College student labor ability cultivation path construction method based on project learning

Abstract

The invention relates to the technical field of labor ability cultivation, and discloses a project learning-based college student labor ability cultivation path construction method, which comprises the following steps of collecting original ability data, and constructing student ability initial measure distribution; project learning is divided into a plurality of stages, and time sequence capability target distribution of labor tasks is formed; establishing a capability evolution path optimality model; outputting a dynamic mapping relation from the students to the tasks; generating a task execution sequence and an ability growth track of each student; the task completion condition and the ability growth track of the student are collected in real time, and the ability measurement distribution of the student is dynamically corrected. According to the invention, accurate task allocation is realized. And an optimal model of a student ability evolution path is constructed, a personalized ability improvement path is provided, and the defects of a linear prediction model are avoided. The real-time accuracy of student ability evaluation is ensured, and the learning bottleneck problem caused by improper task arrangement is avoided by generating a personalized task sequence and a growth track.

Description

Technical Field

[0001] The present invention relates to the technical field of labor ability training, and in particular to a method for constructing a labor ability training path for college students based on project learning. Background Art

[0002] With the continuous deepening of educational reform, personalized education has gradually become a hot topic in the field of education. Traditional teaching methods often rely on the teacher's subjective experience and students' overall performance to assign tasks, lacking the precise identification and dynamic adjustment of individual students' abilities. This approach cannot fully adapt to the changes in students' abilities at different learning stages and during task execution, resulting in unsatisfactory learning outcomes. Therefore, how to dynamically adjust task assignments based on students' actual abilities and provide feedback and corrections after task completion are key to improving learning efficiency.

[0003] Most existing educational technology solutions use fixed competency assessment methods based on standardized assessments. These methods use standardized tests or quantitative indicators to initially assess student competency, then assign tasks based on the results. This approach is prevalent in traditional school education and some online education platforms. Most platforms also use a learning management system (LMS) to match tasks and manage progress, collecting data after students complete tasks to facilitate simple competency assessments. However, the core problem with these technical solutions is their lack of flexibility and real-time performance.

[0004] The shortcoming of existing technical solutions is that they usually adopt a static ability allocation model, which fails to reflect the changes in students' abilities during the task completion process in real time. For example, in traditional methods, students' task allocation is usually based on preliminary ability assessments, ignoring the students' changing ability status during the learning process. This makes students face mismatched task difficulty when performing tasks, resulting in low learning efficiency. In addition, existing solutions also have shortcomings in task feedback and student ability correction. Many systems do not update students' abilities in a timely manner after feedback collection, making it impossible to accurately adjust students' learning paths. This problem is particularly prominent in complex learning tasks and multi-task environments. Students' ability trajectories often deviate due to the lack of dynamic adjustment, which ultimately affects learning outcomes. Summary of the Invention

[0005] In response to the shortcomings of the existing technology, the present invention provides a method for constructing a labor capacity training path for college students based on project learning, which solves the problems of static and unchanged student capacity assessment, inaccurate task allocation and delayed capacity correction in the existing methods of cultivating college students' labor capacity.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: a method for constructing a labor capacity training path for college students based on project learning, comprising the following steps:

[0007] S1. Collect the original ability data of college students and standardize it to form ability vectors, construct the initial measurement distribution of students' abilities, and provide basic ability status information for subsequent task matching;

[0008] S2. Divide project learning into multiple stages and set corresponding target capability vectors for each stage to form a time series capability target distribution of labor tasks;

[0009] S3. Based on the initial measurement distribution of students’ abilities and the distribution of labor task targets, establish an optimal model for the ability evolution path;

[0010] S4. Based on the ability evolution path optimality model, the student's current ability status is optimally matched with the ability requirements of the corresponding stage task, and a dynamic mapping relationship between students and tasks is output;

[0011] S5. Based on the student-task matching results, generate each student's task execution sequence and ability growth trajectory;

[0012] S6. Collect students’ task completion status and ability growth trajectory in real time, dynamically correct students’ ability measurement distribution, and thus update subsequent ability evolution paths.

[0013] Preferably, the step of constructing the initial measurement of student abilities includes:

[0014] Collect students' original scores on multiple ability dimensions;

[0015] Standardize the scores and construct the ability vector;

[0016] The capability measure distribution is constructed based on the capability vector, and the form is:

[0017]

[0018] Among them, μ0 is the initial measurement of the student group's ability, x i is the ability vector of the i-th student, w i is the weight, is the Dirac measure.

[0019] Preferably, constructing the labor task time series capability target distribution includes:

[0020] Divide the labor project into successive phases;

[0021] Set target capability vectors and establish metrics for each phase of the task;

[0022] Generate target capability distribution time series, expressed as:

[0023]

[0024] Among them, y h (t) is the j-th task target capability vector, v j (t) is its distribution weight, M is the total number of task subunits, and T is the total task time.

[0025] Preferably, the capability evolution path model construction includes:

[0026] Define the continuous time measure flow μ of student ability evolution t ;

[0027] Constructing capability path cost function The form is:

[0028]

[0029] Set the path evolution constraint to the continuity equation:

[0030]

[0031] Where: μ t is the distribution of student ability measurement at time t, v t (x) is the velocity vector field of the student’s ability at position x, It is a d-dimensional ability space.

[0032] Preferably, the optimal path problem is solved by discretizing the time domain and adopting a variational method combined with a Lagrange multiplier method, and ensuring that the path evolution satisfies the continuity constraint.

[0033] Preferably, the optimal match includes:

[0034] Time discretization of capability evolution path;

[0035] The Sinkhorn iterative algorithm is used to achieve the optimal matching between students and tasks;

[0036] The matching metric is the regularized Wasserstein distance:

[0037]

[0038] where μ and ν are the current student ability distribution and task goal distribution, respectively; γ is the joint transition probability distribution between μ and ν; ε is the regularization coefficient; KL(·) is the Kullback-Leibler divergence; and Γ(μ,ν) is the set of joint measures where all edges are μ and ν.

[0039] Preferably, the task execution sequence generation step includes:

[0040] Record each student's task matching results;

[0041] Generate task execution sequence in, represents the matching task of the i-th student in the k-th stage;

[0042] Organize task lists and archive task schedules for system and teacher reference.

[0043] Preferably, the capability growth trajectory is represented as a path in the capability space:

[0044]

[0045] in, is the ability status of the i-th student at the k-th time node. This trajectory is used to evaluate the changing trend of students' abilities at different stages.

[0046] Preferably, dynamically modifying the capability measure distribution comprises the following sub-steps:

[0047] Collect capability assessment feedback data after task completion;

[0048] Constructing an observed distribution of student abilities

[0049] Update the capability distribution based on the feedback to generate a revised capability measure:

[0050]

[0051] in, is the distribution of model prediction capabilities, is the feedback capability measure, and λ∈[0,1] is the fusion coefficient.

[0052] A system for constructing a path for cultivating the labor capacity of college students based on project learning is applied to the method for constructing a path for cultivating the labor capacity of college students based on project learning according to any one of claims 1 to 9, and is characterized by comprising:

[0053] Ability measurement modeling module, used to collect student ability data and construct initial probability measures of the ability space;

[0054] Mission objective modeling module, used to analyze project structure and construct phased mission capability objective distribution;

[0055] Path modeling and optimization module, used to establish the optimal model of capability evolution path and solve it based on minimum action;

[0056] Numerical calculation and matching module, used to perform path optimization and dynamically match students and tasks;

[0057] Growth path output module, used to generate task sequences and capability growth trajectories;

[0058] Path feedback module, used to collect task feedback and update capability status;

[0059] Visual display module, used to output student growth reports and graphic path visualization information.

[0060] The present invention provides a method for constructing a path for cultivating college students' labor capacity based on project learning. It has the following beneficial effects:

[0061] 1. This invention utilizes a dynamic matching technology solution based on preliminary student ability measurements and the distribution of task target abilities, achieving precise matching between tasks and students. Compared to the static task allocation method commonly used in the prior art, this invention can adjust task allocation based on students' real-time ability status, resolving the problem of traditional methods' inability to flexibly adjust task matching as students' abilities change. Through this dynamic matching mechanism, students can complete tasks within an appropriate difficulty range, effectively promoting the improvement of their abilities.

[0062] 2. By constructing an optimal model for the evolution of student abilities, this invention allows personalized optimization of the ability-improvement path, achieving the technical effect of accurately predicting student growth trajectories. Compared to the simple linear ability prediction models used in existing technologies, this invention can provide an ability-improvement path that better reflects students' actual needs in a changing learning environment. This avoids the shortcomings of traditional methods that cannot cope with fluctuations in student abilities and task diversity, ensuring maximum effectiveness in improving student abilities.

[0063] 3. The present invention adopts a technical solution based on feedback-based ability path correction, which effectively solves the drawback of traditional methods of static student ability assessment. By collecting task feedback in real time and dynamically correcting students' ability distribution, the present invention can timely adjust the learning path during task execution to ensure that students' ability assessment remains accurate. Compared with the existing technology that relies only on initial data for student ability assessment, the present invention can flexibly adjust according to students' actual learning progress, improving the accuracy and adaptability of the assessment.

[0064] 4. The present invention's task sequence and growth trajectory generation technology ensures that every student can continuously progress along the optimal path during the learning process. Compared to existing solutions that lack personalized task arrangement and ability development paths, the present invention formulates personalized learning trajectories by considering the evolution of students' abilities and the changes in task difficulty, effectively avoiding learning bottlenecks caused by improper task arrangement. This precise path planning not only improves students' learning outcomes, but also provides teachers with a powerful teaching guidance basis. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 Schematic diagram of the method flow of the present invention;

[0066] Figure 2 Schematic diagram of the system architecture of the present invention. DETAILED DESCRIPTION

[0067] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the present specification. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0068] Please see the attached Figure 1 The embodiment of the present invention provides a method for constructing a labor capacity cultivation path for college students based on project learning, including:

[0069] S1. Collect the original ability data of college students and standardize it to form ability vectors, construct the initial measurement distribution of students' abilities, and provide basic ability status information for subsequent task matching;

[0070] In this embodiment, step S1 aims to achieve systematic quantification of the individual labor capacity status of college students. Specifically, through the three processes of capacity data collection, capacity vector construction and measurement modeling, an initial measurement distribution μ0 of the student's capacity status is formed as the starting point for modeling the capacity evolution path.

[0071] This step can be implemented by the "ability measurement modeling module" in the system. This module is embedded in the college student labor ability path construction system provided by the present invention, supports docking with the student information management platform, and completes the automatic extraction and processing of ability-related data.

[0072] In this step, we first define a capability space containing d capability dimensions. Each dimension corresponds to a core competency indicator that is highly relevant to labor education.

[0073] Preferably, competency dimensions may include, but are not limited to, organizational and coordination skills, collaborative skills, hands-on practical skills, task management skills, problem-solving skills, and self-regulation skills. Each competency dimension should be designed based on the university's labor education goals and the qualities required for project-based learning tasks, and should be preset by the faculty team or system strategy module.

[0074] Subsequently, the system collects the original ability data of each student in the above-mentioned ability dimensions. Data collection can preferably be achieved in the following ways:

[0075] Course performance data acquisition: The original evaluation is formed through students' stage performance and homework completion quality in relevant labor-related or practical courses;

[0076] Mentor evaluation system: Course instructors or project mentors will grade students based on preset indicators;

[0077] Self-evaluation and peer evaluation: Students submit their self-evaluation through the project system, and group members can evaluate each other;

[0078] Platform log extraction: extracting behavioral data such as task completion and participation activity from labor project platform records;

[0079] Questionnaire scoring: Combined with the standardized labor education questionnaire, students' self-evaluation and subjective experience data are collected.

[0080] In order to make data from different sources and different dimensions comparable, the system needs to standardize the raw scores after data collection. Preferably, the Z-score standardization method is used to process the data of each ability dimension, and the processing method is as follows:

[0081]

[0082] in,

[0083] represents the original score of the i-th student on the k-th ability dimension,

[0084] μ k is the mean value of the k-th ability dimension among all students,

[0085] σ k is the standard deviation of the dimension,

[0086] is the standardized score of the i-th student on the k-th dimension.

[0087] For each student i, the standardized scores on all d ability dimensions form a d-dimensional ability vector:

[0088]

[0089] The vector x i It is a quantitative expression of the student’s current ability status in the ability space. The vector set {x1,…,x N} constitutes the ability status dataset of the student group, where N represents the total number of students.

[0090] In order to model the ability status of the student group as a mathematical model and use it for path evolution calculation, the system constructs a discrete probability measure based on the above vector set, which is specifically expressed as:

[0091]

[0092] in:

[0093] μ0: The initial probability measure distribution of student ability, which is defined in the ability space Discrete measures on ;

[0094] x i : The ability vector of the i-th student is the standardized ability status;

[0095] Dirac measure, which at point x i The value at the position is 1, and the other positions are 0, which is used to represent the point distribution of students' ability status;

[0096] w i : The weight coefficient in the student ability measurement is preferably That is, equal weight distribution, or weighted processing can be performed based on actual task participation, data reliability, or student role;

[0097] N: represents the number of student samples participating in the modeling.

[0098] This initial measurement distribution μ0, as the mathematical modeling result of the capability state, has the following characteristics:

[0099] It is defined in the capability space Discrete measures on ;

[0100] Can reflect the distribution trend of the entire student group in various ability dimensions;

[0101] It can provide deterministic initial state input for subsequent capability evolution modeling.

[0102] During the system implementation process, the measurement modeling process can be implemented through the following process modules:

[0103] Data interface layer: responsible for connecting with the student affairs system, course platform, and survey system to collect original ability data;

[0104] Standardization module: performs mean-variance standardization on the original data and outputs the capability vector;

[0105] Ability vector aggregation module: aggregates all students' ability vectors and calculates measurement weights;

[0106] Measure generation engine: Construct μ0 as a discrete probability measure in the capability space and output it as the input interface of subsequent modules.

[0107] Through the above technical solutions, the system can realize structured modeling of the labor capacity status of college students and transform discrete individual capacity data into mathematical objects that can be used for analysis, evolution, and matching.

[0108] In summary, step S1: the implementation method of constructing the initial measurement distribution of students' abilities is based on technical links such as ability dimension definition, data collection, standardization processing, ability vector construction and measurement modeling, which systematically realizes the complete conversion from original student behavior and evaluation data to ability measurement model.

[0109] The implementation of this step not only provides key starting point data for the subsequent ability evolution path solving and task matching modules of the present invention, but also lays a modeling foundation for realizing the visual management and system feedback of students' ability development trajectory.

[0110] S2. Divide project learning into multiple stages and set corresponding target capability vectors for each stage to form a time series capability target distribution of labor tasks;

[0111] In this embodiment, the technical goal of step S2 is to complete the measurement expression of the target capability of the labor task, and provide a clear mathematical form for the target distribution in the subsequent capability evolution path model. By converting the capability structure required for each stage of the labor project into a discrete target capability distribution, the present invention can achieve the following goal in a unified capability space. A quantitative mapping between students' ability status and task ability requirements is established to provide a clear direction for ability evolution.

[0112] To achieve the above goals, this step is preferably performed by the "task goal modeling module" in the system, which has functions such as task stage splitting, target ability setting, and measurement weight allocation, and can cooperate with the project learning management system and the teacher input terminal.

[0113] During implementation, the labor projects must first be structured into phases.

[0114] Preferably, a labor project P can be divided into M time-continuous or logically progressive task stages {P1, P2, ..., P M Each stage corresponds to a specific task sub-goal. This stage division can be defined by the project course leader or automatically generated based on a preset project template.

[0115] Each phase task P j Corresponding to certain capability requirements, it is necessary to determine the ideal capability target vector y in the capability space j (t), whose structure is the same as the student ability vector x in step S1 i Consistent:

[0116]

[0117] in,

[0118] j is the task sub-stage number, satisfying j = 1, 2, ..., M;

[0119] d represents the number of ability dimensions, which is consistent with the dimension of the student's ability vector;

[0120] represents the target score of the jth task sub-stage on the kth ability dimension at time t.

[0121] The above capability target vector can be determined as follows:

[0122] Expert experience input: Teachers or project designers set stage competency requirements based on teaching objectives;

[0123] Template library matching: The system has several built-in typical task capability templates and recommends them based on task type and content;

[0124] Historical task analysis: extract capability distribution statistics from historical similar tasks and calculate the central capability vector;

[0125] Fusion generation mechanism: Combines expert input with data-driven models to dynamically adjust vector components and improve adaptability.

[0126] In order to further transform from single point vector to distribution modeling, in this embodiment, the capability target set {y1(t),…,y M (t)} is assigned a weight coefficient v j (t), constitutes the phased measurement representation of the mission capability target. This distribution is a discrete probability measure, specifically expressed as follows:

[0127]

[0128] in:

[0129] ν(t): represents the target capacity distribution of the labor task at time t;

[0130] Indicates that at vector point y j The Dirac measure at (t) indicates the concentration location of the capability target;

[0131] v j (t): is the measurement weight of the j-th task subunit at time t, preferably satisfying Form a normalized distribution;

[0132] M: The number of sub-goals divided in the task stage;

[0133] T: The total duration planned for the entire labor task.

[0134] The target distribution ν(t) serves as the target endpoint in the capability evolution path optimization model and needs to be in the same capability space as the initial capability measure μ0 in step S1, ensuring that the evolution path can be defined as In the dual of the measure space on .

[0135] During system implementation, the construction process of ν(t) may include the following module-level processing:

[0136] Task structure analysis module: parses the task text structure and determines the phase division;

[0137] Target capability generation module: generates target vectors by calling expert templates or historical data according to the stage task content;

[0138] Weight allocation module: Generate weight v based on factors such as task importance, completion time ratio, learning goal requirements, etc. j (t);

[0139] Measurement assembly module: combines the stage target capability set and weights to form a measurement structure ν(t) and performs time discrete management.

[0140] In summary, the method disclosed in step S2, based on a phased task structure, constructs a discrete target capability measurement distribution ν(t) unified within the capability space, providing a quantitative target for modeling capability evolution paths. At the implementation level, a combination of manual setting and data-driven approaches completes the process of capability vector generation, weight assignment, and measurement expression, ensuring the closed-loop operability and goal-driven nature of the proposed path planning model.

[0141] S3. Based on the initial measurement distribution of students’ abilities and the distribution of labor task targets, establish an optimal model for the ability evolution path;

[0142] In this embodiment, the core task of step S3 is to introduce continuity constraints in the process of modeling the student ability evolution path, so that the process of the probability measure describing the ability status of the student group evolving over time satisfies the conservation law, conforms to the evolutionary characteristics of actual ability growth, and constitutes one of the basic constraints of the ability path optimality solution model.

[0143] Under the capability space modeling framework proposed in this invention, the capability status of a student group at a certain time t∈[0,T] can be defined by a The probability measure μ on t The representation is called the capability state measurement flow. This measurement flow describes the distribution structure of group capabilities in the capability space at each moment.

[0144] At the same time, in order to describe the migration trend of the ability state over time, a time-dependent velocity field v is introduced t (x), that is:

[0145]

[0146] The velocity field represents a point in the capability space. The capability state of the system is the unit time change trend at time t, that is, the evolution direction and rate of the local capability vector.

[0147] To ensure the above measurement flow μ t The evolution process of is continuous and smooth in the mathematical sense, and meets the practical requirements of no sudden change or jump in the physical sense. It is necessary to introduce a continuity constraint equation to force the change of the measure over time to be driven by the velocity field and to meet the principle of mass conservation:

[0148]

[0149] in:

[0150] Represents the measure μ t The local partial derivative with time t reflects the evolution speed of the capability state in the time dimension;

[0151] v t (x): the speed of change of local ability at point x in the ability space at time t;

[0152] μ t (x): is the probability density function, which represents the probability density of the ability at position x;

[0153] is the divergence term, which represents the outflow trend of the capability state at that point;

[0154] This equation as a whole indicates that in the absence of capacity source terms or external injection / loss, the transfer of capacity follows the internal migration drive and the overall capacity quality of the system is conserved.

[0155] The above equation originates from the continuity modeling criterion for measure evolution in optimal transport theory and is one of the fundamental constraints of path planning problems.

[0156] In the present invention, the solution of the capability evolution path aims to minimize the action (path cost), and its mathematical model is defined in step S3 as follows:

[0157]

[0158] in:

[0159] The total cost function representing the capability evolution path;

[0160] ||v t (x)|| 2 : is the energy change of unit capacity mass in unit time;

[0161] dμ t (x): represents the capacity density contribution at position x.

[0162] The above optimization objectives need to be solved under continuity constraints, namely:

[0163] μ0 is known, μ T ≈ν T

[0164] At the technical implementation level, in order to make the continuity equation applicable to numerical solutions, the present invention discretizes the time axis [0, T] to form an equidistant or adaptive time node sequence:

[0165] {t0=0,t1,t2,…,t L =T},Δt=t k+1 -t k ;

[0166] At each discrete time step t k , capability measure With velocity field are represented as discrete distributions and vector field sample points. The system implements the modeling of continuity constraints in discrete form through the following strategies:

[0167] Approximate the time derivative term using forward differencing:

[0168]

[0169] Use the grid to discretize the power space and calculate the divergence at each grid node And make corrections to each measure update item;

[0170] If the capability measure is expressed in particle form (i.e. ), we can use the particle method to update the particle position x i (t) is used to measure the flow, and the velocity v t It is determined by the optimal transport vector between particles.

[0171] In addition, to improve the stability and efficiency of the numerical solution, a dual problem structure can be introduced in the path solution, and the following joint optimization strategy can be adopted:

[0172] Construct Lagrangian:

[0173]

[0174] Where ψ(t,x) is the Lagrange multiplier, which represents the dual variable of the constraint function;

[0175] Solve the saddle point problem of the Lagrangian to ensure the dual conditions of cost minimization and constraint satisfaction;

[0176] The model can be solved using variational optimization methods such as the Sinkhorn iterative algorithm, the alternating direction method of multipliers (ADMM), and the Proximal operator method.

[0177] At the system level, the continuity modeling involved in this step is encapsulated as a “path modeling and constraint control module”, which is responsible for maintaining the dynamic consistency of the capability measurement state during the evolutionary solution process, supporting the transition from the initial measurement μ0 to the target distribution v t and provides a basis for state update for each discrete time step.

[0178] Through the constraint management of this module, it can be ensured that during the path evolution process, the changes in students' ability status are driven by internal structures, without mutations or jumps, which is in line with the realistic characteristics of "gradual growth of ability" in educational scenarios.

[0179] In summary, step S3 achieves systematic guarantees for the physical consistency, mathematical solvability and system feasibility of the capability evolution path through the complete construction, mathematical definition, parameter normalization and comprehensive disclosure of the numerical solution form of the continuity equation. It is one of the core constraints in the optimal path modeling of the present invention, and constitutes a unified optimization structure with the path cost function, jointly supporting the implementation basis of the subsequent capability dynamic matching and feedback correction module.

[0180] S4. Based on the ability evolution path optimality model, the student's current ability status is optimally matched with the ability requirements of the corresponding stage task, and a dynamic mapping relationship between students and tasks is output;

[0181] In this embodiment, the core of step S4 is to dynamically achieve the optimal matching relationship between students and staged labor tasks based on the solved student ability evolution path. This step inherits the student initial ability measurement μ0, task ability target distribution ν(t), evolution path model established in steps S1 to S3. and its continuity constraints, aiming to organically connect the capacity improvement process with the project task execution link.

[0182] This step constructs an optimal matching problem between the distribution of student ability states and the distribution of task objectives, outputting a mapping solution between student groups and corresponding task sets at each time point. This matching solution not only performs static allocation based on ability structure similarity but also introduces a regularized optimal transportation model to minimize the cost of ability migration and resource allocation while meeting task requirements, thereby improving the consistency and systematic nature of path execution.

[0183] In terms of modeling, this embodiment uses each discrete time point t l The capability state and task goal on ∈[0,T] are expressed as probability measures:

[0184] The status measurement of the student group in the ability space, x i (t l ) represents the i-th student at time t l The capability state vector, w i is the measurement weight (preferably w i =1 / N);

[0185] The target capability measure of the task set in the capability space, y j (t l ) is the jth task at time t l The target capability vector, v j (t l ) is the weight of the task stage.

[0186] In order to establish a reasonable matching strategy between the above two measures, an optimal transportation model with regularization term is introduced. Specifically, at each time node t l , the optimal matching mapping between students and tasks is obtained by solving the regularized Wasserstein distance of the following form:

[0187]

[0188] in,

[0189] γ is the joint transition probability distribution between μ and ν, which is used to describe the mapping relationship from student ability status x to task ability requirement y;

[0190] Γ(μ,ν): is the joint measure set of all edges μ and ν respectively;

[0191] ||xy|| 2 : represents the Euclidean square distance between the student's ability status and the task goal, as the matching cost;

[0192] ε: Regularization coefficient, used to control the influence of the entropy regularization term;

[0193] KL(·||·): represents the Kullback-Leibler divergence, which is used to measure the degree of information deviation between the current matching strategy and the independent matching strategy;

[0194] represents an independent joint measure between the student and the task.

[0195] The matching model aims to find a set of optimal mapping strategies to minimize the total cost of migrating students' ability states to task target states, while avoiding the matching strategies from being too centralized or deviating from the distribution structure.

[0196] In terms of system implementation, the matching problem is transformed into a solvable discrete optimization problem. The specific process is as follows:

[0197] Constructing a student competency status matrix and the Mission Capability Objective Matrix

[0198] Constructing a cost matrix The i,j items are C ij =||x i -y j || 2 , represents the Euclidean square distance between the student and the task;

[0199] The Sinkhorn algorithm is used to optimize the following objective function:

[0200]

[0201] in,<C,γ> =∑ i,j C ij γ ij is the sum of matching costs, H(γ)=∑ i,j γ ij (logγ ij -1) is the negative entropy term.

[0202] Using the Sinkhorn-Knopp iterative method, the pairing matrix γ is iteratively updated to satisfy:

[0203] γ (k+1) =diag(u (k) )Kdiag(v (k) );

[0204] Where K = exp(-C / ε), u, v are normalized vectors, and the iteration is done until the edge converges to the μ, ν distribution.

[0205] The final output γ *Represents the optimal matching relationship between students and tasks, where each non-zero element It can be interpreted as the probability or weight of the assignment between student i and task j.

[0206] After the matching is completed, the system will extract the maximum weight pairing or generate a student task assignment list based on the matching threshold. l superior

[0207] Student task allocation function:

[0208]

[0209] in Indicates that the i-th student at time t l The assigned task objectives. This mapping will be stored by the system as a task execution plan and used to generate a complete student task sequence and growth trajectory in step S5.

[0210] Preferably, the system can also perform extended matching for the following additional constraints:

[0211] Task capacity constraint: Set the maximum number of students that can be matched for task j Enforce satisfaction in matching;

[0212] Student Workload Balancing: Towards γ ij Add constraints on the total amount of student work;

[0213] Stage dependency constraints: Manage the time sequence of task assignments to avoid skipping stages or violating project processes;

[0214] Ability domain relevance screening: limit tasks to recruit students only in certain ability space areas to improve matching fit.

[0215] The above constraints can be solved by introducing weight penalty terms, constraint functions or penalty multiplier mechanisms into the objective function for joint optimization.

[0216] In the system architecture, the matching process is centrally scheduled by the "Dynamic Matching Scheduling Module", whose main functions include:

[0217] Get path model (μ t ,v t ) Output the current capability measurement status;

[0218] Call the ν(t) distribution provided by the "Mission Capability Target Modeling Module";

[0219] Build matching models and solvers to perform regularized optimal matching;

[0220] Output matching matrix γ and task mapping function Φ t, for downstream modules to call.

[0221] The matching results are written into the "task allocation database" in the form of structured data and will be used in subsequent processes such as task scheduling, evaluation feedback, and path correction.

[0222] In summary, step S4 achieves a dynamic mapping between students and tasks driven by the ability evolution path by constructing an optimal matching model based on the regularized Wasserstein distance. This step is theoretically based on the principle of optimal transportation, relies on efficient numerical optimization algorithms, and forms an automated matching mechanism within the system architecture. It is a key technical step in achieving personalized and precise labor capacity development.

[0223] S5. Based on the student-task matching results, generate each student's task execution sequence and ability growth trajectory;

[0224] In this embodiment, step S5 generates a task execution sequence based on the optimal student-task match output from step S4. It then tracks students' task execution to establish their competence growth trajectory. This step aims to provide students with a clear competence development path and provide teachers with real-time feedback to facilitate necessary adjustments and optimizations.

[0225] In this step, the system first records the task assignments of each student at each time node based on the student-task matching results in step S4. The task execution sequence of each student consists of the tasks performed at each stage, which can be expressed as:

[0226] T i =[t i,1 ,t i,2 ,…,t i,K ];

[0227] in:

[0228] T i represents the task execution sequence of the i-th student;

[0229] t i,k represents the task performed by the i-th student in the k-th stage. Each task t i,k They are all obtained through the optimal matching algorithm in step S4 and are adapted to the ability requirements of the task and the ability status of the students.

[0230] The process of generating a task sequence includes the following steps:

[0231] Record task matching results: Output γ according to the optimal matching in step S4 *The system records the task assignments of each student at each stage and generates task execution records. The task assignments of each student will be optimized based on their current ability status and the required ability of the task.

[0232] Generate task execution sequence: By analyzing the matching results, the system generates a task execution sequence T for each student i ,This sequence indicates the task arrangement that students should participate in throughout the training process. The order of each task is dynamically adjusted according to the students' ability improvement and task requirements.

[0233] Ensure that task allocation is consistent with both students' ability development and the sequence of tasks.

[0234] Organize and archive task lists: Once the student's task sequence is determined, the system will organize all students' task execution status into a complete task list and archive it according to the order of the task stages. This task list will become an important reference for teachers' teaching and project management.

[0235] Provide task arrangements for teachers: The organized task list can be viewed and adjusted by teachers in real time to ensure the orderly progress of teaching activities and adjust the arrangement of subsequent tasks according to changes in students' abilities.

[0236] The growth trajectory is a path that describes the changes in students' abilities at each stage. It records the ability status of students at each stage and provides teachers and the system with visual feedback on their ability improvement. In this step, the system constructs each student's ability growth trajectory G based on the task execution status and student ability feedback. i , expressed as:

[0237] G i =[x i (t1),x i (t2),…,x i (t T )];

[0238] in:

[0239] G i is the growth trajectory of the i-th student, describing the student at time nodes t1, t2,…, t T the evolution of capabilities in the field;

[0240] x i (t l ) represents the ability state vector of the i-th student at the l-th time node. The ability state is adjusted by the tasks performed by the student at that stage and the feedback of the tasks.

[0241] The calculation formula for capability status is:

[0242]

[0243] in:

[0244] x i (t l ) is the student at time t l The ability status of the above;

[0245] T i (t l ) is the i-th student at time t l the tasks performed;

[0246] At time t l The initial distribution of students’ abilities may be input to the system or updated based on the feedback of the aforementioned ability evolution function;

[0247] f(·) is the ability evolution function, which represents the transformation pattern of the student's ability status. This function integrates factors such as skill accumulation after task completion, learning feedback, and student performance in the task.

[0248] Growth Trajectory G i The generation of can be achieved through the following steps:

[0249] The impact of task execution on competency status: After completing each stage of a task, a student's competency status changes. This change depends not only on the difficulty and requirements of the task but also on factors such as student feedback, adjustments, and learning progress. Using the competency evolution function f(·), the system updates each student's competency status based on their task execution results and competency improvement path.

[0250] Continuous tracking of capability evolution path: At each time point t l , the system will record the student's ability status x i (t l ) and add it to the student's growth trajectory G i The changes in students’ abilities at different time points can help teachers analyze their growth trajectory and adjust task arrangements based on that path.

[0251] Assessment of ability change trends: By analyzing students' growth trajectories, teachers can intuitively see the trend of students' ability changes at each stage, determine whether students are improving their abilities according to the predetermined path, and whether further adjustments are needed to their task execution to achieve the best learning results.

[0252] After generating the task sequence and growth trajectory, the system will be able to:

[0253] Provide teachers with visual feedback: Through student growth trajectories, teachers can intuitively see the changing trends of students' abilities at different stages, thereby adjusting and optimizing teaching content;

[0254] Provide students with personalized ability improvement plans: Based on the students' task execution and ability evolution path, the system can provide students with personalized learning plans to ensure that each student can achieve the greatest degree of ability improvement during the project learning process;

[0255] Optimize subsequent task arrangements: Based on the students' growth trajectory, the system can dynamically adjust task allocation in subsequent stages to ensure that students' tasks at different stages are of reasonable difficulty and challenge.

[0256] Step S5 generates task sequences and growth trajectories, ensuring that students' task execution processes align with their competency development paths. This provides real-time competency development feedback and a basis for optimizing subsequent teaching and learning paths. Through precise task matching and competency tracking, the system can provide students with personalized competency development plans and provide teachers with effective competency assessment tools. This process fully discloses the generation mechanisms of task sequences and growth trajectories and ensures that the details of each technical feature are fully disclosed, providing clear guidance for implementation and subsequent adjustments.

[0257] S6. Collect students’ task completion status and ability growth trajectory in real time, dynamically correct students’ ability measurement distribution, and thus update subsequent ability evolution paths.

[0258] In this embodiment, the goal of step S6 is to modify the student's ability path based on the ability assessment feedback after task completion. By collecting the student's ability assessment data after different task stages and combining it with the ability status predicted by the system model, the system can dynamically adjust the student's ability measurement distribution, thereby optimizing the student's learning path.

[0259] The key to this step is to correct the ability distribution through real-time feedback so that the student's ability measurement can reflect his or her actual ability changes throughout the learning process, ensuring that the ability improvement path is always in line with the student's actual development level.

[0260] In this embodiment, the system first collects the ability assessment feedback data after the task is completed in the following ways:

[0261] Task performance feedback: students’ specific performance in completing tasks, including the quality, accuracy, time efficiency, etc.

[0262] Teacher evaluation: A subjective rating given by teachers based on students’ task completion, learning attitude, and participation;

[0263] Self-evaluation: students’ self-evaluation of the results of task execution based on their personal perception;

[0264] Peer assessment: Other students or team members evaluate the student's performance in performing the task.

[0265] These data provide the system with multi-dimensional feedback information, reflecting the students' actual performance during task execution, and further enhancing the authenticity and accuracy of ability measurement.

[0266] Based on the feedback data after the task is completed, the system constructs the observed distribution of students' abilities This distribution is based on the actual performance of the task and its feedback results to correct the student’s ability. Specifically, the observation distribution reflects the student’s l The ability status of the task is taken into account, taking into account the actual effect of task execution and evaluation feedback.

[0267] The observed distribution of capability can be expressed as follows:

[0268]

[0269] in:

[0270] Indicates that at time node t l Revised ability distribution;

[0271] is the preliminary capacity distribution predicted by the system model;

[0272] f feedback (T i ,y i ) is a correction item generated based on the task execution results and feedback data, which represents the students’ actual ability feedback at this task stage;

[0273] g(·) is a correction function, which is used to integrate feedback data into the ability distribution and correct the original ability measure.

[0274] Through this correction, the system can generate a capability distribution based on actual task performance and feedback. More accurately reflect the students' ability status.

[0275] After constructing the observed distribution of student abilities, the system will update the original ability distribution based on feedback to generate a revised ability measurement distribution μ new (t l This correction achieves dynamic updating of the capability path by weighted fusion of the original predicted capability distribution and the new feedback observation distribution.

[0276] The revised capability distribution formula is expressed as:

[0277]

[0278] in, is the distribution of model prediction capabilities, is the feedback capability measure, and λ∈[0,1] is the fusion coefficient.

[0279] The value of λ is dynamically adjusted based on the feedback from different tasks. Typically, λ fluctuates within a reasonable range to reflect the distribution of model prediction capabilities and the accuracy of feedback data. When feedback data is reliable, λ will be smaller, indicating greater reliance on feedback data. When feedback data is unstable or incomplete, λ will be larger, indicating greater reliance on model predictions.

[0280] By updating the capability measure based on feedback The system can dynamically adjust students' ability paths. This ability path will serve as the basis for subsequent task allocation and ability improvement, ensuring that each student's task execution and ability development plan match their actual ability development.

[0281] Specifically, the revised capability measures will be used to:

[0282] Dynamically update capacity forecast: capacity distribution at each time node It will become the basis for task matching in subsequent stages, ensuring that students’ task assignments in future stages accurately reflect their ability status.

[0283] Personalized task assignment: By dynamically adjusting the competency path, the system can assign appropriate tasks to students based on their current competency profile. These tasks will help students further improve their abilities and ensure the continuity and stability of their learning process.

[0284] Teachers can adjust teaching plans in real time: Based on the revised competency path, teachers can adjust teaching content and task requirements to ensure that students receive appropriate challenges and support at each stage.

[0285] Step S6 dynamically adjusts the student's competency path by continuously collecting competency feedback after task completion and combining it with the competency observation distribution generated based on task performance and assessment feedback. By updating and integrating competency measures, the system accurately tracks each student's competency improvement process and provides personalized task assignments and learning paths at each stage. This dynamic adjustment mechanism ensures more precise and flexible student competency development and provides teachers with a powerful real-time feedback tool.

[0286] Through this feedback and correction mechanism of ability paths, the present invention can effectively cope with the complexity of changes in students' abilities during the education process and provide dynamic and personalized learning support solutions for education administrators, teachers and students.

[0287] The system for constructing a path for cultivating the labor capacity of college students based on project learning described below and the method for constructing a path for cultivating the labor capacity of college students based on project learning described above can be used in correspondence with each other.

[0288] See also Figure 2 The present invention also provides a system for constructing a path for cultivating the labor capacity of college students based on project learning, including:

[0289] Ability measurement modeling module, used to collect student ability data and construct initial probability measures of the ability space;

[0290] Mission objective modeling module, used to analyze project structure and construct phased mission capability objective distribution;

[0291] Path modeling and optimization module, used to establish the optimal model of capability evolution path and solve it based on minimum action;

[0292] Numerical calculation and matching module, used to perform path optimization and dynamically match students and tasks;

[0293] Growth path output module, used to generate task sequences and capability growth trajectories;

[0294] Path feedback module, used to collect task feedback and update capability status;

[0295] Visual display module, used to output student growth reports and graphic path visualization information.

[0296] The device of this embodiment can be used to execute the above method embodiment, and its principles and technical effects are similar, so they will not be repeated here.

[0297] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A method for constructing a path for cultivating college students' labor capacity based on project learning, characterized by: The following steps are involved: S1. Collect the original ability data of college students and standardize it to form ability vectors, construct the initial measurement distribution of students' abilities, and provide basic ability status information for subsequent task matching; S2. Divide project learning into multiple stages and set corresponding target capability vectors for each stage to form a time series capability target distribution of labor tasks; S3. Based on the initial measurement distribution of students’ abilities and the distribution of labor task targets, establish an optimal model for the ability evolution path; S4. Based on the ability evolution path optimality model, the student's current ability status is optimally matched with the ability requirements of the corresponding stage task, and a dynamic mapping relationship between students and tasks is output; S5. Based on the student-task matching results, generate each student's task execution sequence and ability growth trajectory; S6. Collect students’ task completion status and ability growth trajectory in real time, dynamically correct students’ ability measurement distribution, and thus update subsequent ability evolution paths.

2. The method for constructing a path for cultivating college students' labor ability based on project learning according to claim 1 is characterized in that: The steps for constructing the initial measurement of student abilities include: Collect students' original scores on multiple ability dimensions; Standardize the scores and construct the ability vector; The capability measure distribution is constructed based on the capability vector, and the form is: Among them, μ0 is the initial measurement of the student group's ability, x i is the ability vector of the i-th student, w i is the weight, is the Dirac measure.

3. The method for constructing a path for cultivating college students' labor ability based on project learning according to claim 1 is characterized in that: Constructing the labor task time series capability target distribution includes: Divide the labor project into successive phases; Set target capability vectors and establish metrics for each phase of the task; Generate target capability distribution time series, expressed as: Among them, y j (t) is the j-th task target capability vector, v j (t) is its distribution weight, M is the total number of task subunits, and T is the total task time.

4. The method for constructing a path for cultivating college students' labor ability based on project learning according to claim 1 is characterized in that: The capability evolution path model construction includes: Define the continuous time measure flow μ of student ability evolution t ; Constructing capability path cost function The form is: Set the path evolution constraint to the continuity equation: Where: μ t is the distribution of student ability measurement at time t, v t (x) is the velocity vector field of the student’s ability at position x, It is a d-dimensional ability space.

5. The method for constructing a path for cultivating college students' labor ability based on project learning according to claim 4 is characterized in that: The optimal path problem is solved by discretizing the time domain and using a variational method combined with the Lagrange multiplier method, ensuring that the path evolution satisfies the continuity constraint.

6. The method for constructing a path for cultivating college students' labor ability based on project learning according to claim 1 is characterized in that: The optimal match includes: Time discretization of capability evolution path; The Sinkhorn iterative algorithm is used to achieve the optimal matching between students and tasks; The matching metric is the regularized Wasserstein distance: where μ and ν are the current student ability distribution and task goal distribution, respectively; γ is the joint transition probability distribution between μ and ν; ε is the regularization coefficient; KL(·) is the Kullback-Leibler divergence; and Γ(μ,ν) is the set of joint measures where all edges are μ and ν.

7. The method for constructing a path for cultivating college students' labor ability based on project learning according to claim 1 is characterized in that: The task execution sequence generation step includes: Record each student's task matching results; Generate task execution sequence in, represents the matching task of the i-th student in the k-th stage; Organize task lists and archive task schedules for system and teacher reference.

8. The method for constructing a path for cultivating college students' labor ability based on project learning according to claim 1 is characterized in that: The capability growth trajectory is represented as a path in the capability space: in, is the ability status of the i-th student at the k-th time node. This trajectory is used to evaluate the changing trend of students' abilities at different stages.

9. The method for constructing a path for cultivating college students' labor ability based on project learning according to claim 1 is characterized in that: Dynamically modifying the capability measure distribution includes the following sub-steps: Collect capability assessment feedback data after task completion; Constructing an observed distribution of student abilities Update the capability distribution based on the feedback to generate a revised capability measure: in, is the distribution of model prediction capabilities, is the feedback capability measure, and λ∈[0,1] is the fusion coefficient.

10. A system for constructing a path for cultivating the labor capacity of college students based on project learning, which is used to execute the method for constructing a path for cultivating the labor capacity of college students based on project learning as described in any one of claims 1 to 9, characterized in that: include: Ability measurement modeling module, used to collect student ability data and construct initial probability measures of the ability space; Mission objective modeling module, used to analyze project structure and construct phased mission capability objective distribution; Path modeling and optimization module, used to establish the optimal model of capability evolution path and solve it based on minimum action; Numerical calculation and matching module, used to perform path optimization and dynamically match students and tasks; Growth path output module, used to generate task sequences and capability growth trajectories; Path feedback module, used to collect task feedback and update capability status; Visual display module, used to output student growth reports and graphic path visualization information.

Images