Engineering big data analysis and integrated learning method based on graph neural network

Abstract

This invention discloses an integrated learning method for engineering big data analysis based on graph neural networks. This invention relates to the field of graph neural network technology and solves the problems of failing to accurately characterize operating habits, optimize instruction indexing logic, and intelligently allocate resources. By quantifying operating instructions into associated points, tracking operating routes, and verifying overlapping features, this invention can accurately extract the scene features and personal characteristics of operators. This achieves a deep binding between "operating instructions and personnel habits" and allows graph neural networks to quickly identify the operating preferences of different personnel. By splitting operating instruction segments and selecting the optimal index path, combined with the priority ranking of scene and personal features, high-frequency, highly associated operating instruction segments receive better index resource allocation.

Description

Technical Field

[0001] This invention relates to the field of graph neural network technology, specifically to an integrated learning method for engineering big data analysis based on graph neural networks. Background Technology

[0002] In the current field of engineering modeling, with the increasing complexity of projects and the expansion of multi-source data, traditional engineering big data analysis methods are gradually revealing significant shortcomings: On the one hand, engineering operations rely on the experience of operators, but there are significant differences in the operating habits and scenario adaptation logic of different personnel. Existing tools mostly adopt the "general command response" mode, which is difficult to match personalized operation needs, resulting in a lot of operational redundancy, low response efficiency, and even increased risk of operational errors due to the mismatch between tools and personnel habits. On the other hand, the indexing of engineering instructions and the allocation of computing power are mostly "undifferentiated," which neither take into account the high-frequency characteristics of the operation scenario nor the personal operation preferences of the personnel. This not only causes inefficient waste of computing resources, but also makes it difficult for the indexing response speed of core operations to meet the real-time requirements of large-scale engineering modeling.

[0003] Meanwhile, existing applications of graph neural networks in engineering often focus on macro-level scenarios such as structural analysis and data fusion, lacking in-depth modeling of the micro-interaction link of "operator-operation command-tool response". This makes it difficult to transform the implicit operational experience of personnel into the explicit intelligent logic of tools, resulting in a "fit gap" between engineering tools and actual operational needs. This makes it impossible to fully leverage the correlation modeling advantages of graph neural networks, and also makes it difficult to support the upgrade of engineering modeling from "tool assistance" to "intelligent collaboration".

[0004] Against this backdrop, there is an urgent need for an engineering big data analysis and integrated learning method that can accurately characterize operating habits, optimize instruction indexing logic, and intelligently allocate resources, in order to fill the gaps in existing technologies in personalized operation adaptation and efficient instruction response. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides an integrated learning method for engineering big data analysis based on graph neural networks, which solves the problems of failing to accurately characterize operating habits, optimize instruction indexing logic, and intelligently allocate resources.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for integrated learning and analysis of engineering big data based on graph neural networks, comprising the following steps: Step 1: Confirm the operation commands associated with the engineering modeling data in sequence, and set several associated points within the preset association framework, so that the associated points correspond to the operation commands, generating relevant standards belonging to the corresponding engineering modeling data. The specific method is as follows: Identify the existing operation commands from the engineering modeling data and record the total number G of operation commands. Generate G associated points within the set association framework and associate each associated point with each operation command in sequence. The association framework is a preset framework. The associated framework after binding the associated points with the operation instructions is recorded as the relevant standard for the corresponding engineering modeling data; Step 2: Confirm the operation records generated by individual operators within the traceability period, identify the sequence of operation instructions associated with different operation scenarios from the operation records, confirm the corresponding operation routes within relevant standards, confirm the scenario characteristics of the same operation scenario from the operation routes, and then confirm the individual characteristics of the operator from different operation scenarios. The specific method is as follows: Using the current time as the base time, a set of traceability cycles is identified, which are preset cycles. The operation records generated by the corresponding operators within the traceability cycle are extracted. Single records associated with the same operation scenario are extracted from the operation records, and the associated operation instructions are extracted from the single records. Several sets of operation instructions are sorted according to their time sequence. The operation instruction sequence associated with the corresponding single record is identified. The first operation instruction in the operation instruction sequence is identified at the associated point corresponding to the relevant standard. The associated point is used as the initial point. The associated points corresponding to other operation instructions are identified in turn. According to the sorting method of the operation instructions in the operation instruction sequence, the associated points corresponding to the subsequent operations are identified in turn. Starting from the initial point, the subsequently identified associated points are connected in turn to generate the operation route associated with the corresponding operation instruction sequence. The operation route associated with each single record is confirmed sequentially, and several operation routes are checked for overlap to identify overlapping segments and record the number of overlaps (CH). k Where k represents different overlapping road segments, and the total number of operation routes is recorded as ZS, satisfying: (CH k Sections with an overlap of ≥70% (÷ZS) are designated as habitual sections. The operation instructions associated with these habitual sections are sorted to confirm the habitual operation sequence. The confirmed sets of habitual operation sequences are then recorded as the scene characteristics of the corresponding operators in the corresponding operation scenarios. The specific methods for confirming personal characteristics in different operational scenarios are as follows: From multiple sets of habitual operation sequences associated with different operational scenarios, the habitual route segments associated with each habitual operation sequence are identified. The identified habitual route segments are then subjected to overlap verification to confirm overlapping habitual route segments. The number of overlaps generated by these overlapping habitual route segments is marked as CS. q Where q represents different habitually overlapping road segments, and the total number of habitual routes is recorded as XZ, satisfying (CS) qThe overlapping road segments with a ratio of ÷XZ)≥30% are recorded as characteristic road segments. The operation instructions associated with the characteristic road segments are sorted to confirm the habitual characteristic sequence. The confirmed sets of habitual characteristic sequences are recorded as the personal characteristics associated with the corresponding operators. Step 3: Identify several operation command segments from the scene features and personal characteristics confirmed by the operator, and confirm the path location associated with each operation command segment. Based on the path features of adjacent operation commands, confirm and record the index features of corresponding adjacent operation command segments. The specific method is as follows: From the personal characteristics of the corresponding operator, identify the sequence of operation instructions associated with each personal characteristic, and identify the operation instruction segments from front to back in the sequence of operation instructions. Each operation instruction segment contains only two operation instructions, and the identified operation instruction segments are recorded as personal instruction segments. Using the same confirmation method as individual command segments, the operation command segments associated with the scene features are confirmed and recorded as scene command segments; For individual command segments and scene command segments, determine the path positions of the preceding and following operation commands, and determine the index path associated between the two sets of path positions. If there is only one set of index paths, record it as the index feature of the corresponding command segment. If there are multiple sets of index paths, the index duration associated with each set of index paths is confirmed from the historical records, and the minimum value is selected from the associated index durations. The index path associated with the minimum value is recorded as the index feature of the corresponding instruction segment. The indexing time is the average time of multiple indexing processes; Step 4: Based on the characteristics and repetition frequency of the operation instruction segments, prioritize several operation instruction segments. After sorting, allocate resource proportions according to the index characteristics associated with each operation instruction segment. The specific method is as follows: Determine the feature to which the operation command segment belongs. If the operation command segment belongs to the personal feature, assign a weight Y and Y=3. If the operation command segment belongs to the scene feature, assign a weight Y and Y=1. Identify the number of repetitions of operation command segments in personal and scene features (CF) n Where n represents different operation instruction segments, using: CF n +Y=PX n Confirm sort value PX n And based on the different sorting values ​​PX associated with different operation instruction segments. n Sort the data from largest to smallest to confirm the sorting sequence of the operation instruction segments. Within the sorting sequence, the priority decreases from front to back. The specific method for allocating resource proportions for each operation instruction segment corresponding to the index feature is as follows: Confirm the index duration associated with the corresponding index feature and simultaneously assign a sorting weight to each index feature. The sorting weight is 1, 1.1, 1.2, ..., m, where m is a positive integer. When the operation instruction segment associated with the index feature is at the first position of the sorting sequence, its sorting weight is 1, and so on. The sorting weights associated with subsequent features gradually increase. The following formula is used: (index duration × C1) ÷ sort weight = proportion feature, to confirm the proportion feature associated with the corresponding index feature, where C1 is a preset fixed coefficient factor; Next, the ratios of the percentage features associated with several index features are processed to confirm the ratio columns of several index features and the specific ratio associated with the corresponding percentage feature. The computing resources corresponding to the ratio are extracted from the set computing resources as the percentage resources associated with the corresponding index features, and the computing resources associated with each subsequent index feature are allocated sequentially.

[0007] This invention provides an integrated learning method for engineering big data analysis based on graph neural networks. Compared with existing technologies, it has the following advantages: By quantifying operation instructions into associated points, tracking operation routes, and verifying overlapping features, the system can accurately extract the scene features and personal features of operators. This achieves a deep binding between "operation instructions and personnel habits" and enables graph neural networks to quickly identify the operation preferences of different personnel. By splitting operation instruction segments and filtering the optimal index path (based on minimizing historical indexing time), and combining the priority ranking of scenarios and individual characteristics, high-frequency and highly relevant operation instruction segments can obtain better index resource allocation. By prioritizing the sorting based on "weight + repetition count" and combining the indexing duration with the weighting ratio, dynamic and precise allocation of computing resources is achieved. High-priority (personal feature association, high-frequency repetition) operation instruction segments can receive more computing power support, avoiding resource waste while ensuring the stable and efficient execution of core operations. This makes the graph neural network more lightweight and targeted in engineering big data scenarios. Attached Figure Description

[0008] Figure 1 This is a schematic diagram of the method flow of the present invention. Detailed Implementation

[0009] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0010] Please see Figure 1 This application provides an integrated learning method for engineering big data analysis based on graph neural networks, including the following steps: Step 1: Confirm the operation instructions associated with the engineering modeling data in sequence, and set several associated points within the preset association framework so that the associated points correspond to the operation instructions, generating relevant standards for the corresponding engineering modeling data. Specifically, in order to make the learning efficiency of the corresponding neural network more efficient, the associated operation instructions are quantified into points, and the trajectory routes between the quantified points are confirmed based on the instruction sequence generated by the relevant operators in the operation process. The trajectory routes are then verified, compared, and analyzed to identify and confirm repeated routes, thereby confirming the corresponding operator's operating habits, and thus generating the operation profile associated with the corresponding operator. The specific method for generating the relevant standards is as follows: Identify the existing operation instructions from the engineering modeling data and record the total number G of operation instructions. Generate G associated points within the set association framework and associate each associated point with each operation instruction in sequence. The association framework is a preset framework, which can be understood as a square. Several points are set within the square, and each point corresponds to a set of operation instructions. The associated framework after binding the associated points with the operation instructions is recorded as the relevant standard for the corresponding engineering modeling data; Specifically, in order to identify the operational standards associated with each operator, the consistency between operational instructions is confirmed during the corresponding modeling process, and the route is confirmed within the corresponding association framework to facilitate subsequent quantitative overlap verification. Step 2: Confirm the operation records generated by a single operator within the traceability period, identify the operation instruction sequences associated with different operation scenarios from the operation records, confirm the operation routes corresponding to the operation instruction sequences within the relevant standards, confirm the scenario characteristics of the same operation scenario from the operation routes, and confirm the personal characteristics of the operator from different operation scenarios. The specific method for confirming the corresponding scene features within the same operation scenario is as follows: Using the current time as the base time, a set of traceability cycles is identified. The traceability cycle is a preset cycle, generally 720 hours. The operation records generated by the corresponding operators within the traceability cycle are extracted. Single records associated with the same operation scenario are extracted from the operation records, and the associated operation instructions are extracted from the single records. Several sets of operation instructions are sorted according to their time sequence. The operation instruction sequence associated with the corresponding single record is identified. The first operation instruction in the operation instruction sequence is identified at the associated point corresponding to the relevant standard. Using the associated point as the initial point, the associated points corresponding to other operation instructions are identified in turn. According to the sorting method of the operation instructions in the operation instruction sequence, the associated points corresponding to the subsequent operations are identified in turn. Starting from the initial point, the subsequently identified associated points are connected in turn to generate the operation route associated with the corresponding operation instruction sequence. The operation route associated with each single record is confirmed sequentially, and several operation routes are checked for overlap to identify overlapping segments and record the number of overlaps (CH). k Where k represents different overlapping road segments, and the total number of operation routes is recorded as ZS, satisfying: (CH k Sections with an overlap of ≥70% (÷ZS) are designated as habitual sections. The operation instructions associated with these habitual sections are sorted to confirm the habitual operation sequence. The confirmed sets of habitual operation sequences are then recorded as the scene characteristics of the corresponding operators in the corresponding operation scenarios. Specifically, individual operators exhibit different operational characteristics in different operational scenarios. Within the same operational scenario, there are also corresponding operational instructions. By analyzing the ordering process of these instructions, the corresponding instruction sequence can be identified. Furthermore, by analyzing the habitual routes associated with these instruction sequences and performing overlap verification, the scene characteristics associated with the corresponding habitual routes can be confirmed. This allows the neural network to respond quickly, optimize index paths, and improve the usability for relevant operators during subsequent operations. The specific methods for confirming personal characteristics from different operational scenarios are as follows: From multiple sets of habitual operation sequences associated with different operational scenarios, the habitual route segments associated with each habitual operation sequence are identified. The identified habitual route segments are then subjected to overlap verification to confirm overlapping habitual route segments. The number of overlaps generated by these overlapping habitual route segments is marked as CS. q Where q represents different habitually overlapping road segments, and the total number of habitual routes is recorded as XZ, satisfying (CS) q The overlapping road segments with a ratio of ÷XZ)≥30% are recorded as characteristic road segments. The operation instructions associated with the characteristic road segments are sorted to confirm the habitual characteristic sequence. The confirmed sets of habitual characteristic sequences are recorded as the personal characteristics associated with the corresponding operators. Specifically, in order to facilitate the learning process, scene features and personal features are identified from each operational feature. For each individual operator, the scene features and personal features are different. Therefore, in order to enable the neural network to respond quickly, the instruction index logic associated with each operator is planned and confirmed to improve the overall learning and utilization effect of the neural network.

[0011] Step 3: Identify several operation instruction segments from the scene features and personal features confirmed by the operator, and identify the path location associated with each operation instruction segment. Based on the path features of adjacent operation instructions, identify and record the index features of the corresponding adjacent operation instruction segments. The specific method for confirming the operation command segment is as follows: From the personal characteristics of the corresponding operator, identify the sequence of operation instructions associated with each personal characteristic, and identify the operation instruction segments from front to back in the sequence of operation instructions. Each operation instruction segment contains only two operation instructions, and the identified operation instruction segments are recorded as personal instruction segments. Using the same confirmation method as individual command segments, the operation command segments associated with the scene features are confirmed and recorded as scene command segments; For individual instruction segments and scene instruction segments, the path positions of the preceding and following operation instructions are identified, and the index paths associated with the two sets of path positions are identified (that is, one path position reaches another set of path positions through the index path, and there are corresponding and numerous associated paths in the corresponding neural network). If there is only one set of index paths, it is recorded as the index feature of the corresponding instruction segment. If there are multiple sets of index paths, the index duration associated with each set of index paths is identified from the historical records (its index duration is the average time of multiple indexing processes), and the minimum value is selected from the associated index durations. The index path associated with the minimum value is recorded as the index feature of the corresponding instruction segment. Its index feature is the optimal index path segment, which has a faster indexing rate, shorter duration, higher efficiency, and can achieve the best indexing processing effect. Step 4: Based on the characteristics and repetition frequency of the operation instruction segments, prioritize several operation instruction segments. After sorting, allocate resource proportions to the index characteristics associated with each operation instruction segment to facilitate convenient and fast execution of the corresponding operation instruction segments in subsequent indexing processes. The specific method for prioritizing several operation instruction segments is as follows: Determine the feature to which the operation command segment belongs. If the operation command segment belongs to the personal feature, assign a weight Y and Y=3. If the operation command segment belongs to the scene feature, assign a weight Y and Y=1. Identify the number of repetitions of operation command segments in personal and scene features (CF) nWhere n represents different operation instruction segments, using: CF n +Y=PX n Confirm sort value PX n And based on the different sorting values ​​PX associated with different operation instruction segments. n Sort the data from largest to smallest to confirm the sorting sequence of the operation instruction segments. Within the sorting sequence, the priority decreases from front to back. The specific method for allocating resource proportions for each operation instruction segment corresponding to the index feature is as follows: Confirm the index duration associated with the corresponding index feature and simultaneously assign a sorting weight to each index feature. The sorting weight is 1, 1.1, 1.2, ..., m, where m is a positive integer. When the operation instruction segment associated with the index feature is at the first position of the sorting sequence, its sorting weight is 1, and so on. The sorting weights associated with subsequent features gradually increase. The formula is: (index duration × C1) ÷ sorting weight = proportion feature. This determines the proportion feature associated with the corresponding index feature. C1 is a preset fixed coefficient factor. Its specific value is determined by the operator based on experience. Its value is determined based on different neural network models and the discrete range between values. If the value range is large, the value is small; if the value range is small, the value is large. Next, the ratio of the percentage features associated with several index features is processed to confirm the ratio column of several index features and the specific ratio associated with the corresponding percentage feature. The computing resources corresponding to the ratio are extracted from the set computing resources as the percentage resources associated with the corresponding index features, and the computing resources associated with each subsequent index feature are allocated sequentially. Specifically, there are three sets of index features from beginning to end, and the associated sorting weights are 1, 1.1, and 1.2. Each index feature is associated with a corresponding index duration. Based on the confirmed sorting weights, the proportion of each index feature can be determined. By processing the ratios of the three proportion features, a set of ratio columns can be determined. The computing resources are evenly distributed based on this ratio column, and the computing resources associated with each ratio are confirmed and allocated again to ensure that each index feature can be effectively allocated the corresponding computing resources.

[0012] Some of the data in the above formulas are numerical calculations with dimensions removed, and the contents not described in detail in this specification are all prior art known to those skilled in the art.

[0013] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A method for integrated learning and analysis of engineering big data based on graph neural networks, characterized in that: Includes the following steps: Step 1: Confirm the operation commands associated with the engineering modeling data in sequence, and set several associated points within the preset associated framework so that the associated points correspond to the operation commands, thereby generating relevant standards belonging to the corresponding engineering modeling data. Step 2: Confirm the operation records generated by a single operator within the traceability period, identify the sequence of operation instructions associated with different operation scenarios from the operation records, confirm the corresponding operation route within the relevant standards, confirm the scenario characteristics of the same operation scenario from the operation route, and then confirm the personal characteristics of the operator from different operation scenarios. Step 3: Identify several operation instruction segments from the scene features and personal features confirmed by the operator, and identify the path location associated with each operation instruction segment. Based on the path features of adjacent operation instructions, identify and record the index features of the corresponding adjacent operation instruction segments. Step 4: Based on the characteristics and repetition frequency of the operation instruction segments, prioritize several operation instruction segments, and after sorting, allocate resource proportions to the index characteristics associated with each operation instruction segment.

2. The engineering big data analysis and integrated learning method based on graph neural networks according to claim 1, characterized in that, In step one, the specific method for generating the relevant standards is as follows: Identify the existing operation commands from the engineering modeling data and record the total number G of operation commands. Generate G associated points within the set association framework and associate each associated point with each operation command in sequence. The association framework is a preset framework. The associated framework, after binding the associated points with the operation instructions, is recorded as the relevant standard for the corresponding engineering modeling data.

3. The engineering big data analysis and integrated learning method based on graph neural networks according to claim 1, characterized in that, In step two, the specific method for confirming the scene characteristics of the same operation scenario is as follows: Using the current time as the base time, a set of traceability cycles is identified, which are preset cycles. The operation records generated by the corresponding operators within the traceability cycle are extracted. Single records associated with the same operation scenario are extracted from the operation records, and the associated operation instructions are extracted from the single records. Several sets of operation instructions are sorted according to their time sequence. The operation instruction sequence associated with the corresponding single record is identified. The first operation instruction in the operation instruction sequence is identified at the associated point corresponding to the relevant standard. The associated point is used as the initial point. The associated points corresponding to other operation instructions are identified in turn. According to the sorting method of the operation instructions in the operation instruction sequence, the associated points corresponding to the subsequent operations are identified in turn. Starting from the initial point, the subsequently identified associated points are connected in turn to generate the operation route associated with the corresponding operation instruction sequence. The operation route associated with each single record is confirmed sequentially, and several operation routes are checked for overlap to identify overlapping segments and record the number of overlaps (CH). k Where k represents different overlapping road segments, and the total number of operation routes is recorded as ZS, satisfying: (CH k Road segments with an overlap of ≥70% (÷ZS) are designated as habitual road segments. The operation instructions associated with these habitual road segments are sorted to confirm the habitual operation sequence. The confirmed sets of habitual operation sequences are then recorded as the scene characteristics of the corresponding operators in the corresponding operation scenarios.

4. The engineering big data analysis and integrated learning method based on graph neural networks according to claim 3, characterized in that, In step two, the specific methods for confirming personal characteristics from different operational scenarios are as follows: From multiple sets of habitual operation sequences associated with different operational scenarios, the habitual route segments associated with each habitual operation sequence are identified. The identified habitual route segments are then subjected to overlap verification to confirm overlapping habitual route segments. The number of overlaps generated by these overlapping habitual route segments is marked as CS. q Where q represents different habitually overlapping road segments, and the total number of habitual routes is recorded as XZ, satisfying (CS) q Road segments with an overlap of ≥30% (÷XZ) are designated as characteristic road segments. The operation instructions associated with the characteristic road segments are sorted to confirm the habitual characteristic sequence. The confirmed sets of habitual characteristic sequences are then recorded as the personal characteristics associated with the corresponding operators.

5. The engineering big data analysis and integrated learning method based on graph neural networks according to claim 1, characterized in that, In step three, the specific method for confirming the operation instruction segment is as follows: From the personal characteristics of the corresponding operator, identify the sequence of operation instructions associated with each personal characteristic, and identify the operation instruction segments from front to back in the sequence of operation instructions. Each operation instruction segment contains only two operation instructions, and the identified operation instruction segments are recorded as personal instruction segments. Using the same confirmation method as individual command segments, the operation command segments associated with the scene features are confirmed and recorded as scene command segments; For individual command segments and scene command segments, determine the path positions of the preceding and following operation commands, and determine the index path associated between the two sets of path positions. If there is only one set of index paths, record it as the index feature of the corresponding command segment.

6. The engineering big data analysis and integrated learning method based on graph neural networks according to claim 5, characterized in that, If there are multiple sets of index paths, the index duration associated with each set of index paths is confirmed from the historical records, and the minimum value is selected from the associated index durations. The index path associated with the minimum value is recorded as the index feature of the corresponding instruction segment. The indexing time is the average time of multiple indexing processes.

7. The engineering big data analysis and integrated learning method based on graph neural networks according to claim 1, characterized in that, In step four, the specific method for prioritizing several operation instruction segments is as follows: Determine the feature to which the operation command segment belongs. If the operation command segment belongs to the personal feature, assign a weight Y and Y=3. If the operation command segment belongs to the scene feature, assign a weight Y and Y=1. Identify the number of repetitions of operation command segments in personal and scene features (CF) n Where n represents different operation instruction segments, using: CF n +Y=PX n Confirm sort value PX n And based on the different sorting values ​​PX associated with different operation instruction segments. n The values ​​are sorted from largest to smallest to confirm the sorting sequence of the operation instruction segment. Within the sorting sequence, the priority decreases from front to back.

8. The engineering big data analysis and integrated learning method based on graph neural networks according to claim 7, characterized in that, In step four, the specific method for allocating resource proportions for the index features corresponding to each operation instruction segment is as follows: Confirm the index duration associated with the corresponding index feature and simultaneously assign a sorting weight to each index feature. The sorting weight is 1, 1.1, 1.2, ..., m, where m is a positive integer. When the operation instruction segment associated with the index feature is at the first position of the sorting sequence, its sorting weight is 1, and so on. The sorting weights associated with subsequent features gradually increase. The following formula is used: (index duration × C1) ÷ sort weight = proportion feature, to confirm the proportion feature associated with the corresponding index feature, where C1 is a preset fixed coefficient factor; Next, the ratios of the percentage features associated with several index features are processed to confirm the ratio columns of several index features and the specific ratio associated with the corresponding percentage feature. The computing resources corresponding to the ratio are extracted from the set computing resources as the percentage resources associated with the corresponding index features, and the computing resources associated with each subsequent index feature are allocated sequentially.

Images