios with high timeliness requirements, such as software development or production line task management.

[0062] The specific implementation of the conflict detection algorithm can be a time-window-based resource occupancy check. For example, it checks whether the time windows of tasks C and D overlap. If so, it calculates whether the resource requirements exceed two servers. The conflict detection algorithm is intuitive and easy to implement, making it particularly suitable for scenarios with limited resources. During linear programming optimization, a virtual resource allocation scenario can be introduced. For example, if an additional server is added, the analysis can be performed to see whether the time required for task C can be shortened, thereby optimizing the overall schedule. Linear programming optimization provides greater flexibility in resource allocation while maintaining consistency with the core solution.

[0063] The multi-task collaborative processing method of the industrial vehicle-mounted computer of the present invention extracts the original time series sequence from the historical driving data, pre-processes the original time series data to obtain a smooth driving behavior sequence, then extracts the original features from the smooth driving behavior sequence, and then accurately extracts the driver habit feature vector containing the driver's driving habits from the original features through principal component analysis. Then, by performing two behavioral classifications on the driver habit feature vector, a number of driving behavior patterns are obtained. Such driving behavior patterns can provide personalized driving suggestions for the driver, which helps to optimize the classification and management of driving behavior. Next, a corresponding priority allocation rule is constructed for each driving behavior pattern, and then real-time driving features are extracted from the real-time driving data to be used as the basis for the real-time driving behavior. The characteristics are matched with the driving behavior patterns in the behavior pattern library to obtain the matching driving behavior patterns and, based on the matching driving behavior patterns, retrieve the corresponding priority allocation rules to obtain the first task execution order. The third task execution order is then optimized in combination with the task timeliness requirements and the optimal task execution time to obtain the final third task execution order. The third task execution order can reasonably allocate resource capacity, effectively avoid resource waste and ensure the completion of high-priority tasks, realize high-speed collaborative processing of multiple tasks, and can adaptively adjust the vehicle system task scheduling strategy according to the driver's behavior characteristics and the current driving environment. While ensuring the real-time performance of key tasks, it improves the utilization rate of system resources and effectively balances the real-time performance and resource efficiency of the vehicle system.

[0064] See also Figure 2 , the method may further comprise the following steps: S2011: Acquire historical driving data and extract original time series from the historical driving data. The original time series can be directly extracted from the vehicle's CAN bus, etc.

[0065] S2012: Perform data cleaning on the original time series sequence, delete duplicate values and outliers in the original time series sequence, and mark missing points in the original time series sequence to obtain a first time series sequence. Data cleaning can delete duplicate values and outliers in the original time series sequence, mark missing points in the original time series sequence, and thus obtain a first time series sequence. The first time series sequence thus obtained has been preliminarily normalized, with redundant and erroneous data points deleted, ensuring the accuracy of subsequent data analysis.

[0066] S2013: Perform a sliding average filter on the first time series to remove noise data in the first time series, thereby obtaining a second time series. The first time series is denoised using a sliding average filter. If the difference between a point in the first time series and the preceding and following points exceeds a preset difference, the point is replaced with the average of the preceding and following points, thereby obtaining the second time series. This denoising method smoothes out the sudden changes in the first time series, reflecting a more realistic historical driving behavior trend and helping to reduce misjudgments in subsequent data analysis.

[0067] S2014: Perform linear interpolation on the second time series to fill in missing points in the second time series, thereby obtaining a third time series. Linear interpolation is used to fill in missing points in the second time series. If there are marked missing points in the second time series, linear interpolation is calculated based on the values of the preceding and following points to obtain the third time series. This method of filling in missing points preserves data continuity, avoids abrupt jumps, and improves the credibility of the analysis.

[0068] S2015. Perform time series data analysis on the third time series sequence to extract driving behavior characteristics. Perform fast Fourier transform on the driving behavior characteristics to determine a frequency distribution of the driving behavior characteristics. Smoothed driving behavior characteristics are obtained based on the frequency distribution. In this embodiment, the driving behavior characteristics include the distribution of acceleration change rate, steering angle change rate, and braking frequency.

[0069] Frequency analysis of the acceleration rate of change may reveal a predominance of low-frequency components, indicating smooth driving behavior. A high concentration of high-frequency components in the acceleration rate of change may indicate frequent abrupt maneuvers. Sudden acceleration or braking can cause a sharp increase in the rate of change, while smooth driving results in gradual changes. Based on this, in this embodiment, when performing time series data analysis on the third time series, the primary features extracted are the acceleration rate of change (e.g., the rate of change from 1 to 2), the steering angle rate of change (e.g., the rate of change from 10 to 15), and the distribution characteristics of braking frequency (e.g., a braking frequency of 2 to 3 times per minute). These driving behavior characteristics reflect the smoothness of driving. By analyzing the third time series data, it is possible to identify periodic patterns in driving behavior, which can be helpful in evaluating driver style or optimizing autonomous driving algorithms.

[0070] Fast Fourier transform can convert these driving behavior characteristics into frequency distribution. Driving behavior characteristics that reflect the smoothness of driving can be used to determine whether a driver tends to operate smoothly. This can be used for insurance pricing or driver training, not only improving data utilization but also providing reliable support for intelligent transportation systems.

[0071] From step S2011 to step S2015, a progressive data processing approach is adopted from cleaning to feature extraction, ensuring data quality while mining the deep information contained in the data.

[0072] S202: Extracting original features from the smooth driving behavior sequence and determining the driver's habit feature vector from the original features. For details, refer to step S102.

[0073] S203: Classify the driver's habit feature vector to obtain a number of driving behavior patterns, and construct a corresponding priority allocation rule for each driving behavior pattern. For details, refer to step S103.

[0074] S204: Real-time driving data is acquired, real-time driving features are extracted from the real-time driving data, and the real-time driving features are matched with driving behavior patterns in the behavior pattern library to obtain a matching driving behavior pattern. A corresponding priority allocation rule is retrieved based on the matching driving behavior pattern, and the first task execution order is determined based on the priority allocation rule. For details, refer to step S104.

[0075] S205: Determine whether there is a circular dependency between tasks based on the priority allocation rule, adjust the execution order of the first tasks based on the circular dependency and the priority weights, obtain a second task execution order, and allocate resources based on the second task execution order and the priority allocation rule to obtain an initial resource allocation plan. For details, refer to step S105.

[0076] S206: Obtain the resource occupation threshold of the task according to the initial resource allocation plan, adjust the initial resource allocation plan according to the resource occupation threshold, and obtain an optimized resource allocation plan. For details, refer to step S106.

[0077] S207: Adjust the second task execution order according to the optimized resource allocation plan to obtain a third task execution order, and execute the tasks according to the third task execution order. For details, refer to step S107.

[0078] See also Figure 3 , the method may further comprise the following steps: S301: Acquire historical driving data, extract original time series from the historical driving data, and pre-process the original time series data to obtain a smoothed driving behavior sequence. For details, refer to step S101.

[0079] S3021: Align the data in the smooth driving behavior sequence onto a unified time axis, and extract original features from the smooth driving behavior sequence after data alignment.

[0080] In this embodiment, a timestamp alignment algorithm is used to process the time deviation between different on-board sensor data. Since different on-board sensors may have slight time synchronization errors, such as accelerometer data lagging behind brake data by 0.02 seconds, timestamp alignment will align all data to a unified time axis through linear interpolation or nearest neighbor method.

[0081] For example, at a certain moment, the acceleration is 2.3m / s² and the steering angle is 12°, but the braking data is missing. In this case, the value can be estimated to be 0 by interpolating the front and rear braking states. Timestamp alignment, a data alignment method, ensures the accuracy of subsequent feature calculations.

[0082] Alternatively, a sliding window approach can be used to calculate the acceleration rate of change, steering angular velocity, and braking frequency to obtain the required raw features, which can then be aggregated to form a raw feature set. For example, the acceleration rate of change is calculated by dividing the acceleration difference within a window by the time interval. For example, if the acceleration changes from 2.0m / s² to 2.5m / s² within the window, the acceleration rate of change is 0.5m / s³. Steering angular velocity is calculated similarly: if the steering angle changes from 10° to 15° within the window, the steering angular velocity is 5° / s³. Braking frequency is calculated by counting the number of brakes within the window. This feature extraction method effectively captures the dynamic changes in driving behavior.

[0083] S3022: Normalize the acceleration rate of change and steering angular velocity in the original features, and discretize and normalize the braking frequency to obtain standardized features. In this embodiment, the acceleration rate of change and steering angular velocity in the original feature set are normalized using a z-score normalization method, and the braking frequency is discretized and normalized using a frequency interval partitioning method to obtain standardized features. These standardized features are then aggregated to obtain a standardized feature set.

[0084] Preferably, the acceleration change rate and steering angular velocity are normalized by the z-score normalization method to eliminate dimensional differences and facilitate subsequent analysis. Assuming that the mean value of the acceleration change rate is 0.4m / s³, the standard deviation is 0.1m / s³, and the value after normalization is 1; the braking frequency is divided into three frequency intervals: low, medium, and high. Discretization simplifies the processing of frequency characteristics. For example, 0-1Hz is set as low frequency, 1-3Hz is set as medium frequency, and above 3Hz is set as high frequency. If a certain frequency is actually calculated to be 2Hz, it can be classified into the medium frequency range.

[0085] S3023. Determine the variance between the standardized features. If it is determined that the variance between the standardized features exceeds the preset variance, calculate the feature weight corresponding to each standardized feature using the entropy method. If it is determined that the variance between the standardized features does not exceed the preset variance, equally divide the feature weight corresponding to each standardized feature.

[0086] In this embodiment, a preset variance, such as 0.8, can be pre-set. If the variance of the standardized feature set exceeds the preset variance, the weight is calculated using the entropy method. The entropy method is based on the information entropy of the feature. Lower entropy indicates higher feature discrimination and a higher weight. If the variance does not exceed the preset variance, the weight is evenly distributed to 0.33. This dynamic weighting method highlights the contribution of key features.

[0087] S3024. Perform dimensionality reduction processing on the standardized features with feature weights to obtain reduced-dimensionality features, determine the information retention rate of the reduced-dimensionality features, and use the reduced-dimensionality features whose information retention rate exceeds a preset retention rate as the driver habit feature vector.

[0088] In this embodiment, a preset retention rate, such as 90%, can be set in advance. The standardized feature set with assigned feature weights is subjected to dimensionality reduction processing by the principal component analysis algorithm, and the principal components with an information retention rate higher than the preset retention rate are retained. Assuming that the standardized feature set contains three dimensions: acceleration change rate, steering angular velocity, and braking frequency, the principal component analysis may project the data onto two principal components, retaining 95% of the information. For example, the feature vector after dimensionality reduction may be expressed as [0.7, 0.4], reflecting the driver's stability and steering habits. This dimensionality reduction can reduce the computational complexity while retaining the core driving characteristics. In particular, if the data dimension of the dimensionality reduction feature is higher than the preset dimension, the high-dimensional dimensionality reduction feature is further regularized to obtain a feature vector of unified dimension. In this way, overfitting of high-dimensional data can be avoided and the computational complexity can be effectively reduced.

[0089] S303: Classify the driver's habit feature vector to obtain several driving behavior patterns, and construct a corresponding priority allocation rule for each driving behavior pattern. For details, refer to step S103.

[0090] S304: Real-time driving data is acquired, real-time driving features are extracted from the real-time driving data, and the real-time driving features are matched with driving behavior patterns in the behavior pattern library to obtain a matching driving behavior pattern. A corresponding priority allocation rule is retrieved based on the matching driving behavior pattern, and the first task execution order is determined based on the priority allocation rule. For details, refer to step S104.

[0091] S305: Determine whether there is a circular dependency between tasks based on the priority allocation rule, adjust the execution order of the first tasks based on the circular dependency and the priority weights, obtain a second task execution order, and allocate resources based on the second task execution order and the priority allocation rule to obtain an initial resource allocation plan. For details, refer to step S105.

[0092] S306: Obtain the resource occupation threshold of the task according to the initial resource allocation plan, adjust the initial resource allocation plan according to the resource occupation threshold, and obtain an optimized resource allocation plan. For details, refer to step S106.

[0093] S307: Adjust the second task execution order according to the optimized resource allocation plan to obtain a third task execution order, and execute the tasks according to the third task execution order. For details, refer to step S107.

[0094] See also Figure 4 , the method may further comprise the following steps: S401: Acquire historical driving data, extract original time series from the historical driving data, and pre-process the original time series data to obtain a smoothed driving behavior sequence. For details, refer to step S101.

[0095] S402: Extracting original features from the smooth driving behavior sequence and determining the driver's habit feature vector from the original features. For details, refer to step S102.

[0096] S4031. Cluster the driver's habit feature vector to obtain several clusters. If the variance within the cluster is determined to be lower than a preset threshold, the clusters with variances lower than the preset threshold are merged to obtain the initial behavior pattern. It can be understood that Each cluster represents a pattern within the initial behavioral patterns, and each pattern contains several driver habit feature vectors. Clustering algorithms in unsupervised learning generally serve as the basis for analysis by other algorithms. Clustering data can provide a holistic analysis of certain characteristics of the data. Specifically, a K-means clustering algorithm is used for unsupervised learning to partition driver habit feature vectors into multiple clusters based on internal similarities, maximizing intra-cluster similarity and minimizing inter-cluster similarity. Preferably, the initial number of clusters is set to three, representing aggressive, smooth, and conservative driving modes, respectively. An aggressive driving mode may be characterized by large speed fluctuations, high acceleration peaks, and frequent braking; a smooth driving mode exhibits gentle speed fluctuations and low braking frequency. After clustering, if the variance within a cluster is less than 0.01, it indicates that the samples within the cluster are highly similar and can be merged into adjacent clusters. For example, two conservative driving clusters can be merged into one cluster because their speed fluctuations are both less than 5 km / h. This merging improves the differentiation of behavioral patterns.

[0097] S4032. Calculate the behavioral weight of each driver habit feature vector in the initial behavioral pattern using the entropy method to analyze the contribution of the feature to the clustering result. Then, divide the initial behavioral pattern with the behavioral weight into patterns using the hierarchical clustering method to obtain the driving behavior pattern.

[0098] S4033. Determine the distance between each driving behavior pattern according to the behavior weight. If the distance is determined to be less than a first preset value, merge the driving behavior patterns whose mutual distance is less than the first preset value to obtain a final driving behavior pattern set.

[0099] In this embodiment, a first preset value can be set in advance, such as 0.05. When the initial behavior pattern is divided twice by the hierarchical clustering method, the distance between the driving behavior patterns can be calculated based on the feature weight. For example, if the speed feature weights of the two driving behavior patterns are both 0.3, the difference in acceleration weights is less than 0.1, and the distance is lower than the second preset value, they are merged into a single pattern. Assuming that there are 5 driving behavior patterns initially, they are merged into 3 driving behavior patterns after the secondary division to generate a driving behavior pattern set including aggressive, smooth, and mixed modes. This setting refines the granularity of the behavior classification. In particular, if the behavior weight of a driver's habit feature vector is lower than the pre-set second preset value, the corresponding feature is deleted to obtain the weight of each driver's habit feature vector in the initial behavior pattern.

[0100] In this embodiment, a second preset value, such as 0.25, can be pre-set. Assuming that the braking frequency weight is lower than the second preset value, the feature is deleted, and a feature weight set containing only speed, acceleration, and steering angle is generated. This screening method ensures that only the most discriminative features of the behavior pattern are retained. Taking the analysis of urban taxi driving behavior as an example, a driver frequently brakes during peak hours and has large acceleration changes, which is classified as an aggressive driving mode; while the speed is stable during low traffic hours at night, which is classified as a stable driving mode. By dividing the driving behavior patterns, personalized driving suggestions can be provided to the driver, which helps to optimize the classification and management of driving behavior.

[0101] S4034, constructing corresponding priority allocation rules for each driving behavior pattern, and storing the driving behavior pattern in a pre-set behavior pattern library and the priority allocation rules in a pre-set rule library. S404: Real-time driving data is acquired, real-time driving features are extracted from the real-time driving data, and the real-time driving features are matched with driving behavior patterns in the behavior pattern library to obtain a matching driving behavior pattern. A corresponding priority allocation rule is retrieved based on the matching driving behavior pattern, and a first task execution order is determined based on the priority allocation rule. For details, refer to step S104.

[0102] S405: Determine whether there is a circular dependency between tasks based on the priority allocation rule, adjust the execution order of the first tasks based on the circular dependency and the priority weights, obtain a second task execution order, and allocate resources based on the second task execution order and the priority allocation rule to obtain an initial resource allocation plan. For details, refer to step S105.

[0103] S406: Obtain the resource occupation threshold of the task according to the initial resource allocation plan, adjust the initial resource allocation plan according to the resource occupation threshold, and obtain an optimized resource allocation plan. For details, refer to step S106.

[0104] S407: Adjust the second task execution order according to the optimized resource allocation plan to obtain a third task execution order, and execute the tasks according to the third task execution order. For details, refer to step S107.

[0105] See also Figure 5 , the method may further comprise the following steps: S501: Acquire historical driving data, extract original time series from the historical driving data, and pre-process the original time series data to obtain a smoothed driving behavior sequence. For details, refer to step S101.

[0106] S502: Extracting original features from the smooth driving behavior sequence and determining the driver's habit feature vector from the original features. For details, refer to step S102.

[0107] S503: Classify the driver's habit feature vector to obtain a number of driving behavior patterns, and construct a corresponding priority allocation rule for each driving behavior pattern. For details, refer to step S103.

[0108] S504: Real-time driving data is acquired, real-time driving features are extracted from the real-time driving data, and the real-time driving features are matched with driving behavior patterns in the behavior pattern library to obtain a matching driving behavior pattern. A corresponding priority allocation rule is retrieved based on the matching driving behavior pattern, and a first task execution order is determined based on the priority allocation rule. For details, refer to step S104.

[0109] S5051. Determine the dependencies between tasks according to the priority allocation rule, and determine the estimated duration of each task based on the dependencies and the completion duration of the tasks.

[0110] In this embodiment, the task order and priority weight of the task are obtained from the task priority table. The task order and priority weight can reflect the dependency relationship between tasks. Then, the estimated duration of each task is calculated using the state transition equation, where the state transition equation is defined as: in, Indicates the The estimated duration of each task; Indicates the The estimated duration of the dependent tasks with dependencies between the first and second tasks. A task may have multiple dependent tasks with dependencies; Indicates the The completion time of a task.

[0111] For example, in a task scheduling scenario for an autonomous driving system, assume there are three tasks, A, B, and C: Task A is path planning, Task B is obstacle detection, and Task C is speed adjustment. The task priority table shows that Task A has a priority weight of 0.8, Task B has a priority weight of 0.7, and Task C has a priority weight of 0.5. The task order indicates that Task B depends on the results of Task A, while Task C depends on the output of Task B. Based on this, the state transition equation can be used to calculate the estimated durations of these three tasks. Specifically, assuming Task A takes 3 seconds to complete, Task B takes 2 seconds to complete, and Task C takes 1 second to complete, then Task B's estimated duration is 5 seconds, equal to Task A's 3 seconds plus Task C's 2 seconds. Task C's estimated duration is 6 seconds, equal to Task B's 5 seconds plus Task C's 1 second. This approach clearly quantifies task durations and dependency order, ensuring rigorous scheduling logic.

[0112] S5052. Construct a topologically sorted sequence of tasks based on the estimated duration and dependency order of the tasks, determine whether there is a cyclic dependency among the tasks based on the topologically sorted sequence, and if it is determined that there is a dependency cycle, adjust the execution order of the first tasks based on the dependency cycle and the priority weights to obtain a loop-free second task execution order.

[0113] When constructing a topological sorting sequence, if there are no circular dependencies, the first task execution order is directly Task A → Task B → Task C. Suppose a new Task D is added, and Task D depends on Task C, which in turn depends on Task D, forming a circular dependency. Priority weights must be adjusted. However, if Task C's priority weight of 0.5 is lower than Task D's priority weight of 0.6, Task D can be executed first, breaking the original dependency cycle. The first task execution order is adjusted to Task A → Task B → Task D → Task C, resulting in a new second task execution order. This adjustment ensures a non-cyclic task execution sequence, solves the task loop problem, optimizes the priority execution of critical tasks, and improves system response efficiency.

[0114] S5053. Obtain resource capacity, and allocate resources using a forward recursive strategy based on the second task execution order and resource capacity. If it is determined that the resource capacity is insufficient, reallocate the resources based on the estimated task duration and priority weight to obtain an initial resource allocation plan.

[0115] After obtaining the execution order and resource capacity of the second task, the forward recursive strategy can be used to allocate resources. Assume that the system has 10 computing units, and Task A requires 4 computing units, Task B requires 3 computing units, Task C requires 2 computing units, and Task D requires 3 computing units. The forward recursive strategy starts with Task A, allocating 4 computing units, leaving 6 computing units. Task B is then allocated 3 computing units, leaving 3 computing units. Task C is then allocated 2 computing units, leaving 1 computing unit. Task D requires 3 computing units but has insufficient resources. At this point, reallocation is performed based on the estimated duration and priority of the tasks. For example, reducing the number of computing units for Task C to 1 and freeing up 1 unit for D to ensure the smooth execution of all tasks.

[0116] In this embodiment, after the initial resource allocation plan is generated, its feasibility needs to be further verified. For example, if task A is a critical task, insufficient allocation may cause delays in path planning and affect subsequent tasks. By ensuring sufficient resources for task A through priority weights, other tasks are dynamically adjusted, which not only meets real-time requirements but also improves overall stability. This resource allocation plan effectively avoids resource waste and guarantees the completion of high-priority tasks. It can also adapt to the insertion of sudden tasks and enhance system flexibility. At the same time, by quantifying duration, optimizing sorting and dynamic allocation, it significantly improves the rationality and execution efficiency of task scheduling. For example, during peak hours, quickly adjusting resources to obstacle detection tasks can effectively reduce the risk of collision. This multi-faceted resource allocation plan, from core dependency analysis to resource optimization, forms a consistent scheduling logic to ensure the efficient operation of the autonomous driving system.

[0117] S506: Obtain the resource occupation threshold of the task according to the initial resource allocation plan, adjust the initial resource allocation plan according to the resource occupation threshold, and obtain an optimized resource allocation plan. For details, refer to step S106.

[0118] S507: Adjust the second task execution order according to the optimized resource allocation plan to obtain a third task execution order, and execute the tasks according to the third task execution order. For details, refer to step S107.

[0119] See also Figure 6 , the method may further comprise the following steps: S601: Acquire historical driving data, extract original time series from the historical driving data, and pre-process the original time series data to obtain a smoothed driving behavior sequence. For details, refer to step S101.

[0120] S602: Extracting original features from the smooth driving behavior sequence and determining the driver's habit feature vector from the original features. For details, refer to step S102.

[0121] S603: Classify the driver's habit feature vector to obtain a number of driving behavior patterns, and construct a corresponding priority allocation rule for each driving behavior pattern. For details, refer to step S103.

[0122] S604: Real-time driving data is acquired, real-time driving features are extracted from the real-time driving data, and the real-time driving features are matched with driving behavior patterns in a behavior pattern library to obtain a matching driving behavior pattern. A corresponding priority allocation rule is retrieved based on the matching driving behavior pattern, and a first task execution order is determined based on the priority allocation rule. For details, refer to step S104.

[0123] S605: Determine whether there is a circular dependency between tasks based on the priority allocation rule, adjust the execution order of the first tasks based on the circular dependency and the priority weights, obtain a second task execution order, and allocate resources based on the second task execution order and the priority allocation rule to obtain an initial resource allocation plan. For details, refer to step S105.

[0124] S606: Obtain a resource occupation threshold of the task according to the initial resource allocation plan, and adjust the initial resource allocation plan according to the resource occupation threshold to obtain an optimized resource allocation plan.

[0125] S6061. Obtain the resource occupancy ratio of all tasks whose response time does not exceed the preset time according to the initial resource allocation plan to obtain a resource occupancy threshold. In this embodiment, a high-timeliness task refers to a task with strict response time requirements, such as real-time data processing or critical business operations, that is, a task whose response time does not exceed the preset time. Specifically, receive the allocated amount of resource capacity for high-timeliness tasks in the initial allocation plan, and obtain the above-mentioned resource occupancy threshold by traversing the resource capacity occupancy ratio of each allocated task. If the resource occupancy threshold is lower than the preset standard, a genetic algorithm is used to initialize the population and iteratively optimize to generate an optimized allocation plan.

[0126] Consider a task scheduling system where Task A is real-time log analysis and is initially allocated 4 CPU cores. Task B is database synchronization and is initially allocated 2 CPU cores. Recording the resource allocation for each task provides baseline data for subsequent analysis. This approach ensures transparency in resource allocation and facilitates subsequent optimization. When analyzing resource allocation and capacity utilization ratios, each task's resource allocation can be compared to the total system resources. Assuming the system has 16 CPU cores, Task A utilizes 4 / 16 = 25%, and Task B utilizes 2 / 16 = 12.5%. Assuming both Task A and Task B are time-sensitive tasks, the resource utilization ratios of all time-sensitive tasks can be summed to determine the resource utilization threshold, such as 25% + 12.5% = 37.5%.

[0127] S6062: When it is determined that the resource occupancy threshold is lower than the preset standard, a genetic algorithm is used to reallocate resource capacity to obtain an optimized resource allocation solution.

[0128] It is understandable that if the new resource occupancy threshold is close to the preset standard, it indicates that the system resource utilization is more efficient, avoiding resource waste or task delays. In this embodiment, a preset standard can be set in advance, such as 50%. If the obtained resource occupancy threshold is lower than the preset standard, a genetic algorithm is used to optimize the allocation. The genetic algorithm initializes a set of resource allocation schemes as a population by simulating natural selection, and each generation iteratively optimizes through crossover and mutation operations. For example, the initial population includes Scheme 1: Task A is allocated 5 CPU cores and Task B is allocated 1 CPU core; Scheme 2: Task A is allocated 3 CPU cores and Task B is allocated 3 CPU cores. After iteration, a new scheme is generated: Task A is allocated 4 CPU cores and Task B is allocated 2 CPU cores, and the overall occupancy threshold is increased to 43.75%. Iterative optimization is a multi-round optimization process, through which resource allocation efficiency and task timeliness are balanced.

[0129] Furthermore, resource requirements can be mapped to resource pools based on the optimized allocation plan, taking into account the dynamic nature of the resource pool. A resource pool may contain different types of compute nodes, such as high-performance nodes and low-power nodes. Suppose the optimized plan requires that Task A be allocated 4 CPU cores. High-performance nodes are prioritized to meet timeliness requirements, while Task B is allocated 2 CPU cores, which can be handled by low-power nodes. This mapping process ensures maximum resource utilization while meeting task performance requirements. After reallocating resource capacity, the resource allocations for time-sensitive tasks are reviewed again to calculate the new resource utilization ratio. Assuming Task A still has 4 CPU cores, while Task B is adjusted to 3 CPU cores, the new utilization threshold is 4 / 16 + 3 / 16 = 43.75%. This ensures that resource allocation is more aligned with task requirements. By optimizing resource allocation for time-sensitive tasks, a closed-loop process is formed. This closed-loop design ensures continuous improvement of the resource allocation plan to adapt to dynamic task demands.

[0130] S607: Adjust the second task execution order according to the optimized resource allocation plan to obtain a third task execution order, and execute the tasks according to the third task execution order. For details, refer to step S107.

[0131] After step S107, the method may further include: S108. According to the third task execution order, the delay of the task with a priority weight lower than the preset priority is obtained. When it is determined that the delay exceeds the preset degradation threshold, the resource capacity of the task with a priority weight lower than the preset priority is reduced to obtain the fourth task execution order, and the task is executed according to the fourth task execution order. Specifically, the delay of the task with a priority weight lower than the preset priority, i.e., the low-priority task, is obtained from the third task execution order. If the delay exceeds the preset degradation threshold, a trigger signal can be generated. Next, the rule engine is activated according to the trigger signal. The rule engine is responsible for obtaining the resource capacity pre-allocated to the above-mentioned low-priority task from the third task execution order, and then dynamically trimming the resource capacity through the computational amount trimming algorithm to determine the streamlined resource capacity. A lightweight task execution topology can be generated based on the streamlined resource capacity, and then a topology optimization tool is used to re-sort the dependencies in the topological sorting sequence to obtain an optimized topological sorting sequence. Thereafter, the task execution order is extracted from the optimized topological sorting sequence to obtain the final optimized task processing flow, i.e., the fourth task execution order. When the computational workload trimming algorithm is in effect, the processing workload of secondary data is trimmed, such as retaining only backups of key files, and reducing the computational workload, that is, the required resource capacity, to 50GB and 2,000 queries.

[0132] The lightweight topology reduces unnecessary dependency waiting and avoids idle resources. This adaptive dynamic optimization is particularly suitable for scenarios with tight resources or large fluctuations in task volume, and has strong practical value.

[0133] For example, suppose a low-priority task is data backup, and its normal delay is 10 minutes. The real-time monitoring tool collects delay data every 1 minute and finds that the delay of a certain execution reaches 15 minutes. The preset degradation threshold is 12 minutes, which has obviously exceeded the threshold. In this way, the above-mentioned trigger signal can be generated.

[0134] Preferably, the trigger signal can also be generated by comparing the real-time latency with historical latency. For example, by comparing it with the historical latency average (e.g., the average value of 9 minutes over the past week), the adaptive degradation condition can be determined. This comparison method is not only based on single-time data but also incorporates historical trends. This method can more accurately reflect the abnormal status of the task.

[0135] The following describes a system provided by an embodiment of the present invention. The system described below and the method described above can refer to each other.

[0136] See also Figure 7 , Figure 7 A schematic diagram of the structure of a multi-task collaborative processing system for an industrial vehicle-mounted computer according to an embodiment of the present invention is shown. The system may include: The sequence acquisition module 10 is used to acquire historical driving data, extract original time series from the historical driving data, and pre-process the original time series data to obtain a smooth driving behavior sequence.

[0137] The habit acquisition module 20 is used to extract original features from the smooth driving behavior sequence and determine the driver's habit feature vector from the original features.

[0138] The behavior classification module 30 is used to classify the driver's habit feature vector into behavior categories to obtain a number of driving behavior patterns, and to construct a corresponding priority allocation rule for each driving behavior pattern, wherein the priority allocation rule includes the included tasks, the order of each task, and the priority weight.

[0139] The first sequence module 40 is used to obtain real-time driving data, extract real-time driving features from the real-time driving data, match the real-time driving features with driving behavior patterns in the behavior pattern library to obtain matching driving behavior patterns, retrieve corresponding priority allocation rules based on the matching driving behavior patterns, and obtain the first task execution order based on the priority allocation rules.

[0140] The second order module 50 is configured to determine whether there is a circular dependency between tasks based on the priority allocation rule, adjust the execution order of the first tasks based on the circular dependency and the priority weights, and obtain a second task execution order. Resources are allocated based on the second task execution order and the priority allocation rule to obtain an initial resource allocation plan.

[0141] The resource optimization module 60 is configured to obtain a resource occupation threshold of a task according to the initial resource allocation plan, and adjust the initial resource allocation plan according to the resource occupation threshold to obtain an optimized resource allocation plan.

[0142] The sequence determination module 70 is configured to adjust the second task execution sequence according to the optimized resource allocation solution to obtain a third task execution sequence, and execute the tasks according to the third task execution sequence.

[0143] The multi-task collaborative processing system of the industrial vehicle-mounted computer of the present invention extracts the original time series sequence from the historical driving data, pre-processes the original time series data to obtain a smooth driving behavior sequence, then extracts the original features from the smooth driving behavior sequence, and then accurately extracts the driver habit feature vector containing the driver's driving habits from the original features through principal component analysis. Then, by performing two behavioral classifications on the driver habit feature vector, a number of driving behavior patterns are obtained. Such driving behavior patterns can provide personalized driving suggestions for the driver, which helps to optimize the classification and management of driving behavior. Next, a corresponding priority allocation rule is constructed for each driving behavior pattern, and then real-time driving features are extracted from the real-time driving data to be used as the basis for the real-time driving behavior. The characteristics are matched with the driving behavior patterns in the behavior pattern library to obtain the matching driving behavior patterns and, based on the matching driving behavior patterns, retrieve the corresponding priority allocation rules to obtain the first task execution order. The third task execution order is then optimized in combination with the task timeliness requirements and the optimal task execution time to obtain the final third task execution order. The third task execution order can reasonably allocate resource capacity, effectively avoid resource waste and ensure the completion of high-priority tasks, realize high-speed collaborative processing of multiple tasks, and can adaptively adjust the vehicle system task scheduling strategy according to the driver's behavior characteristics and the current driving environment. While ensuring the real-time performance of key tasks, it improves the utilization rate of system resources and effectively balances the real-time performance and resource efficiency of the vehicle system.

[0144] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A multi-task collaborative processing method for an industrial vehicle-mounted computer, characterized by: The method comprises: Obtain historical driving data, extract the original time series from the historical driving data, preprocess the original time series data, and obtain a smooth driving behavior sequence; Extracting original features from the smooth driving behavior sequence, and determining a driver habit feature vector carrying the driver's driving habits from the original features; The driver's habit feature vector is classified into several driving behavior patterns, and a corresponding priority allocation rule is constructed for each driving behavior pattern. The priority allocation rule includes the tasks involved, the order of each task, and the priority weight. acquiring real-time driving data, extracting real-time driving features from the real-time driving data, matching the real-time driving features with driving behavior patterns in a behavior pattern library to obtain a matching driving behavior pattern, retrieving a corresponding priority allocation rule based on the matching driving behavior pattern, and obtaining a first task execution order based on the priority allocation rule; Determine whether there is a circular dependency between tasks according to the priority allocation rule, adjust the execution order of the first tasks according to the circular dependency and the priority weights to obtain the execution order of the second tasks, allocate resources according to the execution order of the second tasks and the priority allocation rule, and obtain an initial resource allocation plan; Obtain the resource occupation threshold of the task according to the initial resource allocation plan, adjust the initial resource allocation plan according to the resource occupation threshold, and obtain the optimized resource allocation plan; The second task execution order is adjusted according to the optimized resource allocation scheme to obtain a third task execution order, and the tasks are executed according to the third task execution order.

2. The multi-task collaborative processing method for an industrial vehicle-mounted computer according to claim 1, characterized in that: The driver habit feature vector is subjected to behavioral classification to obtain a number of driving behavior patterns, and a corresponding priority allocation rule is constructed for each driving behavior pattern, specifically including: The driver habit feature vectors are clustered to obtain several clusters. When the variance within the cluster is determined to be lower than a preset threshold, the clusters with variances lower than the preset threshold are merged to obtain the initial behavior pattern. Each cluster represents a pattern in the initial behavior pattern, and each pattern contains several driver habit feature vectors. Determine the behavior weight of each driver's habit feature vector in the initial behavior pattern, divide the initial behavior pattern with the behavior weight into patterns using a hierarchical clustering method, and obtain the driving behavior pattern; determining the distance between each driving behavior pattern according to the behavior weight, and merging the driving behavior patterns whose mutual distance is less than the first preset value when the distance is determined to be less than a first preset value; A corresponding priority allocation rule is constructed for each driving behavior pattern, and the driving behavior pattern is stored in a pre-set behavior pattern library and the priority allocation rule is stored in a pre-set rule library.

3. The multi-task collaborative processing method of an industrial vehicle-mounted computer according to claim 2, characterized in that: The acquiring of real-time driving data, extracting real-time driving features from the real-time driving data, matching the real-time driving features with driving behavior patterns in a behavior pattern library to obtain a matching driving behavior pattern, retrieving a corresponding priority allocation rule based on the matching driving behavior pattern, and obtaining a first task execution order based on the priority allocation rule specifically includes: Acquire real-time driving data by a time series sampling method to obtain a first data stream carrying timestamp information; Performing moving average filtering on the first data stream to obtain a second data stream; Performing principal component analysis on the second data stream to obtain a third data stream containing real-time feature vectors; determining a Euclidean distance of real-time feature vectors in the third data stream, and if it is determined that the Euclidean distance exceeds a preset environment complexity threshold, performing vector normalization processing on the real-time feature vectors that exceed the preset environment complexity threshold to generate a real-time feature vector; Determine the Euclidean distance between the real-time driving feature and each driving behavior pattern in the behavior pattern library, and determine the driving behavior pattern that matches the real-time driving feature based on the Euclidean distance; According to the matching driving behavior pattern, the corresponding priority allocation rule is retrieved from the rule library, a task priority table is generated according to the priority allocation rule, and the first task execution order is obtained according to the task priority table.

4. The multi-task collaborative processing method of an industrial vehicle-mounted computer according to claim 1, characterized in that: The determining whether there is a circular dependency among tasks according to the priority allocation rule, adjusting the execution order of the first tasks according to the circular dependency and the priority weights to obtain the execution order of the second tasks, and allocating resources according to the second task execution order and the priority allocation rule to obtain the initial resource allocation plan specifically includes: Determine the dependencies between tasks according to the priority allocation rules, and determine the estimated duration of each task based on the dependencies and the completion time of the tasks; Based on the estimated duration and dependency order of the tasks, a topological sorting sequence of the tasks is constructed. Based on the topological sorting sequence, whether there is a cyclic dependency between the tasks is determined. If a dependency cycle is determined to exist, the execution order of the first tasks is adjusted based on the dependency cycle and the priority weights to obtain the execution order of the second tasks. Obtain resource capacity, allocate resources using a forward recursive strategy based on the second task execution order and resource capacity, and if it is determined that resource capacity is insufficient, reallocate resources based on the estimated task duration and priority weight to obtain an initial resource allocation plan.

5. The multi-task collaborative processing method of an industrial vehicle-mounted computer according to claim 1, characterized in that: The step of obtaining a resource occupation threshold of a task according to the initial resource allocation plan and adjusting the initial resource allocation plan according to the resource occupation threshold to obtain an optimized resource allocation plan specifically includes: According to the initial resource allocation plan, the resource occupancy ratio of all tasks whose response time does not exceed the preset time is obtained to obtain the resource occupancy threshold; When it is determined that the resource occupancy threshold is lower than the preset standard, a genetic algorithm is used to reallocate resource capacity to obtain an optimized resource allocation plan.

6. The multi-task collaborative processing method of an industrial vehicle-mounted computer according to claim 1, characterized in that: The adjusting the second task execution order according to the optimized resource allocation scheme to obtain a third task execution order, and executing the tasks according to the third task execution order specifically includes: According to the optimized resource allocation plan and the conflict detection algorithm, it is determined whether there is resource competition between tasks. If resource competition or conflict exists, a conflict-free task execution path is determined. Through the conflict-free task execution path and the estimated duration of the task, a linear programming algorithm is used to adjust the allocation of resource capacity to obtain the allocation efficiency with the highest resource utilization. According to the allocation efficiency and the conflict-free task execution path, the second task execution order is adjusted to obtain the third task execution order.

7. The multi-task collaborative processing method of an industrial vehicle-mounted computer according to claim 1, characterized in that: The method further comprises the following steps: According to the third task execution order, the delay of the task whose priority weight is lower than the preset priority is obtained. When it is determined that the delay exceeds the preset degradation threshold, the resource capacity of the task whose priority weight is lower than the preset priority is reduced to obtain the fourth task execution order, and the tasks are executed according to the fourth task execution order.

8. The multi-task collaborative processing method of an industrial vehicle-mounted computer according to claim 1, characterized in that: The step of extracting original features from the smooth driving behavior sequence and determining a driver habit feature vector carrying the driver's driving habits from the original features specifically includes: Align the data in the smooth driving behavior sequence to a unified time axis, and extract original features from the aligned smooth driving behavior sequence; the original features include acceleration change rate, steering angular velocity, and braking frequency; Normalize the acceleration rate of change and steering angular velocity, and discretize and normalize the braking frequency to obtain standardized features; Determine the variance between the standardized features. If it is determined that the variance between the standardized features exceeds a preset variance, calculate the feature weight corresponding to each standardized feature by using an entropy method. If it is determined that the variance between the standardized features does not exceed the preset variance, equally divide the feature weight corresponding to each standardized feature. The standardized features with feature weights are subjected to dimensionality reduction processing to obtain dimensionality reduction features, the information retention rate of the dimensionality reduction features is determined, and the dimensionality reduction features whose information retention rate exceeds the preset retention rate are used as the driver habit feature vectors.

9. The multi-task collaborative processing method of an industrial vehicle-mounted computer according to claim 1, characterized in that: The acquisition of historical driving data, extraction of original time series from the historical driving data, and preprocessing of the original time series data to obtain a smoothed driving behavior sequence specifically include: Obtain historical driving data and extract original time series from the historical driving data; Perform data cleaning on the original time series, delete duplicate values and outliers in the original time series, and mark missing points in the original time series to obtain the first time series; Performing a sliding average filtering process on the first time series to remove noise data in the first time series to obtain a second time series; Performing linear interpolation on the second time series to fill in missing points in the second time series to obtain a third time series; The third time series is subjected to time series data analysis to extract driving behavior characteristics, which are then processed by fast Fourier transform to determine the frequency distribution of the driving behavior characteristics. Smoothed driving behavior characteristics are obtained based on the frequency distribution. The driving behavior characteristics include the distribution of acceleration change rate, steering angle change rate, and braking frequency.

10. A multi-task collaborative processing system for an industrial vehicle-mounted computer, characterized by: The system comprises: The sequence acquisition module is used to obtain historical driving data, extract the original time series sequence from the historical driving data, and preprocess the original time series data to obtain a smooth driving behavior sequence; A habit acquisition module is used to extract original features from the smooth driving behavior sequence and determine a driver habit feature vector carrying the driver's driving habits from the original features; The behavior classification module is used to classify the driver's habit feature vectors into several driving behavior patterns and construct corresponding priority allocation rules for each driving behavior pattern; the priority allocation rules include the included tasks, the order of each task, and the priority weight; a first sequence module, configured to obtain real-time driving data, extract real-time driving features from the real-time driving data, match the real-time driving features with driving behavior patterns in a behavior pattern library to obtain matching driving behavior patterns, retrieve corresponding priority allocation rules based on the matching driving behavior patterns, and obtain a first task execution order based on the priority allocation rules; a second order module, configured to determine whether there is a circular dependency among tasks according to a priority allocation rule, adjust the execution order of the first tasks according to the circular dependency and the priority weights to obtain a second task execution order, and allocate resources according to the second task execution order and the priority allocation rule to obtain an initial resource allocation plan; The resource optimization module is used to obtain the resource occupation threshold of the task according to the initial resource allocation plan, adjust the initial resource allocation plan according to the resource occupation threshold, and obtain the optimized resource allocation plan; The sequence determination module is used to adjust the second task execution sequence according to the optimized resource allocation plan to obtain a third task execution sequence, and execute the tasks according to the third task execution sequence.

Images