An information system engineering supervision resource configuration optimization management method and system

Abstract

The application discloses an information system engineering supervision resource configuration optimization management method and system, relates to the technical field of information system engineering supervision, and comprises the following steps: collecting supervision resource basic information, constructing a multi-technical dimension capability index system, and forming a resource capability index framework; based on the resource capability index framework, extracting the capability characteristics of each supervisor, performing capability grade division, and obtaining a hierarchical capability vector; dynamically correcting the hierarchical capability vector according to real-time work performance to obtain an updated capability vector; based on the updated capability vector, generating a supervision resource multi-dimensional capability portrait model, outputting a fine-grained capability portrait, and taking the fine-grained capability portrait as the basic input of resource matching and distribution. The application effectively solves the problems of lacking standardized and quantitative capability evaluation, lacking a dynamic adjustment mechanism and lacking a scientific resource matching and optimization algorithm in the prior art.

Description

Technical Field

[0001] This invention relates to the field of information system engineering supervision technology, and more specifically, to a method and system for optimizing the allocation of information system engineering supervision resources. Background Technology

[0002] As the complexity of information systems engineering continues to increase, the role of project supervision is becoming increasingly important. Supervisors are not only responsible for overseeing the project execution process to ensure timely, high-quality, and budgetary completion, but also for providing early warnings and managing potential problems. However, how to efficiently and accurately allocate supervision resources and ensure that personnel capabilities match project needs remains a challenge in information systems engineering management. Traditional methods of allocating supervision resources largely rely on human experience and intuition, lacking scientific quantitative indicators and dynamic adjustment mechanisms, resulting in low accuracy and efficiency in resource allocation. Current methods of supervising resource management typically suffer from the following problems:

[0003] 1. Lack of a standardized and quantifiable competency assessment system. Currently, many information system engineering supervision resource allocation methods rely on the limited background information and experience of supervisors, making it difficult to comprehensively and objectively assess their technical capabilities and actual work performance. Existing methods largely depend on manual assessment, often failing to provide accurate and systematic competency evaluation standards, especially when considering technical skills and overall qualities, which often rely on subjective judgment. This subjectivity leads to inefficient and inaccurate resource allocation, hindering the achievement of refined personnel competency distribution.

[0004] 2. Lack of a dynamic adjustment mechanism. Traditional supervision resource allocation is often based on static information, failing to reflect changes in the capabilities of supervision personnel during the project process. For example, the capabilities of supervision personnel may change during project progress (such as increased technical proficiency or fluctuations in work performance), but traditional methods fail to update and adjust personnel capability assessments in real time, making it impossible to optimize the allocation of supervision resources in a timely manner as project needs change. Such resource allocation methods easily lead to a mismatch between personnel capabilities and task requirements in the project, thereby affecting project quality and schedule.

[0005] 3. Lack of scientific resource matching and optimization algorithms. Traditional methods of allocating supervision resources lack data-driven intelligent optimization mechanisms, often relying on manual allocation or rules of thumb for personnel scheduling. This not only consumes a lot of time and energy but also makes it difficult to achieve optimal allocation. Existing methods struggle to accurately match personnel to specific project needs (such as project type, task requirements, etc.), cannot dynamically adjust the allocation strategy for supervision personnel, and cannot optimize and adjust the tasks of supervision personnel in a timely manner based on project progress and real-time performance. Summary of the Invention

[0006] To overcome the aforementioned deficiencies in the prior art, embodiments of the present invention provide an information system engineering supervision resource allocation optimization management method and system. Through a data-driven capability assessment system, a dynamic capability correction mechanism, and a reinforcement learning optimization algorithm, it effectively solves the problems of lack of standardized and quantitative capability assessment, lack of dynamic adjustment mechanism, and lack of scientific resource matching and optimization algorithm in the prior art.

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] Firstly, this application provides a method for optimizing the allocation and management of information system engineering supervision resources. This method includes: collecting basic information on supervision resources and constructing a multi-technical-dimensional capability index system to form a resource capability index framework; based on the resource capability index framework, extracting capability characteristics for each supervisor and classifying their capability levels to obtain a hierarchical capability vector; dynamically correcting the hierarchical capability vector based on real-time work performance to obtain an updated capability vector; and generating a multi-dimensional capability profile model of supervision resources based on the updated capability vector, outputting a fine-grained capability profile, and using the fine-grained capability profile as the basic input for resource matching and allocation.

[0009] In one embodiment, basic information on supervision resources is collected, and a multi-technical-dimensional capability indicator system is constructed to form a resource capability indicator framework. This includes: defining a set of basic information fields for supervision resources, and collecting and standardizing structured information from each supervisor to obtain an original resource dataset; constructing a set of technical dimensions from the original resource dataset using a data-driven approach; defining basic capability indicators for each dimension based on the set of technical dimensions to form an initial capability indicator set; calculating a comprehensive capability score for each dimension based on the initial capability indicator set to form a capability quantification rule set; and establishing a multi-dimensional capability indicator matrix for each supervisor based on the original resource dataset and the capability quantification rule set, and summarizing these to obtain the resource capability indicator framework.

[0010] In one embodiment, a set of technical dimensions is constructed from the original resource dataset using a data-driven approach. This includes: extracting supervision texts and preprocessing them to form a standardized text set; using the standardized text set as input, establishing an LDA topic model to mine several topics and obtain an initial set; based on the initial set, constructing a personnel-topic matrix and performing normalization to obtain a standardized matrix; using the standardized matrix as input, performing principal component analysis to obtain several principal components, and selecting the top few principal components that meet preset conditions; projecting the features of each supervisor on the standardized matrix onto the top few principal component spaces to obtain a personnel-principal component matrix; using the personnel-principal component matrix as input, performing a K-means clustering algorithm for verification. If the distance between different clusters in the principal component space is greater than the distance between samples within the cluster, it has good discriminative ability; otherwise, adjustments are made until a stable clustering result is obtained; the dimensions verified by K-means are determined as the final set of technical dimensions.

[0011] In one embodiment, based on the resource capability index framework, capability features are extracted for each supervisor and capability levels are divided to obtain a hierarchical capability vector. This includes: extracting features based on the resource capability index framework to obtain an initial capability vector; and introducing a dynamic threshold setting method based on quantiles based on the initial capability vector to automatically classify the capability score of each supervisor.

[0012] In one embodiment, a dynamic threshold setting method based on quantiles is used to automatically classify the competency scores of each supervisor, including: sorting the initial competency vectors of each supervisor to obtain a competency score list; calculating 25%, 50%, and 75% quantile thresholds based on the competency score list, and mapping the competency scores to different competency levels through the quantile thresholds; and obtaining a graded competency vector by classifying all technical dimensions.

[0013] In one embodiment, the hierarchical capability vector is dynamically corrected based on real-time work performance to obtain an updated capability vector. This includes: collecting real-time work performance data and performing standardization processing; constructing a work performance-capability impact model based on the real-time work performance data and calculating the correction amount for each technical dimension; using a smoothing update function to correct the current hierarchical capability vector based on the correction amount to obtain an updated hierarchical capability vector; and normalizing the updated hierarchical capability vector to obtain an updated capability vector.

[0014] In one embodiment, the job performance-capability impact model further includes dynamic weight adjustment based on complex network analysis, including: constructing a capability dimension network, simulating the information transmission process in the capability dimension network through a graph propagation algorithm, and calculating the weight of each node; adjusting the node weights in the capability dimension network according to job performance data to obtain the impact weights.

[0015] In one embodiment, a multi-dimensional capability profile model for supervisory resources is generated based on the updated capability vector, outputting a fine-grained capability profile. This fine-grained capability profile is then used as the basic input for resource matching and allocation. The process includes: decomposing the updated capability vector into multiple fine-grained capability dimensions using an adaptive dimensionality partitioning and optimization method based on reinforcement learning; combining the updated capability vector with its decomposed fine-grained capability dimensions to construct a multi-dimensional capability profile model; generating a fine-grained capability profile for each supervisor using the multi-dimensional capability profile model; and using the fine-grained capability profile as the basic input for resource matching and allocation to guide the role allocation, resource configuration, and task arrangement of supervisors in different projects.

[0016] In one embodiment, based on the updated capability vector, it is decomposed into multiple fine-grained capability dimensions using an adaptive dimension partitioning and optimization method based on reinforcement learning. This includes: constructing a reinforcement learning model, where the state space represents the current updated capability vector of the supervisor in each technical dimension, the action space represents the strategy for partitioning the supervisor's capability dimensions, and the reward function includes the supervisor's project success rate, task quality, and customer satisfaction; initializing the capability dimension partitioning strategy; learning the updated capability vector of each supervisor through the reinforcement learning model and adjusting the partitioning strategy for each dimension; decomposing the supervisor's updated capability vector into multiple fine-grained capability dimensions based on the optimized capability dimension partitioning strategy; dynamically adjusting the capability dimension partitioning according to changes in project progress and supervisor performance; and outputting the fine-grained capability dimensions of the supervisor based on the final optimized dimension partitioning strategy.

[0017] Secondly, this application provides an information system engineering supervision resource allocation optimization management system, which includes: a framework construction module for collecting basic information on supervision resources and constructing a multi-technical-dimensional capability index system to form a resource capability index framework; a hierarchical module for extracting capability features from each supervisor based on the resource capability index framework and classifying capability levels to obtain hierarchical capability vectors; a correction module for dynamically correcting the hierarchical capability vectors based on real-time work performance to obtain updated capability vectors; and a profile generation module for generating a multi-dimensional capability profile model of supervision resources based on the updated capability vectors, outputting fine-grained capability profiles, and using the fine-grained capability profiles as the basic input for resource matching and allocation.

[0018] As can be seen from the above technical solutions, the embodiments of this application have the following advantages:

[0019] A comprehensive and quantitative assessment system for supervisory resource capabilities was constructed using a data-driven approach. By collecting and standardizing multi-dimensional information on supervisory personnel, technical dimensions can be scientifically and objectively categorized. Based on indicators such as project experience, technical depth, quality performance, evaluation feedback, and text matching, personnel capabilities are precisely quantified. Through techniques such as LDA topic modeling, PCA principal component analysis, and K-means clustering, the technical dimensions are automatically and dynamically adjusted, avoiding human subjectivity. Automatic quantile-based grading and dynamic performance-based correction methods improve the accuracy and real-time nature of capability assessment. Furthermore, by optimizing the capability dimension categorization through reinforcement learning, precise resource matching and task allocation can be achieved according to project needs, effectively improving the efficiency of supervisory resource allocation and the accuracy of decision-making, ensuring smooth project execution and optimal resource utilization. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a schematic diagram of a method for optimizing the allocation of information system engineering supervision resources, provided as an embodiment of this application.

[0022] Figure 2 This is a schematic diagram of the structure of an information system engineering supervision resource allocation optimization management system provided in an embodiment of this application. Detailed Implementation

[0023] To enable those skilled in the art to better understand the technical solutions in this application, the technical solutions in the embodiments of this application will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0024] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, "a plurality of" or "several" means two or more, unless otherwise explicitly specified.

[0025] It should also be noted that, in this document, terms such as “comprising,” “including,” or any other variations thereof are intended to cover non-exclusive inclusion, such that an article or device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such an article or device. Without further limitation, an element defined by the phrase “comprising one…” does not exclude the presence of other identical elements in the article or device that includes the aforementioned element.

[0026] Reference Figure 1 As shown in the diagram, the present invention provides a flowchart of an information system engineering supervision resource allocation optimization management method, which includes the following steps:

[0027] S1. Collect basic information on supervision resources and construct a multi-technical-dimensional capability indicator system to form a resource capability indicator framework, including:

[0028] A predefined set of basic information fields for supervision resources is provided, which includes fields for personnel identity information, job title, professional background, years of service, list of projects participated in, supervision roles undertaken, historical work results, historical performance evaluation, training and certification records, and historical supervision documents (weekly reports, monthly reports, review opinions, acceptance reports, and risk records).

[0029] Based on the basic information field set of supervision resources, structured information was collected from each supervisor, and the data was standardized to obtain the original resource dataset.

[0030] The standardization process includes unifying job title descriptions, professional title level codes, project type classifications, supervisor role names, and text formats, and performing word segmentation, noise reduction, and terminology normalization.

[0031] Based on the original resource dataset, a set of technical dimensions is constructed using data-driven methods. The set of technical dimensions includes network architecture review capability, software quality assessment capability, information security audit capability, data governance and compliance review capability, and project progress control capability.

[0032] Based on the set of technical dimensions, basic capability indicators are defined for each dimension to form an initial set of capability indicators. The basic capability indicators include project experience indicators, technical depth indicators, quality performance indicators, evaluation feedback indicators, and text matching indicators.

[0033] The specific calculation formula for the project experience indicators is as follows:

[0034]

[0035] In the formula, For the first Project experience indicators for individual supervisors under the current technical dimensions For the first The number of times each supervisor participates in relevant projects in the current dimension. This represents the maximum number of all supervisors who can participate in the project within the current dimension.

[0036] The specific calculation formula for the aforementioned technical depth indicator is as follows:

[0037]

[0038] In the formula, As a technical depth indicator, This represents the total number of critical tasks in the current dimension. The score is given to the personnel for their performance in the k-th key technical task.

[0039] The specific calculation formula for the aforementioned quality performance indicators is as follows:

[0040]

[0041] In the formula, As a quality performance indicator. This refers to the number of defects in the individual's historical projects. This represents the maximum number of defects for all personnel.

[0042] The specific calculation formula for the evaluation feedback indicators is as follows:

[0043]

[0044] In the formula, To evaluate feedback indicators, Average customer rating This is the average value from peer review.

[0045] The specific calculation formula for the text matching index is as follows:

[0046]

[0047] In the formula, For the first Text matching metrics between individual supervisors and the technical topics of that dimension. For personnel The text vector, This is the standard text vector for this technical dimension.

[0048] Calculate the comprehensive ability score for each dimension based on the initial set of ability indicators to form a set of ability quantification rules;

[0049] The specific formula for calculating the comprehensive ability score is as follows:

[0050]

[0051] In the formula, To calculate the overall ability score, , , , , These are the weighting coefficients.

[0052] Based on the original resource dataset and the set of capability quantification rules, a multidimensional capability index matrix is ​​established for each supervisor. The multidimensional capability index matrices of all personnel are summarized to obtain a resource capability index framework. The resource capability index framework includes a set of technical dimensions, an initial capability index set for each dimension, a set of capability quantification rules, and a standardized personnel capability matrix. In the multidimensional capability index matrix, rows represent personnel, columns represent technical dimensions, and elements are comprehensive capability scores.

[0053] Furthermore, based on the original resource dataset, a set of technical dimensions is constructed using data-driven methods, including:

[0054] All supervision texts are extracted from the original resource dataset and preprocessed to form a standardized text set. The preprocessing includes removing irrelevant stop words, standardizing the expression of professional terms, and performing word segmentation.

[0055] Using the standardized text set as input, an LDA topic model is established to automatically mine several topics, with each topic serving as a candidate technology dimension to obtain an initial set;

[0056] The LDA (Latent Dirichlet Allocation) topic model is a probabilistic generation model used to automatically identify potential topic structures from the standardized project text set. Specifically, the model treats each text as a mixture of several potential topics, with each topic represented by a set of technically relevant words and their probability distribution. Through model training, it can output the distribution of each topic in the text and the degree of association between each supervisor's text and each topic. In this invention, each automatically generated topic is interpreted as a candidate technical dimension, thus forming an initial set of technical dimensions. This makes the division of technical dimensions data-driven, quantifiable, and reproducible, rather than relying on subjective human settings.

[0057] Based on the initial set, obtain the relevance score of each person on each topic (e.g., based on the frequency of relevant keywords in the projects they participated in), and construct a person-topic matrix, where rows represent supervisors, columns represent topic dimensions, and each matrix element represents the person's relevance score on the topic;

[0058] The personnel-topic matrix is ​​normalized to obtain a standardized matrix;

[0059] Using the standardized matrix as input, principal component analysis is performed to obtain several principal components, and the cumulative explained variance is calculated. The top principal components with a cumulative explained variance ≥ 80% are selected to form a validated set of technical dimensions. The cumulative explained variance refers to the sum of the variance contribution rates of each principal component after performing principal component analysis (PCA) on the standardized matrix, in descending order. It is used to measure the proportion of the total variance of the original data that the selected principal components can explain.

[0060] Project the features of each supervisor on the standardized matrix onto the first few principal component spaces to obtain the personnel-principal component matrix;

[0061] Using the personnel-principal component matrix as input, the K-means clustering algorithm is executed to divide all supervisors into K clusters. If the distance between different clusters in the principal component space is greater than the distance between samples within the cluster, showing obvious separation, it indicates that the set of technical dimensions has good discriminative ability. Conversely, if the distance between clusters is significantly less than the distance between samples within the cluster, the number of LDA topics or the number of PCA retained dimensions are adjusted backtrackingly until a stable clustering result is obtained.

[0062] The dimensions discovered by LDA, verified by PCA, and validated by K-means were determined as the final set of technical dimensions.

[0063] It should be noted that by using a data-driven approach to automatically construct the technical dimensions of supervision resources, and combining them with multi-dimensional quantitative indicators such as project experience, technical depth, quality performance, evaluation feedback, and text matching, a standardized and quantifiable multi-dimensional capability matrix and resource capability indicator framework are formed. This enables the scientific and objective division of technical dimensions and the precise quantitative assessment of capabilities, thereby significantly improving the efficiency and accuracy of supervision resource allocation, capability matching, and management decisions.

[0064] S2, based on the resource capability index framework, extracts capability characteristics for each supervisor and classifies them into capability levels, resulting in a hierarchical capability vector, including:

[0065] Based on the resource capability index framework, feature extraction is performed on the technical capabilities of each supervisor to obtain an initial capability vector.

[0066] The initial capability vector consists of multiple basic capability indicators defined in the framework. Each indicator is calculated based on historical project experience, technical task depth, quality performance, evaluation feedback, and text matching quantification indicators through a set of capability quantification rules to form multidimensional capability characteristics of personnel.

[0067] Based on the initial capability vector, a dynamic threshold setting method based on quantiles is introduced to automatically classify the capability scores of each supervisor and map the capability scores of each dimension to the corresponding level range.

[0068] Furthermore, based on a dynamic threshold setting method using quantiles, the competency scores of each supervisor are automatically graded, including:

[0069] The initial capability vectors of each supervisor in each technical dimension are sorted to obtain a capability score list;

[0070] Calculate the 25th, 50th, and 75th percentile thresholds based on the ability score list. , as well as And these quantile thresholds are used to map ability scores to different ability levels;

[0071] Based on the set quantile threshold , as well as The initial capability vector of each supervisor is graded, if the initial capability vector This is the beginner level; if Initial capability vector This is intermediate level; if Initial capability vector , for advanced; initial capability vector > , is at the expert level;

[0072] By classifying all technical dimensions, a hierarchical capability vector is obtained.

[0073] It should be noted that by using automated threshold settings based on quantiles to map competency scores to different levels, the fairness and accuracy of the level classification are ensured. The use of quantiles allows for dynamic adjustment of the classification criteria, adapting to the competency distribution of different supervisors and avoiding the rigidity of static classification.

[0074] S3, dynamically adjusts the hierarchical capability vector based on real-time work performance to obtain an updated capability vector, including:

[0075] Collect real-time work performance data, including task difficulty, completion quality, customer feedback, and task time management.

[0076] Standardize the real-time work performance data to ensure that the data between different projects can be unified and compared. The standardization process includes mapping each indicator to the [0,1] interval, using a unified scoring scale for different project types, and eliminating outliers and filling in missing values.

[0077] Specifically, the standardized real-time work performance data is aligned with the current graded capability vector of the supervisor to clarify which work performance indicators affect which capability dimensions.

[0078] A performance-ability impact model is constructed based on real-time performance data to quantify the impact of task performance on the capabilities of each technical dimension and to calculate the correction amount for each technical dimension.

[0079] The specific calculation formula for the job performance-ability impact model is as follows:

[0080]

[0081] In the formula, This is the capability correction amount for the j-th technical dimension. As for the difficulty of the task, To achieve quality, Customer reviews For task time management, , , , These are the influence weights, satisfying... .

[0082] Based on the correction amount, the current hierarchical capability vector is corrected using a smooth update function to obtain the updated hierarchical capability vector.

[0083] The specific calculation formula for the smooth update function is as follows:

[0084]

[0085] In the formula, This is the updated hierarchical capability vector. This is a hierarchical capability vector. The coefficient is retained for historical data, with a value range of [0.6, 0.9].

[0086] The updated hierarchical capability vector is normalized to obtain the updated capability vector. .

[0087] The normalization process is performed, and the specific calculation formula is as follows:

[0088]

[0089] In the formula, This refers to the overall technical dimension.

[0090] Furthermore, the job performance-ability impact model also includes dynamic weight adjustment based on complex network analysis, including:

[0091] A capability dimension network is constructed, in which each node in the network represents a technology dimension, the edges between nodes represent the relationship between different dimensions, and the weight of the edge represents the strength of the mutual influence between dimensions. The influence relationship is initialized through expert knowledge or historical data. Based on the correlation between different technology dimensions, the initial weight value of each edge is dynamically determined using historical data and expert evaluation.

[0092] The process of information transmission in the capability dimension network is simulated by using graph propagation algorithms (such as PageRank algorithm). The weight of each node (technical dimension) is calculated. The weight of each node is updated according to the weight of its neighboring nodes and the weight of the edges connected to it. After each iteration, the weight of the node gradually converges until it reaches a preset stable value.

[0093] The graph propagation algorithm is the PageRank algorithm, and the node weight update rule is as follows:

[0094]

[0095] In the formula, Let i be the updated weight. Let be all the adjacent nodes connected to node i. Let be the influence coefficient between node i and node j. Let be the current weight of node j.

[0096] Based on work performance data, the node weights in the capability dimension network are adjusted to obtain the influence weights, ensuring that the node weights can reflect the current task environment and the actual capabilities of the supervisors.

[0097] First, performance data reflects the supervisors' performance across various technical dimensions. Based on this data, performance scores for each supervisor in each dimension can be obtained, and these scores are converted into weights using a weighted average method. Then, through a feedback mechanism, the edge weights between corresponding nodes (technical dimensions) in the network are adjusted to reflect the mutual influence between dimensions. For example, if performance in a certain dimension improves, it may increase the influence of that dimension on other dimensions, thus adjusting the edge weights between dimensions accordingly. Finally, the weights of all nodes are recalculated to optimize the relative influence between capability dimensions, ensuring that the node weights in the network match the latest capability status of the supervisors.

[0098] It should be noted that the adaptive updating of the supervisor's competency vector is achieved through the dynamic collection and standardized processing of real-time work performance data. This method, combining a work performance-competency impact model and complex network analysis, quantifies the impact of task performance on competency across various technical dimensions. Through smooth updates and normalization, it makes competency assessment more accurate and real-time. Furthermore, the graph propagation mechanism based on the PageRank algorithm ensures that the influence relationships between competency dimensions can be dynamically adjusted according to actual work performance, thereby optimizing resource allocation and task assignment decisions, improving the accuracy and reliability of supervisor competency assessment, and avoiding the limitations of fixed weights in traditional methods. This method not only improves the flexibility of the assessment but also ensures the continuous reflection of the supervisor's competency status, exhibiting high adaptability and operability.

[0099] S4. Based on the updated capability vector, a multi-dimensional capability profile model of supervision resources is generated, and a fine-grained capability profile is output. The fine-grained capability profile is used as the basic input for resource matching and allocation, and is used for precise skill matching and optimized configuration according to project requirements.

[0100] In this embodiment, a multi-dimensional capability profile model of supervision resources is generated based on the updated capability vector, outputting a fine-grained capability profile. This fine-grained capability profile is then used as the basic input for resource matching and allocation, including:

[0101] Based on the updated capability vector, the adaptive dimension partitioning and optimization method based on reinforcement learning is used to decompose it into multiple fine-grained capability dimensions.

[0102] The updated capability vector is combined with its decomposed fine-grained capability dimensions to construct a multi-dimensional capability profile model. The multi-dimensional capability profile model includes the capability performance of the supervisor in each technical dimension, which can accurately show the comprehensive capability status of the supervisor.

[0103] The multi-dimensional capability profile model generates a fine-grained capability profile for each supervisor, which accurately reflects the supervisor's capability status in various technical dimensions.

[0104] The fine-grained capability profile includes capability scores for each technical dimension, thus providing more specific and clear capability assessment results;

[0105] The fine-grained capability profile is used as the basic input for resource matching and allocation, to guide supervisors in role allocation, resource configuration and task arrangement in different projects;

[0106] Based on the detailed capability profile, the system can optimize task allocation and resource scheduling according to the specific capability status of the supervisors, ensuring that supervisory resources are used most rationally during project execution.

[0107] Furthermore, based on the updated capability vector, and using an adaptive dimension partitioning and optimization method based on reinforcement learning, it is decomposed into multiple fine-grained capability dimensions, including:

[0108] A reinforcement learning model is constructed, in which the state space represents the current updated capability vector of the supervisor in each technical dimension, the action space represents the strategy for dividing the supervisor's capability dimensions, and the reward function is defined by comprehensively considering the supervisor's project success rate, task quality, and customer satisfaction (i.e., weighted summation). The division of capability dimensions is optimized by maximizing these indicators.

[0109] At the outset, an initial capability dimension segmentation strategy is set based on heuristics using historical data or random initialization. For example, capabilities are initially divided into several broad dimensions (such as technical skills, project experience, and communication skills). This strategy serves as the initial state of the reinforcement learning model and is subsequently refined through model optimization.

[0110] The reinforcement learning model learns the updated capability vector for each supervisor. Based on the supervisor's performance in the project (e.g., task completion, customer feedback, project quality), the model adjusts the dimensional partitioning strategy. Q-learning or policy gradient methods of reinforcement learning are used to optimize the dimensional partitioning strategy after multiple iterations, ensuring that the dimensional partitioning gradually approaches the optimal level and adapts to the needs of different projects and tasks.

[0111] Based on the optimized capability dimension segmentation strategy, the updated capability vector of supervisors is further decomposed into multiple fine-grained capability dimensions. Each fine-grained dimension reflects the supervisor's performance in a specific skill or project capability (such as technical depth, project experience, and task execution quality). In this process, the reinforcement learning model dynamically creates and adjusts new capability dimensions according to the supervisor's work performance, project requirements, and task requirements, ensuring that the capability dimension segmentation closely matches the actual work environment.

[0112] As the project progresses and the performance of the supervisors changes, performance data of the supervisors in new tasks or projects is collected in real time, and the capability dimension classification is dynamically adjusted. Whenever the performance of the supervisors changes, the reinforcement learning model automatically adjusts the dimension classification strategy using real-time data to ensure that the capability vector and fine-grained dimensions are consistent with the latest state of the supervisors. This step is carried out through an online learning mechanism of reinforcement learning to ensure that the dimension classification can adapt to changes in the supervisors and the project environment;

[0113] Based on the final optimized dimensional segmentation strategy, fine-grained capability dimensions for supervisors are output. These dimensions not only include project experience and technical skills, but also automatically add or adjust fine-grained dimensions according to the specific needs of the project, such as schedule control capabilities, quality assessment capabilities, and problem-solving capabilities. The final generated fine-grained capability dimensions will serve as the basis for assessing the capabilities of supervisors and will support decisions such as project resource allocation and task assignment.

[0114] The reinforcement learning model is optimized using the Q-learning algorithm or deep Q-network (DQN). The dimension partitioning strategy is continuously adjusted through the state-action value function (Q-value), so that the capability dimension partitioning gradually tends to the optimal solution after each update.

[0115] It should be noted that by adaptively optimizing the division of the supervisor's competency dimensions through reinforcement learning models, fine-grained competency profiles are accurately generated, thereby achieving precise resource matching and task allocation based on the actual competency status of supervisors, which improves resource utilization efficiency and project execution effectiveness.

[0116] Reference Figure 2 As shown in the diagram, the present invention provides a structural schematic of an information system engineering supervision resource allocation optimization management system, including a framework construction module, a hierarchical module, a correction module, and a profile generation module, with connections between the modules:

[0117] The framework construction module is used to collect basic information on supervision resources and build a multi-technical-dimensional capability indicator system to form a resource capability indicator framework.

[0118] The hierarchical module is used to extract the capability characteristics of each supervisor based on the resource capability index framework, and to classify the capability levels to obtain a hierarchical capability vector.

[0119] The correction module is used to dynamically correct the hierarchical capability vector based on real-time work performance to obtain an updated capability vector;

[0120] The profile generation module is used to generate a multi-dimensional capability profile model of supervision resources based on the updated capability vector, output a fine-grained capability profile, and use the fine-grained capability profile as the basic input for resource matching and allocation.

[0121] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0122] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0123] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0124] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0125] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for optimizing the allocation and management of information system engineering supervision resources, characterized in that, include: Collect basic information on supervision resources and construct a multi-technical-dimensional capability indicator system to form a resource capability indicator framework; Based on the resource capability index framework, capability characteristics are extracted for each supervisor, and capability levels are divided to obtain a hierarchical capability vector. The hierarchical capability vector is dynamically adjusted based on real-time work performance to obtain an updated capability vector. Based on the updated capability vector, a multi-dimensional capability profile model of supervision resources is generated, outputting a fine-grained capability profile, which is then used as the basic input for resource matching and allocation.

2. The method for optimizing the allocation and management of information system engineering supervision resources according to claim 1, characterized in that, The aforementioned collection of basic information on supervision resources and the construction of a multi-technical-dimensional capability indicator system form a resource capability indicator framework, including: Define the basic information field set of supervision resources, and collect and standardize the structured information of each supervision personnel to obtain the original resource dataset; A set of technical dimensions is constructed from the original resource dataset using a data-driven approach; Based on the set of technical dimensions, basic capability indicators are defined for each dimension to form an initial set of capability indicators; Based on the initial set of capability indicators, a comprehensive capability score is calculated for each dimension to form a set of capability quantification rules. Based on the original resource dataset and the set of capability quantification rules, a multi-dimensional capability index matrix is ​​established for each supervisor, and the resource capability index framework is obtained by summarizing them.

3. The method for optimizing the allocation and management of information system engineering supervision resources according to claim 2, characterized in that, The process of constructing a set of technical dimensions from the original resource dataset using a data-driven approach includes: Extract the supervision text and preprocess it to form a standardized text set; Using a standardized text set as input, an LDA topic model is built to mine several topics and obtain an initial set. Based on the initial set, a personnel-topic matrix is ​​constructed and normalized to obtain a standardized matrix; Using the standardized matrix as input, principal component analysis is performed to obtain several principal components, and the top few principal components that meet the preset conditions are selected. Project the features of each supervisor on the standardized matrix onto the first few principal component spaces to obtain the personnel-principal component matrix; Using the personnel-principal component matrix as input, the K-means clustering algorithm is executed for verification. If the distance between different clusters in the principal component space is greater than the distance between samples within the cluster, it has good discriminative ability; otherwise, adjustments are made until a stable clustering result is obtained. The dimensions verified by K-means are determined as the final set of technical dimensions.

4. The method for optimizing the allocation and management of information system engineering supervision resources according to claim 1, characterized in that, The resource capability index framework is used to extract capability characteristics for each supervisor and classify their capability levels to obtain a hierarchical capability vector, including: Based on the resource capability index framework, feature extraction is performed to obtain an initial capability vector; Based on the initial capability vector, a dynamic threshold setting method based on quantiles is introduced to automatically classify the capability scores of each supervisor.

5. The method for optimizing the allocation and management of information system engineering supervision resources according to claim 4, characterized in that, The quantile-based dynamic threshold setting method automatically grades the competence score of each supervisor, including: The initial capability vectors of each supervisor are sorted to obtain a capability score list; Based on the list of ability scores, calculate the 25th, 50th, and 75th percentile thresholds, and map the ability scores to different ability levels using the quantile thresholds. By classifying all technical dimensions, a hierarchical capability vector is obtained.

6. The method for optimizing the allocation and management of information system engineering supervision resources according to claim 1, characterized in that, The step of dynamically correcting the hierarchical capability vector based on real-time work performance to obtain an updated capability vector includes: Collect real-time performance data and perform standardized processing; A performance-ability impact model is constructed based on real-time performance data, and the correction amount for each technical dimension is calculated. Based on the correction amount, the current hierarchical capability vector is corrected using a smooth update function to obtain the updated hierarchical capability vector. The updated hierarchical capability vector is normalized to obtain the updated capability vector.

7. The method for optimizing the allocation and management of information system engineering supervision resources according to claim 6, characterized in that, The job performance-ability impact model also includes dynamic weight adjustment based on complex network analysis, including: Construct a capability dimension network, simulate the information transmission process in the capability dimension network using a graph propagation algorithm, and calculate the weight of each node; Based on work performance data, the node weights in the capability dimension network are adjusted to obtain the influence weights.

8. The method for optimizing the allocation and management of information system engineering supervision resources according to claim 1, characterized in that, The process involves generating a multi-dimensional capability profile model for supervision resources based on updated capability vectors, outputting a fine-grained capability profile, and using this fine-grained capability profile as the basic input for resource matching and allocation. This includes: Based on the updated capability vector, it is decomposed into multiple fine-grained capability dimensions using an adaptive dimension partitioning and optimization method of reinforcement learning; The updated capability vector is combined with its decomposed fine-grained capability dimensions to construct a multi-dimensional capability profile model; A fine-grained capability profile is generated for each supervisor through a multi-dimensional capability profile model. The fine-grained capability profile is used as the basic input for resource matching and allocation, guiding supervisors in role allocation, resource configuration, and task arrangement in different projects.

9. The method for optimizing the allocation and management of information system engineering supervision resources according to claim 8, characterized in that, The method described above, based on updating the capability vector, decomposes it into multiple fine-grained capability dimensions using an adaptive dimension partitioning and optimization method derived from reinforcement learning, including: Construct a reinforcement learning model, where the state space represents the current updated capability vector of the supervisor in each technical dimension, the action space represents the strategy for dividing the supervisor's capability dimensions, and the reward function includes the supervisor's project success rate, task quality, and customer satisfaction. Initialization capability dimension segmentation strategy; The updated capability vector of each supervisor is learned through a reinforcement learning model, and the partitioning strategy of each dimension is adjusted accordingly. Based on the optimized capability dimension segmentation strategy, the updated capability vector of supervisors is decomposed into multiple fine-grained capability dimensions; The capability dimensions will be dynamically adjusted based on changes in project progress and the performance of the supervisors. Based on the final optimized dimension partitioning strategy, the fine-grained capability dimensions of the supervisors are output.

10. A system using the information system engineering supervision resource allocation optimization management method as described in any one of claims 1-9, characterized in that, include: The framework construction module is used to collect basic information on supervision resources and build a multi-technical-dimensional capability indicator system to form a resource capability indicator framework. The hierarchical module is used to extract the capability characteristics of each supervisor based on the resource capability index framework, and to classify the capability levels to obtain a hierarchical capability vector. The correction module is used to dynamically correct the hierarchical capability vector based on real-time work performance to obtain an updated capability vector; The profile generation module is used to generate a multi-dimensional capability profile model of supervision resources based on the updated capability vector, output a fine-grained capability profile, and use the fine-grained capability profile as the basic input for resource matching and allocation.

Images