Heavy equipment collaborative design task and personnel matching method considering proficiency

Abstract

The invention belongs to the technical field of complex heavy equipment collaborative design, and particularly relates to a heavy equipment collaborative design task and personnel matching method considering proficiency, which comprises the following steps: S1, constructing a task information table and a personnel information table; s2, calculating the similarity among the tasks, and constructing a task similarity matrix; s3, constructing a proficiency prediction model; s4, taking the minimum total time and the minimum total cost as optimization objectives, and constructing a multi-objective optimization model; s5, solving the multi-objective optimization model to obtain a personnel-task matching scheme; in the solving process, based on a task information table, a personnel information table, a task similarity matrix and a proficiency prediction model, the proficiency of each personnel engaged in each task is solved; and S6, the obtained personnel-task matching scheme is used to carry out personnel scheduling and distribution. According to the method, the defects of an existing static evaluation model can be overcome, and the rationality, efficiency and cost effectiveness of human resource allocation in collaborative design of complex heavy equipment are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of collaborative design of complex heavy equipment, and in particular relates to a method for matching heavy equipment collaborative design tasks and personnel with proficiency taken into consideration. Background Art

[0002] As the manufacturing industry undergoes a profound transformation toward intelligent, networked, and service-oriented manufacturing, the complex heavy equipment industry is facing unprecedented pressure to upgrade and innovation challenges. Faced with increasingly fierce market competition, traditional, relatively isolated equipment design models are increasingly unable to meet the market's stringent demands for cost-effective, efficient delivery, and highly reliable products. Heavy equipment products themselves exhibit significant characteristics such as extremely complex structures, high levels of technological integration, lengthy design cycles, and massive investment. Their successful design relies heavily on the deep integration and efficient collaboration of multidisciplinary expertise (such as mechanics, electrical engineering, hydraulics, control, materials, etc.). This inherent complexity makes the design team's cross-disciplinary collaboration efficiency and the precise use of heterogeneous personnel capabilities key factors in determining the success or failure of the entire design project.

[0003] Focusing on the collaborative design of complex heavy equipment, the design process is generally characterized by a high degree of customization, a high variety of small batches, and a dynamic multidisciplinary and multi-stakeholder approach. In this context, achieving efficient, precise, and dynamic optimization between interdisciplinary design tasks and diverse human resources has become a core issue for improving the overall effectiveness of collaborative design and reducing R&D cycles and costs.

[0004] However, existing task-personnel matching and evaluation methods have the following obvious limitations when dealing with this complex proposition:

[0005] 1. Incomplete coverage of matching factors: Mainstream methods mainly rely on static personnel capability assessment models, and often assign tasks by simply solidifying the standard completion time of tasks or the basic skill labels of personnel. This method seriously ignores a variety of key dynamic factors that affect the actual work efficiency of personnel, such as: Task similarity: Designers who have performed highly similar tasks usually have higher proficiency and efficiency in new tasks. Historical completion frequency and experience accumulation: The number of times personnel have completed a certain type of task and the depth of experience significantly affect their proficiency. Personnel title level and depth of technical expertise: The actual performance of personnel at different levels and with different technical depths when handling complex tasks varies greatly.

[0006] 2. Single and static evaluation dimensions: Existing assessments are overly simplistic and static, failing to establish a multi-dimensional, dynamic evaluation metric (such as predicted actual proficiency) that comprehensively reflects the degree of alignment between task characteristics (complexity, required skill sets, etc.) and personnel dynamic capabilities (proficiency, accumulated experience). This results in the inability to accurately measure and predict the actual performance differences between personnel with the same basic skill set when performing tasks of varying difficulty and nature.

[0007] 3. Models struggle to cope with complex constraints: Actual collaborative design environments present multiple rigid constraints (e.g., skills must fully cover task requirements and individual workload limits), while optimization objectives often involve multiple conflicting dimensions (e.g., minimizing total time and minimizing total cost). Static models struggle to effectively balance multiple objectives while satisfying these complex constraints.

[0008] Due to the inherent complexity of complex heavy equipment design, the multidimensional heterogeneity of personnel capabilities, and the complexity of optimization models (high dimensionality, nonlinearity, and NP-hardness), developing a method that can accurately match tasks and personnel, achieve efficient resource allocation, and significantly improve collaborative efficiency faces extremely great technical challenges.

[0009] Therefore, how to solve the shortcomings of the existing static evaluation model, improve the rationality, efficiency and cost-effectiveness of human resource allocation in the collaborative design of complex heavy equipment, and ultimately support enterprises to agilely respond to market demand and accelerate product innovation has become an urgent problem to be solved. Summary of the Invention

[0010] In response to the above-mentioned deficiencies of the existing technology, the present invention provides a method for matching heavy equipment collaborative design tasks and personnel based on proficiency, which can address the deficiencies of existing static evaluation models and improve the rationality, efficiency and cost-effectiveness of human resource allocation in the collaborative design of complex heavy equipment.

[0011] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0012] A method for matching heavy equipment collaborative design tasks and personnel with proficiency considerations includes the following steps:

[0013] S1. Construct a task information table and a personnel information table; the task information table includes information of various dimensions of tasks; the personnel information table includes information of various dimensions of personnel;

[0014] S2. Calculate the similarity between tasks and construct a task similarity matrix;

[0015] S3. Construct a proficiency prediction model to calculate the proficiency of person i in the jth task based on the information in the person information table and the task similarity matrix;

[0016] S4. Constructing a multi-objective optimization model with minimization of total time and total cost as optimization objectives; the constraints of the multi-objective optimization model include: the skills mastered by the personnel fully cover the task requirements, and the personnel's workload is less than a preset load threshold;

[0017] S5. Solve the multi-objective optimization model to obtain a personnel-task matching solution; in the solution process, the proficiency of each person in each task is solved based on the task information table, the personnel information table, the task similarity matrix, and the proficiency prediction model;

[0018] S6. Use the obtained personnel-task matching solution to schedule and allocate personnel.

[0019] Compared with the prior art, the present invention has the following beneficial effects:

[0020] 1. Accurately predict the dynamic ability of personnel to perform tasks. Different from existing matching methods that rely on fixed task hours or static ability labels, this solution innovatively introduces a proficiency prediction model based on task similarity and multi-dimensional personnel information. This model can dynamically evaluate the expected performance level (proficiency) of personnel on different tasks. It not only considers the static skills they have mastered, but also fully incorporates multi-dimensional factors such as the characteristic association (similarity) of the task itself and the personnel's historical experience (implicit in the information table). This makes the evaluation of personnel's actual effectiveness more realistic, precise, and individualized.

[0021] 2. Leverage task relevance to uncover potential capabilities. By constructing a task similarity matrix, this solution proactively identifies and analyzes the inherent relevance between tasks. This information is effectively incorporated into proficiency predictions, enabling more accurate identification and utilization of the potential capabilities of individuals with prior experience in similar tasks on new tasks, thereby optimizing the basis for matching.

[0022] 3. Improve the overall efficiency and cost-effectiveness of human resource allocation. The optimization objective is to minimize both total time and total cost, while strictly constraining skill coverage and workload thresholds within the model. This means that the resulting person-task matching solution systematically seeks a globally optimal or near-optimal resource allocation solution while satisfying key business rules. Compared to existing methods that rely on simple rules or local optimization, this approach significantly shortens the overall design cycle and reduces labor and management costs.

[0023] 4. Enhance the feasibility and practicality of the matching solution. The explicitly defined constraints (full skill coverage, workload limits) and the comprehensive consideration of global objectives during the optimization process (such as the impact of proficiency on construction period and cost) ensure that the final matching solution not only meets the rigid requirements of actual project management but also effectively improves overall design efficiency and benefits.

[0024] In summary, this method effectively addresses the shortcomings of existing static evaluation models, such as their single evaluation dimension and neglect of task relevance and personnel dynamic capabilities, by introducing task similarity analysis and a multidimensional dynamic proficiency prediction mechanism, integrated into a global optimization framework that comprehensively considers time, cost, and multiple constraints. Compared with traditional methods, this method significantly improves the rationality of human resource allocation in the collaborative design of complex heavy equipment (more accurately matching task requirements with personnel capabilities), efficiency (shortening overall cycle time), and cost-effectiveness (optimizing human resource input), thus providing scientific support for the efficient collaborative design of complex equipment.

[0025] Preferably, in S2, the process of constructing the task similarity matrix includes:

[0026] Construct the skill similarity evaluation matrix q required for the task:

[0027]

[0028] Where q ij Characterizes the matching degree between task i and benchmark task in terms of the skill j dimension required by the task; n is the total number of tasks, u is the total number of required skill dimensions;

[0029] Construct the task type similarity evaluation matrix l:

[0030]

[0031] Where, l ij represents the matching degree between task i and benchmark task in terms of task type j dimension; v is the total number of task type dimensions;

[0032] Construct a comprehensive task similarity matrix:

[0033]

[0034] Where s i =(q i1 ,…,q iu ,l i1 ,…,l iv ) represents the comprehensive similarity code of task i, and n is the total number of design tasks;

[0035] The generalized Jaccard similarity is used to calculate the intersection-over-union ratio of the skill sets of the two tasks as the similarity between the two tasks:

[0036]

[0037] Get the task similarity matrix I:

[0038]

[0039] This setup enables a multi-dimensional fusion assessment of task similarity. Unlike similarity assessments based on a single dimension (e.g., skills alone), this solution innovatively integrates two key dimensions: the required skills (matrix q) and the task type (matrix l) to construct a comprehensive similarity encoding (matrix H). This fusion assessment captures the essential connections between tasks in a more comprehensive and structured manner, avoiding misjudgments of similarity due to missing dimensions and laying a solid foundation for accurate matching.

[0040] 2. Improve the precision and rationality of similarity calculation. The solution uses generalized Jaccard similarity to calculate the similarity value x between specific tasks. ij By calculating the intersection-over-union ratio of skill feature sets, this method effectively quantifies partial and fuzzy matching relationships (not just 0 / 1 matching). This method better reflects the fact that task skill requirements in real-world scenarios often overlap rather than being completely identical. Compared to simple Euclidean distance or cosine similarity, this method is more targeted at characterizing discrete feature sets.

[0041] 3. Build a standardized and computable similarity relationship network. The final output task similarity matrix I has a diagonal of 1 (completely similar to itself) and non-diagonal elements are calculated x ij This matrix fully characterizes the relative similarities of all task pairs in a standardized form, forming a clear and symmetrical relational network data structure. This structured representation facilitates direct and efficient invocation and processing of subsequent optimization algorithms (such as the models in S3-S6), improving the engineering practicality and computational feasibility of the overall approach.

[0042] In summary, this technical solution overcomes the shortcomings of traditional task similarity assessment, such as single dimension, rough calculation method, and difficulty in structured application, through multi-dimensional feature fusion, generalized Jaccard similarity quantification, and standardized relationship matrix construction. It significantly improves the comprehensiveness, accuracy, and computability of task similarity assessment, and provides a solid and reliable basic support for downstream optimization such as personnel matching and resource scheduling based on task relevance in complex collaborative design.

[0043] Preferably, in S3, the expression of the proficiency prediction model is:

[0044]

[0045] Where m represents the historical completion frequency, which is obtained by counting the number of historical executions of similar tasks; o represents the task similarity, which is determined by calculating the mean similarity between historical tasks and current tasks through the similarity matrix; k represents the personnel title level; α, β, γ, and b are the parameters of the fitting curve to be determined.

[0046] This setup 1. achieves the effective integration and quantification of multi-dimensional influencing factors. The model groundbreakingly integrates three key dynamic factors as proficiency prediction inputs: historical completion frequency (m): quantifies the direct practical experience accumulated by personnel in similar tasks; personnel title level (k): reflects the overall ability level and depth of professional knowledge of the personnel; task similarity (o): quantifies the value of transfer experience based on the correlation between historical tasks and current tasks. Compared with traditional static models (which rely only on a single skill label or fixed task time), this model is the first to mathematically structure the integration of the core three elements of "experience-ability-relevance" that affect the actual execution efficiency of personnel, significantly improving the depth of the portrayal of personnel's potential performance.

[0047] 2. Scientifically handle the nonlinear growth effect of experience and ability. The model uses a logarithmic function form (ln(...)) to process the linear combination term (α*m+β*k+γ*o), accurately representing the general growth law of professional ability of "significant improvement in the novice period and slower improvement in the proficient period"; and effectively avoids the marginal contribution of high-frequency or high-level features from being infinitely magnified, which is closer to the real scene. At the same time, the square root function As the denominator, it further balances the scale effect of the growth of cumulative eigenvalues ​​on the prediction results, enhancing the mathematical robustness of the model and the rationality of the results.

[0048] 3. Provides a highly interpretable and adaptable forecasting framework. The design of the model's fitting parameters, α, β, γ, and b, endows the model with trainable optimization properties. This allows for parameter calibration based on specific scenario historical data, improving forecast accuracy and scenario fit. Parameters α, β, and γ clearly quantify the differential contributions of historical experience, professional title level, and task similarity to individual proficiency, providing a quantitative basis for managers to analyze talent effectiveness. The intercept parameter, b, adjusts the baseline level of the forecast. This flexible, structured framework not only supports model migration applications but also ensures the transparency of the forecast logic and its value as a reference for decision-making.

[0049] In summary, this proficiency prediction model creatively integrates and quantifies three core elements: depth of experience (m), competency level (k), and task transfer potential (o). It also employs a logarithmic-square root composite function structure to scientifically model the nonlinear growth characteristics of competency and the balance between features. This fundamentally overcomes the significant flaw of traditional static models that ignore task relevance and the dynamic nature of a person's experience. Compared to existing methods that rely solely on static skill labels or preset task durations, this model achieves refined, personalized, and interpretable predictions of a person's execution effectiveness, establishing a scientific and reliable competency assessment benchmark for subsequent optimization and matching solutions.

[0050] Preferably, when calculating the fitting parameters α, β, γ, and b, the compression ratio of the actual working hours of historical tasks to the standard working hours is used to quantitatively represent the corresponding proficiency p ij , combined with the corresponding m, k, o, the model parameters α, β, γ, and b are iteratively optimized by the least squares method.

[0051] This setup offers 1. Quantifiable proficiency mapping based on real business data. The solution innovatively uses the "compression ratio of actual work hours compared to standard work hours for historical tasks" as a quantitative indicator of proficiency (pij). This method directly links to the company's core business data (actual work hours vs. standard work hours) and truly reflects the actual improvement in personnel's performance in completing historical tasks (a higher compression ratio indicates greater proficiency). It overcomes the one-sidedness of traditional subjective scoring or single completion time metrics, establishing a measurable, traceable, and standardized objective representation of proficiency.

[0052] 2. Build a data-driven parameter optimization mechanism. Use the least squares method to iteratively optimize to determine the key parameters α, β, γ, and b, and effectively utilize the actual rules in historical data (a large number of p ij and corresponding m, k, o records), find the optimal fit through mathematical methods; ensure that the proficiency value P predicted by the model ij and the historical true value p ij The overall deviation is minimized, significantly improving the overall accuracy and generalization ability of the prediction model; the iterative process ensures the optimality or strong approximation of the solution, avoiding the subjectivity and inefficiency of manual parameter adjustment.

[0053] 3. Enhance the engineering adaptability and prediction reliability of the model. This optimization method has a clear mathematical basis and computational feasibility (least squares is a mature optimization technology): the optimized parameters give the model the ability to flexibly adapt to the characteristics of historical data of different enterprises and projects, supporting the migration of the model across scenarios; the optimization results make the internal logic of the model (the weights of α, β, γ) more in line with the actual experience distribution, and the predicted value has a higher reference value for actual production arrangements; it provides a reference for subsequent P-based ij It provides reliable and credible quantitative input basis for task allocation and resource scheduling.

[0054] Preferably, in S4, the objective function of the multi-objective optimization model is:

[0055]

[0056] Where T represents the total time required to complete the entire design task, T max T represents the longest time required for the entire design task to be completed by different personnel. min represents the shortest time required to complete the entire design task by different personnel; C represents the total cost required to complete the entire design task, C max represents the maximum cost of the entire design task due to the use of different personnel to complete it, C min represents the minimum cost required to complete the entire design task by using different personnel; ω1 and ω2 represent the weights of T and C in the objective function respectively, and Z0 represents the objective function value.

[0057] This setting can achieve standardized integration and balance of multi-objective benefits. It innovatively transforms the total time (T) and total cost (C) two objectives with different physical dimensions but the same optimization direction into a single optimizable objective function value Z0 through linear weighted standardization; and Normalization is performed to eliminate dimensional differences and make the contribution value of each target fall within the range of [0,1], significantly improving the comparability and balance between targets; the weight coefficients ω1 and ω2 give the management the ability to dynamically adjust priorities (such as time-sensitive or cost-sensitive projects).

[0058] 2. Provide a transparent scale to quantify the global optimization effect. The construction of the objective function Z0 has clear mathematical meaning and business interpretability: the numerator (T max -T / C max -C) intuitively reflects the absolute improvement of the current allocation scheme compared with the worst historical scheme; the denominator (T max -T min / C max -C min ) represents the theoretical maximum optimization space of the goal; the higher the Z0 value, the stronger the degree to which the comprehensive benefit (time and cost weighted) approaches the optimal solution, providing a quantitative ranking basis for decision makers to select Pareto solutions.

[0059] 3. Support efficient dynamic resource scheduling decisions. This model, as the optimization core, can be combined with the personnel dynamic proficiency (P ijThe influencing factors are closely linked to task working hours (T) and personnel cost coefficient (influencing C); under the premise of meeting skill coverage constraints and load threshold limits, the optimal solution for the comprehensive configuration of limited human resources in the dual dimensions of time and cost is achieved by maximizing Z0; it avoids the common bias in single-objective optimization and is more in line with the multi-objective trade-off needs of actual engineering.

[0060] In summary, this method, through normalized weighted fusion and global utility quantification, overcomes the limitations of traditional single-objective models or manual experience-based allocation, which often struggle to reconcile schedule and cost conflicts and lack a unified benefit evaluation benchmark. The constructed Z0 function provides a highly transparent and comparable decision-making basis for finding optimal human resource allocation solutions under complex constraints, significantly improving the ability to comprehensively manage both timeliness and cost-effectiveness in the collaborative design of complex equipment.

[0061] Preferably, the actual time T for person i to complete the j-th task is aij The calculation formula is:

[0062] T aij =Min{T ej (1-p max ),T ej (1-p ij )};

[0063] Where p max is the preset maximum proficiency; p ij T represents the proficiency of person i in performing the jth type of task; ej (1-p max ) represents the maximum proficiency p based on the preset max Calculate the shortest time, T ej (1-p ij ) indicates that the ij Calculate the theoretical time it takes for person i to perform the jth type of task.

[0064] This setting, 1. Introducing a dual dynamic guarantee mechanism to ensure the rationality of the working time estimate. Use the Min{} function to nest two dynamic calculation items: T ej (1-p max ) Calculate the shortest time based on the preset maximum proficiency; T ej (1-p ij ) calculates the theoretical time required based on the individual worker's individual proficiency (actual ability) for the task. By minimizing the two, we fully leverage the worker's actual potential while strictly avoiding the risk of time underestimation due to extreme proficiency prediction bias, mathematically guaranteeing the safety and feasibility of working hours.

[0065] 2. Realize dynamic personalization and scenario adaptability of work time estimation. The core variable p in the formula ij(The dynamic proficiency of person i for task type j) is derived from the previous prediction model (e.g., step S3). This allows: for the same task, people with different abilities and experience will receive differentiated time estimates; for the same person, tasks with different technical characteristics or similarities will trigger dynamic efficiency expectations. This breaks through the rigidity of traditional static work hour quotas and significantly improves task scheduling accuracy and resource utilization.

[0066] 3. Preset parameters provide a flexible control interface for engineering management. max (Preset maximum proficiency) as a configurable parameter: Managers can flexibly set it (such as 0.3 or 0.5) based on project risk tolerance or industry experience thresholds; provide a safety net for extreme scenarios (such as inexperienced personnel performing difficult tasks); balance efficiency optimization demands with project implementation robustness, and enhance the method's engineering implementation capabilities.

[0067] In summary, this method innovatively integrates personalized proficiency drive (p ij ) and system security boundary control (p max ) and utilizes the Min{} function for dual dynamic verification, fundamentally addressing the drawbacks of traditional fixed-hour models or single-proficiency prediction methods, such as uncontrollable estimation risk, insufficient flexibility, and a lack of safety redundancy. This formula ensures robust and reliable estimation results while maximizing the potential for human efficiency, providing a highly accurate and robust time benchmark input for subsequent resource scheduling and project management.

[0068] Preferably, the actual cost C for person i to complete the j-th task ij The calculation formula is:

[0069] C ij =T aij *[1+ξ(k-1)];

[0070] In the formula, k is the personnel level, and ξ represents the salary coefficient that changes with the promotion of job level.

[0071] Such a setting transforms traditional empirical pricing into a structured mathematical model, truly reflecting the business logic of "high skills = high cost".

[0072] 2. Dynamic working hours T aij (Including the influence of proficiency) Deep coupling: working hours T aij Compression (efficient personnel) can partially offset high-level costs (such as experts completing tasks quickly); avoid the "double high costs" (time-consuming and high unit price) generated by inefficient high-level personnel; provide a micro-computing basis for the goal of minimizing total costs and help optimize global resources.

[0073] 3. Parameter ξ acts as a configurable lever. Companies can adjust ξ values ​​based on project priorities (e.g., cost-sensitive vs. quality-sensitive projects). Industry differences (e.g., high-end equipment vs. general-purpose equipment) can be customized through ξ to achieve cost models. This breaks the rigidity of one-size-fits-all cost accounting and adapts to complex management scenarios.

[0074] In summary, this method, by structurally embedding personnel grades and adjustable compensation coefficients into the cost formula and combining them with dynamic working hours, thoroughly overcomes the shortcomings of traditional methods, such as the disconnect between labor costs and skill value, static pricing that ignores efficiency linkage, and a lack of enterprise control interfaces. While ensuring the rigor of the cost model, it provides a highly accurate and flexible quantitative tool for differentiated human resource pricing and global cost optimization in the collaborative design of complex equipment.

[0075] Preferably, in S5, an elite-preserving genetic algorithm is used to solve the multi-objective optimization model.

[0076] This setup, through the deep integration of an elite solution inheritance mechanism and heuristic global search capabilities, effectively overcomes the inefficiency and unstable solution quality of traditional optimization methods in high-dimensional, strongly constrained, and multi-objective scenarios. The resulting Pareto-optimal solution set provides a library of task-personnel matching solutions for the collaborative design of complex heavy equipment, combining high timeliness, low cost, and management flexibility. This fundamentally improves the scientific nature and adaptability of resource allocation decisions.

[0077] Preferably, in S1, the information in the task information table includes the task name, task ID, required skills, task status, and task type of the design task; the task status includes not executed, executing, and completed.

[0078] This setup, along with the multi-dimensional structured design of the task information table, overcomes the shortcomings of traditional task descriptions, such as missing attributes, dynamic disconnection, and cross-system fragmentation, through the precise definition and coordinated interaction of key fields (skills, type, and status). It provides a complete, real-time, and actionable data foundation for subsequent similarity calculations, proficiency predictions, and multi-objective optimization, significantly improving the sophistication of task modeling and the reliability of process linkage in the collaborative design of complex heavy equipment.

[0079] Preferably, the information in the personnel information table includes the designer's number, technical level, skill type, maximum working time, and personnel status; the personnel status includes idle state, underloaded state, and fully loaded state.

[0080] This setup systematically addresses the core pain points of traditional approaches, such as one-sided personnel profiles, disconnected resource status, and inaccurate constraints, by integrating static capability dimensions (technical level, skill type), dynamic load constraints (maximum working hours), and real-time status tracking (idle / underutilized / fully loaded). It provides a highly accurate and timely personnel data foundation for optimized matching in complex collaborative designs, ensuring that human resource allocation plans are both adaptable and feasible. BRIEF DESCRIPTION OF THE DRAWINGS

[0081] In order to make the purpose, technical solutions and advantages of the invention more clear, the present invention will be further described in detail below with reference to the accompanying drawings, in which:

[0082] Figure 1 Flowchart of this method;

[0083] Figure 2 This is a schematic diagram of the "task-personnel" comparison method in Example 1;

[0084] Figure 3 This is an example of the "mission-personnel" gene encoding in Example 1;

[0085] Figure 4 This is a three-dimensional assembly diagram of the 120MN extruder in Example 2;

[0086] Figure 5 This is the difference curve between the fitted value and the actual value of the proficiency of personnel with intermediate and lower professional titles in Example 2;

[0087] Figure 6 This is a comparison diagram of the fitness curves of the two types of genetic algorithms in Example 2;

[0088] Figure 7 This is a comparison chart of the effects of different algorithms in Example 2. DETAILED DESCRIPTION

[0089] The following is a further detailed description through specific implementation methods:

[0090] Example 1

[0091] like Figure 1 As shown, this embodiment discloses a method for matching heavy equipment collaborative design tasks and personnel considering proficiency, including the following steps:

[0092] S1. Construct a task information table and a personnel information table; the task information table includes information of various dimensions of tasks; the personnel information table includes information of various dimensions of personnel.

[0093] During specific implementation, the information in the task information table includes the task name, task ID, required skills, task status, and task type of the design task; the task status includes not executed, executing, and completed.

[0094] The information in the personnel information table includes the designer's number, technical level, skill type, maximum working time, and personnel status; the personnel status includes idle state, underloaded state, and fully loaded state.

[0095] To help those skilled in the art better understand the task information table and the personnel information table, the following description is given.

[0096] The task (i.e., design task) information table description model can be expressed as follows:

[0097] TaskInfo={ID,TaskTime,TaskSkills,TaskState,TaskType};

[0098] Among them, ID: PName-FTask-STask, the design task number, encodes the design task and includes the product name to which the design task belongs, the names of the first-level subtask and the second-level subtask, and other information. These three types of information together determine the unique number of the subtask.

[0099] PName represents the product name, as shown in Table 1;

[0100] Table 1 Product name code

[0101]

[0102] Ftask represents the name of the first-level subtask, as shown in Table 2:

[0103] Table 2 Level 1 task coding

[0104]

[0105] STask represents the name of the second-level subtask, as shown in Table 3:

[0106] Table 3 Secondary task coding

[0107]

[0108] From the above coding method, it can be seen that the unique task ID of the hydraulic design oil tank and its corresponding component design subtask in the extruder design project is "010201".

[0109] TaskTime: Design task time information, mainly stipulates the standard working time T of the jth task ej and the actual working time T of person i to complete the jth task after considering proficiencyaij , in days, as well as the start and end times of the task.

[0110] TaskSkills: Design task skills, specifying the specific skills required to complete a task. These skills include product type and specific skill requirement information. The combination of product type and specific skill requirements uniquely identifies the skill requirements for a task. Skill requirements reflect the specific technical characteristics of the task. Table 4 lists the specific skill attributes corresponding to skill requirement identifiers.

[0111] Table 4 Task requirement skill identification code

[0112]

[0113] From the above coding method, it can be seen that for the hydraulic design of the oil tank and its related component design subtask in the extruder design project, the task required skills are "drawing / geometric modeling" and "hydraulic structure design", and the unique coding of its task required skills is "0101, 0106".

[0114] TaskState: Design task status, describes the current status of the design task, and has three forms: unexecuted, executing, and completed.

[0115] TaskType: Design task type, which describes the type of design task and is determined by the product type and task code name, as shown in Table 5.

[0116] Table 5 Design task type coding

[0117]

[0118] From the above, we can see that for the hydraulic design of the oil tank and its associated component design subtask in the extruder design project, its task code is "010201", and its task type code is "0102".

[0119] The personnel information table can be expressed as follows:

[0120] esignerInfo={ID,Level,Skills,MaxLoad,DeState};

[0121] Among them, ID: designer number, which encodes the designer and includes the designer's name information.

[0122] Level: Designer level, indicating the designer's professional title level. Different professional levels have different salary coefficients.

[0123] Skills: Designer skills, representing all types of skills that designers possess. The skills mastered by the designer and the skills required for the task are considered here as different expressions of the same attribute. Therefore, the specific attributes of the skills mastered by the designer are shown in Table 4.

[0124] MaxLoad: maximum working time T of the designer i , T i It represents the maximum working time that designer i can allocate in a collaborative design project cycle, in days.

[0125] DeState: Designer status, which indicates the different working states of designers due to the number of design tasks they are currently handling. It can be divided into three states: idle state, underloaded state, and fully loaded state. The workload of designers can be defined as:

[0126]

[0127] Where, T aij / T i It represents the ratio of the time that person i spends completing the jth task to the longest working time of designer i, L i is the accumulated load value of designer i, L0 is the initial load value of designer i, and L is the total load value of designer i.

[0128] S2. Calculate the similarity between tasks and construct a task similarity matrix.

[0129] In specific implementation, the process of constructing the task similarity matrix includes:

[0130] Construct the skill similarity evaluation matrix q required for the task:

[0131]

[0132] Where q ij Characterizes the matching degree between task i and benchmark task in terms of the skill j dimension required by the task; n is the total number of tasks, u is the total number of required skill dimensions;

[0133] Construct the task type similarity evaluation matrix l:

[0134]

[0135] Where, l ij represents the matching degree between task i and benchmark task in terms of task type j dimension; v is the total number of task type dimensions;

[0136] Construct a comprehensive task similarity matrix:

[0137]

[0138] Where s i =(q i1 ,…,q iu ,l i1 ,…,l iv ) represents the comprehensive similarity code of task i, and n is the total number of design tasks;

[0139] The generalized Jaccard similarity is used to calculate the intersection-over-union ratio of the skill sets of the two tasks as the similarity between the two tasks:

[0140]

[0141] Get the task similarity matrix I:

[0142]

[0143] In this way, a multi-dimensional fusion evaluation of task similarity is achieved. Different from the similarity judgment of a single dimension (such as skills only), this solution innovatively integrates the two key dimensions of the skills required for the task (matrix q) and the task type (matrix l) to construct a comprehensive similarity code (matrix H). This fusion evaluation can capture the essential relationship between tasks in a more comprehensive and structured way, avoid similarity misjudgment caused by missing dimensions, and lay a solid foundation for accurate matching. In addition, the precision and rationality of similarity calculation are improved. The solution uses generalized Jaccard similarity to calculate the similarity value x between specific tasks. ij . This method can effectively quantify the relationship between partial matching and fuzzy matching (not just 0 / 1 matching) by calculating the intersection-over-union ratio of the skill feature set, which is more in line with the characteristics of the task skill requirements in actual scenarios where there is overlap rather than complete equality. Compared with simple Euclidean distance or cosine similarity, this method is more targeted in describing discrete feature sets. In addition, a standardized and computable similarity relationship network can be constructed. The final output task similarity matrix I has a diagonal of 1 (completely similar to itself) and non-diagonal elements are the calculated x ij This matrix fully characterizes the relative similarities of all task pairs in a standardized form, forming a clear and symmetrical relational network data structure. This structured representation facilitates direct and efficient invocation and processing of subsequent optimization algorithms (such as the models in S3-S6), improving the engineering practicality and computational feasibility of the overall approach.

[0144] S3. Construct a proficiency prediction model to calculate the proficiency of person i in the jth type of task based on the information in the person information table and the task similarity matrix.

[0145] The expression of the proficiency prediction model is:

[0146]

[0147] Where m represents the historical completion frequency, which is obtained by counting the number of historical executions of similar tasks; o represents the task similarity, which is determined by calculating the mean similarity between historical tasks and current tasks through the similarity matrix; k represents the personnel title level; α, β, γ, and b are the parameters of the fitting curve to be determined.

[0148] In this way, the effective integration and quantification of multi-dimensional influencing factors are achieved. The model has groundbreakingly integrated three key dynamic factors as proficiency prediction inputs: historical completion frequency (m): quantifies the direct practical experience accumulated by personnel in similar tasks; personnel title level (k): reflects the overall ability level and professional knowledge depth of personnel; task similarity (o): quantifies the value of transfer experience based on the correlation between historical tasks and current tasks. Compared with traditional static models (relying only on a single skill label or fixed task time), this model is the first to mathematically structure the integration of the core three elements of "experience-ability-relevance" that affect the actual execution efficiency of personnel, significantly improving the depth of the portrayal of personnel's potential performance. In addition, it scientifically handles the nonlinear growth effect of experience and ability. The model uses a logarithmic function form (ln(...)) to process the linear combination term (α*m+β*k+γ*o), accurately representing the general growth law of professional ability of "significant improvement in the novice period and slower improvement in the proficiency period"; and effectively avoids the marginal contribution of high-frequency or high-level features from being infinitely amplified, which is closer to the real scene. At the same time, the square root function As the denominator, it further balances the scale effect of the growth of the cumulative eigenvalue on the prediction results, enhancing the mathematical robustness of the model and the rationality of the results. In addition, it provides a highly interpretable and flexible prediction framework. The design of the parameters to be fitted in the model, α, β, γ, and b, gives the model trainable optimization properties, and can calibrate parameters based on historical data of specific scenarios to improve the accuracy of predictions and the degree of fit to the scenario; the parameters α, β, and γ clearly quantify the differentiated contribution weights of historical experience, professional title level, and task similarity to individual proficiency, providing a quantitative basis for managers to analyze talent effectiveness; the intercept parameter b can adjust the baseline level of the prediction value. This flexible structured framework not only supports model migration applications, but also ensures the transparency of the prediction logic and its reference value for decision-making.

[0149] In the specific implementation, when calculating the fitting parameters α, β, γ, and b, the compression ratio of the actual working hours of historical tasks compared to the standard working hours is used to quantitatively represent the corresponding proficiency p ij , combined with the corresponding m, k, o, the model parameters α, β, γ, and b are iteratively optimized by the least squares method.

[0150] In this way, the solution makes a breakthrough by using the "compression rate of actual working hours of historical tasks compared to standard working hours" as a quantitative indicator of proficiency pij. This method is directly related to the core business data of the enterprise (actual working hours vs. standard working hours), and truly reflects the actual efficiency improvement level of personnel in completing historical tasks (the higher the compression rate, the stronger the proficiency); it overcomes the one-sidedness of traditional subjective scoring or single completion time indicators, and establishes a measurable, traceable, and standardized objective representation method for proficiency. In addition, the least squares method is used for iterative optimization to determine the key parameters α, β, γ, and b, effectively utilizing the actual rules in historical data (a large number of p ij and corresponding m, k, o records), find the optimal fit through mathematical methods; ensure that the proficiency value P predicted by the model ij and the historical true value p ij The overall deviation is minimized, significantly improving the overall accuracy and generalization ability of the prediction model; the iterative process ensures the optimality or strong approximation of the solution, avoiding the subjectivity and inefficiency of manual parameter adjustment. In addition, this optimization method has a clear mathematical basis and computational feasibility (least squares is a mature optimization technology): the optimized parameters give the model the ability to flexibly adapt to the characteristics of historical data of different enterprises and projects, supporting the migration of the model across scenarios; the optimization results make the internal logic of the model (the weights of α, β, γ) more in line with the actual empirical distribution, and the predicted value has a higher reference value for actual production arrangements; it provides a reference for subsequent P-based ij It provides reliable and credible quantitative input basis for task allocation and resource scheduling.

[0151] In specific implementation, the calculation process of the above task similarity and personnel proficiency is shown in Table 6.

[0152] Table 6 Calculation method of personnel proficiency

[0153]

[0154] The nonlinear regression solution process is as follows:

[0155] (1) Construction of variable system: Establish a multi-dimensional parameter system covering historical completion frequency (m), personnel title level (k), task similarity (o), and personnel proficiency (p). Among them, m is obtained by the statistics of the historical execution times of similar tasks; o is determined by calculating the mean similarity between historical tasks and current tasks through the similarity matrix; p is a quantitative representation based on the compression ratio of the actual working hours of historical tasks compared to the standard working hours;

[0156] (2) Differentiated parameter calculation: Parameter calculation is performed based on a nonlinear least squares algorithm, and parameter optimization is achieved using the curve_fit function. For parameter calculation of intermediate and lower-level personnel, a group data fusion approach is used. By collecting the historical task records of all personnel in this category, a unified training dataset is constructed, and the nonlinear least squares algorithm is used to globally optimize the shared model parameters. For parameter calculation of senior personnel, a personalized strategy is adopted. The group baseline model parameters are injected as initial values ​​into the individual model, and the parameters are then adjusted based on the individual historical data.

[0157] (3) Model visualization verification: Substitute the obtained parameters into the model to generate fitting values, and visually evaluate the model accuracy through the comparison curve between the benchmark value and the fitting value;

[0158] (4) Quantitative accuracy evaluation: using the residual sum of squares (SSR), regression sum of squares (SSE), total sum of squares (SST), determination coefficient (R 2 ) evaluation system to verify the effectiveness of the model.

[0159] S4. Constructing a multi-objective optimization model with minimization of total time and total cost as optimization objectives; the constraints of the multi-objective optimization model include: the skills mastered by the personnel fully cover the task requirements, and the personnel's workload is less than a preset load threshold;

[0160] In specific implementation, the objective function of the multi-objective optimization model is:

[0161]

[0162] Where T represents the total time required to complete the entire design task, T max T represents the longest time required for the entire design task to be completed by different personnel. min represents the shortest time required to complete the entire design task by different personnel; C represents the total cost required to complete the entire design task, C max represents the maximum cost of the entire design task due to the use of different personnel to complete it, C min represents the minimum cost required to complete the entire design task by using different personnel; ω1 and ω2 represent the weights of T and C in the objective function respectively, and Z0 represents the objective function value.

[0163] In this way, the standardized integration and balance of multi-objective benefits can be achieved. Innovatively, the total time (T) and total cost (C), two objectives with different physical dimensions but the same optimization direction, are transformed into a single optimizable objective function value Z0 through linear weighted standardization; the range method is used to and Normalization is performed to eliminate dimensional differences and ensure that the contribution values ​​of each target fall within the range [0,1], significantly improving the comparability and balance between targets. The weight coefficients ω1 and ω2 give management the ability to dynamically adjust priorities (such as time-sensitive or cost-sensitive projects). In addition, a transparent scale is provided to quantify the global optimization effect. The construction of the objective function Z0 has clear mathematical meaning and business interpretability: the numerator (T max -T / C max -C) intuitively reflects the absolute improvement of the current allocation scheme compared with the worst historical scheme; the denominator (T max -T min / C max -C min ) represents the theoretical maximum optimization space of the goal; the higher the Z0 value, the stronger the degree to which the comprehensive benefit (time and cost weighted) approaches the optimal solution, providing a quantitative ranking basis for decision makers to select Pareto solutions. In addition, it supports efficient dynamic resource scheduling decisions. This model, as the optimization kernel, can be combined with the personnel dynamic proficiency (P ij The influencing factors are closely linked to task working hours (T) and personnel cost coefficient (influencing C); under the premise of meeting skill coverage constraints and load threshold limits, the optimal solution for the comprehensive configuration of limited human resources in the dual dimensions of time and cost is achieved by maximizing Z0; it avoids the common bias in single-objective optimization and is more in line with the multi-objective trade-off needs of actual engineering.

[0164] In specific implementation, the actual time T for person i to complete the jth task aij The calculation formula is:

[0165] T aij =Min{T ej (1-p max ),T ej (1-p ij )};

[0166] Where p max is the preset maximum proficiency; p ij T represents the proficiency of person i in performing the jth type of task; ej (1-p max ) represents the maximum proficiency p based on the preset max Calculate the shortest time, T ej (1-p ij ) indicates that the ij Calculate the theoretical time it takes for person i to perform the jth type of task.

[0167] Thus, the present method innovatively integrates personalized proficiency drive (p ij ) and system security boundary control (p max) and utilizes the Min{} function for dual dynamic verification, fundamentally addressing the drawbacks of traditional fixed-hour models or single-proficiency prediction methods, such as uncontrollable estimation risk, insufficient flexibility, and a lack of safety redundancy. This formula ensures robust and reliable estimation results while maximizing the potential for human efficiency, providing a highly accurate and robust time benchmark input for subsequent resource scheduling and project management.

[0168] The actual cost C of person i completing the j-th task ij The calculation formula is:

[0169] C ij =T aij *[1+ξ(k-1)];

[0170] In the formula, k is the personnel level, and ξ represents the salary coefficient that changes with the promotion of job level.

[0171] By structurally embedding personnel grades and adjustable compensation coefficients into cost formulas and combining them with dynamic working hours, this method thoroughly overcomes the shortcomings of traditional methods, such as the disconnect between labor costs and skill value, static pricing that ignores efficiency linkage, and a lack of enterprise control interfaces. While ensuring the rigor of the cost model, it provides a highly accurate and flexible quantitative tool for differentiated human resource pricing and global cost optimization in the collaborative design of complex equipment.

[0172] S5. Solve the multi-objective optimization model to obtain a personnel-task matching solution. In the solution process, the proficiency of each person in each task is solved based on the task information table, the personnel information table, the task similarity matrix and the proficiency prediction model.

[0173] In practice, an elite-preserving genetic algorithm is used to solve multi-objective optimization models. This deep integration of an elite solution inheritance mechanism and heuristic global search capabilities effectively overcomes the inefficiency and unstable solution quality of traditional optimization methods in high-dimensional, strongly constrained, and multi-objective scenarios. The resulting Pareto-optimal solution set provides a library of task-personnel matching solutions that are both time-efficient, cost-effective, and manageably flexible for the collaborative design of complex heavy equipment, fundamentally improving the scientific nature and adaptability of resource allocation decisions.

[0174] To facilitate a better understanding of the technology in this field, the following description is given of the use of an elite-preserving genetic algorithm to solve a multi-objective optimization model.

[0175] This method introduces an elite retention strategy, establishing a fitness ranking mechanism to forcibly retain the top-ranked solutions in each generation, effectively accelerating the population's convergence rate. By simulating genetic operations such as selection, crossover, and mutation during biological evolution, the algorithm significantly improves global optimization capabilities while maintaining solution diversity.

[0176] (1) Coding

[0177] The core of the "design task-designer" matching mechanism is to achieve accurate adaptation between task requirements and personnel capabilities. This study constructed a genetic algorithm model based on discrete chromosome encoding. The encoding rules contain three constraints:

[0178] 1) Task encoding mechanism: The task set is represented by an n-dimensional chromosome, and the loci are arranged in ascending order by task number to eliminate temporal interference;

[0179] 2) Personnel mapping mechanism: Gene values ​​are encoded using personnel unique identifiers to construct a task-personnel mapping relationship matrix;

[0180] 3) Solution space constraints: The chromosome length is strictly equal to the total number of tasks, and the value space of each gene locus is limited to the set of candidate candidates who are qualified to perform the corresponding task.

[0181] This encoding method constructs a task allocation model through discrete gene sequences, which not only meets the needs of multi-task parallel processing, but also avoids the interference of timing parameters on optimization targets in traditional methods. Each gene locus corresponds to a specific task unit. During the allocation process, the system compares the skill set required for the task with the skill set mastered by the personnel: first, the task requirement skill list is retrieved based on the gene locus index, and then combined with the personnel information, when it is detected that the skills mastered by a certain person completely cover the task requirements, the system compares the skill set required for the task with the skill set mastered by the personnel. When the task is completed, the code is written into the corresponding gene site; otherwise, the personnel are replaced until the task requirements are met.

[0182] After completing the task assignment, the workload of each designer is checked by the preset load threshold. When it is found that a person's workload exceeds his or her maximum load, the personnel replacement program is triggered - the task position is reallocated under the premise of meeting the skill matching, and the workload of all designers is adjusted to be within the normal range, such as Figure 2 shown.

[0183] Through this coding method, task information and personnel information can be presented on a chromosome. When all tasks are assigned to appropriate designers, the genetic coding of a chromosome is completed, and a "task-personnel" matching solution is obtained, such as Figure 3 shown.

[0184] (2) Fitness function value

[0185] For each "task-personnel" matching scheme, the corresponding total time T and total cost C are calculated. The objective function is to find the comprehensive minimum of total time and total cost. According to the standard normalization formula above, when T and C are smaller, Z0 is larger. Therefore, the fitness function can be expressed as follows:

[0186] f(i)=Max(Z0);

[0187] (3) Selection

[0188] Genetic algorithms require a selection process based on fitness to determine the next generation's parent individuals. This study employs a roulette wheel selection method combined with an elite retention strategy. The implementation process consists of two stages: first, the top five individuals in the current population by fitness are retained directly to the next generation, forming an elite population; second, a probabilistic selection strategy based on fitness ratio is applied to the remaining individuals. The calculation formula is as follows, where the probability of individual selection is positively correlated with its fitness value.

[0189]

[0190] Where f(i) is the fitness of individual i, n-5 represents the population size after elite retention, and P(i) is the probability that individual i is selected as the parent.

[0191] (4) Cross

[0192] With a certain probability, the parent generation undergoes genetic recombination according to the crossover genetic operator, producing a new generation of candidate solutions. However, due to the randomness of crossover, after the crossover is completed, the designer may not meet the skills required for the corresponding task, and may also be overloaded. Therefore, load verification and skill screening are required for crossover points that do not meet the requirements until the genes of the offspring after the crossover meet the task requirements.

[0193] (5) Variation

[0194] After the crossover operation, the newly generated individuals are mutated based on the mutation probability parameters. This study uses a single-site gene replacement strategy, randomly selecting mutation sites within the chromosome coding sequence to randomly change the corresponding personnel number. After the mutation is complete, the newly mutated individuals are subjected to load verification and skill screening to ensure that the new genes meet the mission requirements.

[0195] (6) Generate a new population

[0196] After three stages of genetic operations (selection, crossover, and mutation), an optimized offspring population is generated. The new population consists of the top five individuals with the highest fitness in the previous generation, individuals selected by probability, and some offspring individuals generated through crossover and mutation.

[0197] S6. Use the obtained personnel-task matching solution to schedule and allocate personnel.

[0198] Unlike existing matching methods that rely on fixed task hours or static competency labels, this solution innovatively introduces a proficiency prediction model based on task similarity and multidimensional personnel information. This model dynamically assesses a person's expected performance level (proficiency) on different tasks, taking into account not only their static skills but also factors such as the task's inherent characteristics (similarity) and the person's historical experience (implicit in the information table). This makes the assessment of a person's actual effectiveness more realistic, precise, and personalized. Furthermore, by constructing a task similarity matrix, this solution proactively identifies and analyzes the inherent correlations between tasks. This information is effectively incorporated into proficiency prediction, allowing the potential skills of people with similar historical task experience to be more accurately identified and utilized in new tasks, thereby optimizing the basis for matching. Furthermore, the optimization objectives are to simultaneously minimize total time and total cost, while strictly constraining skill coverage and workload thresholds within the model. This means that the resulting person-task matching solution systematically seeks a globally optimal or near-optimal resource allocation solution while satisfying key business rules. Compared to existing methods that rely on simple rules or local optimization, this approach significantly shortens the overall design cycle and reduces labor and management costs. Furthermore, the explicitly defined constraints (full skill coverage and workload limits) and the comprehensive consideration of global objectives during the optimization process (such as the impact of proficiency on construction period and cost) ensure that the resulting matching solution not only meets the rigid requirements of actual project management but also effectively improves overall design efficiency and effectiveness.

[0199] Specifically, the algorithm design process of the method for matching heavy equipment collaborative design tasks and personnel based on proficiency of the present invention can be shown in Table 7, as follows:

[0200] Table 7 Algorithm example of the matching method of heavy equipment collaborative design tasks and personnel considering proficiency of the present invention

[0201]

[0202] In summary, this method effectively addresses the shortcomings of existing static evaluation models, such as their single evaluation dimension and neglect of task relevance and personnel dynamic capabilities, by introducing task similarity analysis and a multidimensional dynamic proficiency prediction mechanism, integrated into a global optimization framework that comprehensively considers time, cost, and multiple constraints. Compared with traditional methods, this method significantly improves the rationality of human resource allocation in the collaborative design of complex heavy equipment (more accurately matching task requirements with personnel capabilities), efficiency (shortening overall cycle time), and cost-effectiveness (optimizing human resource input), thus providing scientific support for the efficient collaborative design of complex equipment.

[0203] Example 2

[0204] In order for those skilled in the art to better understand the effect of this method, the following examples are provided.

[0205] Take the 120MN extruder design task published by Z company on the collaborative manufacturing platform as an example. The 3D assembly drawing of the design project is as follows: Figure 4 As shown in the figure, one of the task sets consists of 9 independent subtasks. The number of designers available in the collaborative design resource pool is 16, and the salary of designers at level 1 is set at 300 yuan / day.

[0206] The design task information, designer information, and the frequency of designer completed tasks are shown in Table 8, Table 9, and Table 10. ij The value of represents the historical frequency of person i engaging in the jth type of task.

[0207] Table 8 Initial design task information

[0208]

[0209] Table 9 Initial designer information

[0210]

[0211] Table 10: Frequency of designers completing tasks in history

[0212]

[0213] The comparison of salary coefficients of designers with different professional titles in the extruder collaborative design project can be found in Table 11.

[0214] Table 11 Salary coefficient table corresponding to designer grades

[0215]

[0216] After obtaining the above task and personnel information, we can construct a similarity judgment matrix between tasks and use Jacarrd similarity calculation to obtain task similarity. Among them, the task required skill similarity evaluation matrix q is shown in Table 12:

[0217] Table 12. Evaluation matrix of similarity of skills required for tasks

[0218]

[0219] The task type similarity evaluation matrix l is shown in Table 13:

[0220] Table 13 Task type similarity evaluation matrix

[0221]

[0222] The task similarity evaluation matrix H is shown in Table 14:

[0223] Table 14 Task similarity evaluation matrix

[0224]

[0225] Through the above three matrices, the similarity values ​​between the tasks in the design task set are calculated, and the results are shown in Table 15:

[0226] Table 15 120MN extruder task similarity value matrix

[0227]

[0228] By using Python and calling the curve_fit function in the SciPy library, we can fit the parameters of the proficiency curve for designers with intermediate and lower professional titles: α = 0.92, β = 2.69, γ = 0.12, and b = 0.78 based on the task similarity value, historical completion frequency, personnel level, and baseline proficiency value. The proficiency calculation equation for this type of personnel is as follows:

[0229]

[0230] The comparison chart between the fitted value and the actual value can be obtained through the fitting equation, such as Figure 5 shown.

[0231] The residual squares, regression sum of squares, total sum of squares, and determination coefficient of the fitting equation are shown in Table 16:

[0232] Table 16 Fitting effect indicators

[0233]

[0234] For senior designers, the fitting parameters are solved based on their personal historical data. The relevant parameters and error index values ​​are shown in Table 17.

[0235] Table 17 Fitting parameters and error index values ​​of senior designer proficiency equation

[0236]

[0237] By analyzing the fitness curve of each generation of the population obtained by this method and the fitness curve obtained by the traditional genetic algorithm, it can be seen that when dealing with the same problem, this method requires fewer iterations to reach convergence, about 80 times, while the traditional genetic algorithm requires about 120 times to reach convergence, which is a reduction of 33.3%. Figure 6 shown.

[0238] In addition, this study uses the particle swarm optimization algorithm to compare the effectiveness of the algorithms. The total time and total cost of completing the design task using different methods are as follows:

[0239] The cost of using the genetic algorithm based on the elite strategy is 6831.44 yuan and the time is 18.59 days; the cost of using the particle swarm algorithm is 7466.72 yuan and the time is 19.55 days; the cost of randomly generating combinations is 7685.75 yuan and the time is 20.38 days. Figure 7 shown.

[0240] By comparison, the genetic algorithm using the elite retention strategy reduces the cost by about 8.51% and the time by about 4.91% compared with the particle swarm optimization algorithm; and reduces the cost by about 10.06% and the time by about 8.78% compared with the random generation combination.

[0241] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit the technical solutions. Those skilled in the art should understand that modifications or equivalent replacements of the technical solutions of the present invention that do not depart from the purpose and scope of the technical solutions of the present invention should be included in the scope of the claims of the present invention.

Claims

1. A method for matching heavy equipment collaborative design tasks and personnel considering proficiency, characterized by: The following steps are involved: S1. Construct a task information table and a personnel information table; the task information table includes information of various dimensions of tasks; the personnel information table includes information of various dimensions of personnel; S2. Calculate the similarity between tasks and construct a task similarity matrix; S3. Construct a proficiency prediction model to calculate the proficiency of person i in the jth task based on the information in the person information table and the task similarity matrix; S4. Taking the minimization of total time and total cost as the optimization objectives, a multi-objective optimization model is constructed; The constraints of the multi-objective optimization target model include: the skills mastered by the personnel fully cover the task requirements, and the personnel's workload is less than a preset load threshold; S5. Solve the multi-objective optimization model to obtain a personnel-task matching solution; in the solution process, the proficiency of each person in each task is solved based on the task information table, the personnel information table, the task similarity matrix, and the proficiency prediction model; S6. Use the obtained personnel-task matching solution to schedule and allocate personnel.

2. The method for matching heavy equipment collaborative design tasks and personnel based on proficiency as claimed in claim 1, characterized in that: In S2, the process of constructing the task similarity matrix includes: Construct the skill similarity evaluation matrix q required for the task: Where q ij Characterizes the matching degree between task i and benchmark task in terms of the skill j dimension required by the task; n is the total number of tasks, u is the total number of required skill dimensions; Construct the task type similarity evaluation matrix l: Where, l ij represents the matching degree between task i and benchmark task in terms of task type j dimension; v is the total number of task type dimensions; Construct a comprehensive task similarity matrix: Where s i =(q i1 ,…,q iu ,l i1 ,…,l iv ) represents the comprehensive similarity code of task i, and n is the total number of design tasks; The generalized Jaccard similarity is used to calculate the intersection-over-union ratio of the skill sets of the two tasks as the similarity between the two tasks: Get the task similarity matrix I:

3. The method for matching heavy equipment collaborative design tasks and personnel based on proficiency as claimed in claim 2, characterized in that: In S3, the expression of the proficiency prediction model is: Where m represents the historical completion frequency, which is obtained by counting the number of historical executions of similar tasks; o represents the task similarity, which is determined by calculating the mean similarity between historical tasks and current tasks through the similarity matrix; k represents the personnel title level; α, β, γ, and b are the parameters of the fitting curve to be determined.

4. The method for matching heavy equipment collaborative design tasks and personnel based on proficiency as claimed in claim 3, characterized in that: When calculating the fitting parameters α, β, γ, and b, the compression ratio of the actual working hours of historical tasks compared to the standard working hours is used to quantitatively represent the corresponding proficiency p ij , combined with the corresponding m, k, o, the model parameters α, β, γ, and b are iteratively optimized by the least squares method.

5. The method for matching heavy equipment collaborative design tasks and personnel based on proficiency as claimed in claim 4, characterized in that: In S4, the objective function of the multi-objective optimization model is: Where T represents the total time required to complete the entire design task, T max T represents the longest time required for the entire design task to be completed by different personnel. min Represents the shortest time required to complete the entire design task by using different personnel; C represents the total cost required to complete the entire design task, C max represents the maximum cost of the entire design task due to the use of different personnel to complete it, C min represents the minimum cost required to complete the entire design task by using different personnel; ω1 and ω2 represent the weights of T and C in the objective function respectively, and Z0 represents the objective function value.

6. The method for matching heavy equipment collaborative design tasks and personnel based on proficiency as claimed in claim 5, characterized in that: The actual time T for person i to complete the jth task aij The calculation formula is: T aij =Min{T ej (1-p max ),T ej (1-p ij )}; Where p max is the preset maximum proficiency; p ij T represents the proficiency of person i in performing the jth type of task; ej (1-p max ) represents the maximum proficiency p based on the preset max Calculate the shortest time, T ej (1-p ij ) indicates that the ij Calculate the theoretical time it takes for person i to perform the jth type of task.

7. The method for matching heavy equipment collaborative design tasks and personnel based on proficiency as claimed in claim 6, characterized in that: The actual cost C of person i completing the j-th task ij The calculation formula is: C ij =T aij *[1+ξ(k-1)]; In the formula, k is the personnel level, and ξ represents the salary coefficient that changes with the promotion of job level.

8. The method for matching heavy equipment collaborative design tasks and personnel based on proficiency as claimed in claim 1, characterized in that: In S5, the elite-preserving genetic algorithm is used to solve the multi-objective optimization model.

9. The method for matching heavy equipment collaborative design tasks and personnel based on proficiency as claimed in claim 1, characterized in that: In S1, the information in the task information table includes the task name, task ID, required skills, task status, and task type of the design task; the task status includes not executed, executing, and completed.

10. The method for matching heavy equipment collaborative design tasks and personnel based on proficiency as claimed in claim 9, characterized in that: The information in the personnel information table includes the designer's number, technical level, skill type, maximum working time, and personnel status; the personnel status includes idle state, underloaded state, and fully loaded state.

Images