A virtual simulation technology driven integration of production and teaching teaching system

Abstract

The present application relates to the field of educational information technology, in particular to a virtual simulation technology driven production and teaching integration teaching system, acquires behavior sequence and calculates correlation degree, extracts high-frequency behavior chain node, marks failure stage to generate identification result, optimizes semantic scene configuration, reconstructs task execution time sequence, adjusts access permission level, and generates integration scheduling path result. The present application extracts high-frequency behavior chain by constructing behavior sequence, clarifies task path key node order, and enhances behavior data identification accuracy; adjusts node display strategy based on display frequency and failure proportion, improves task content and scene semantic matching degree; rearranges task node order according to adaptation degree, optimizes execution path combined with role time period and task priority; dynamically adjusts access level through permission difference matrix, and realizes permission allocation refinement.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of educational informatization, and in particular to a virtual simulation technology driven production and teaching integration teaching system. BACKGROUND

[0002] The technical field of educational informatization includes a systematic method of utilizing digital technology and deep integration with educational scenarios to realize optimized allocation of teaching resources and intelligent management of teaching processes. The content involves constructing a virtual and real interactive teaching environment through virtual simulation, cloud computing, and Internet of Things technology, integrating industrial practice data and educational theory models, and solving problems such as lagging teaching content, insufficient practical resources, and disconnection between skill training and job requirements caused by traditional separation of production and teaching. It covers digital modeling of teaching scenarios, real-time synchronization of multi-source data, design of cross-platform collaboration mechanisms, and establishment of dynamic feedback evaluation systems.

[0003] Among them, a virtual simulation technology driven production and teaching integration teaching system refers to constructing a high-precision industrial scene model based on a virtual simulation engine, converting enterprise production processes, equipment operation specifications, and job skill requirements into interactive teaching units, realizing dynamic adaptation of teaching resources and industrial standards through data-driven models, and completing real-time operation feedback between teachers, students, enterprise personnel, and simulation environments through collaborative interaction interfaces. The main technical means include: two-way mapping modeling of teaching scenarios and industrial environments, teaching behavior data collection based on standardized protocols, logic arrangement and conflict resolution mechanism of multi-role collaboration tasks, and real-time operation verification and path optimization based on rule engines.

[0004] In the prior art, behavior data processing is limited to static recording, lacking extraction of task path evolution rules, affecting the effectiveness of dynamic adjustment of tasks; task configuration relies on preset rules, unable to adjust the display strategy according to failed nodes, reducing the relevance of teaching guidance; path scheduling does not consider the matching of adaptation degree and time resources, often resulting in task accumulation or resource waste; the lack of dynamic mechanism for permission allocation causes a disconnection between task permission configuration and actual participation; node scheduling ignores the degree of role period overlap, making collaborative tasks prone to conflict, limiting the smooth execution of teaching tasks and resource allocation efficiency. SUMMARY

[0005] The purpose of the present application is to solve the shortcomings in the prior art and to propose a virtual simulation technology driven production and teaching integration teaching system and method.

[0006] In order to achieve the above-mentioned purpose, the present application adopts the following technical solution: a virtual simulation technology driven production and teaching integration teaching system, the system comprising:

[0007] The behavior recognition module obtains interaction behavior data of students and industry roles, constructs a behavior sequence according to teaching task numbers and time sequences, calculates a behavior correlation degree according to operation time differences, counts a successful number of teaching tasks according to feedback information, classifies the behavior sequence and calculates an occurrence frequency, extracts a behavior node order in a behavior chain with a frequency exceeding a sequence average, marks a teaching task stage corresponding to a continuous failure of the behavior sequence, and generates a behavior pattern recognition result;

[0008] The semantic scene optimization module maps teaching tasks and semantic labels based on semantic content of task nodes in the behavior pattern recognition result, reads a total display time and a number of times of each task node in the behavior sequence and calculates an average value, filters a node set with an average display number of times lower than a median value in the task nodes, counts a failure number of times of students in a corresponding behavior path and calculates a failure number ratio, reallocates a display time of a task node with a failure ratio exceeding an average failure ratio and adjusts an order of the task node, and generates a semantic scene configuration optimization result;

[0009] The time sequence scheduling module calls a total execution time and a total completion number of time of a task node interaction record in the semantic scene configuration optimization result, calculates an average execution time and a completion ratio according to node grouping, sorts the task nodes based on the calculation result, extracts nodes with an adaptation degree exceeding a set adaptation degree threshold, and cross-compares a role available time period and a task priority, and generates a task execution time sequence reconstruction result;

[0010] The permission ruling module reads an access permission level of a role corresponding to each task node and a task permission requirement of the task node in the task execution time sequence reconstruction result, constructs a role permission difference matrix and counts a permission satisfaction number of times of the role in the task, extracts a role set with a permission satisfaction number of times lower than a median value in a task total number, obtains a task completion rate of the role set in all task records, sorts and adjusts the access permission level of the role according to the completion rate, and forms a permission ruling result;

[0011] The production and teaching integration task scheduling module re-reads a role available time period and a teaching task priority according to the role with an adjusted permission in the permission ruling result, sorts a task node adaptation degree in the teaching task and constructs a task priority execution sequence, extracts a task node combination with an available time period overlap degree exceeding an overlap degree threshold between the roles in the priority sequence, rearranges the integration scheduling path according to the time sequence, and outputs a production and teaching integration task scheduling result.

[0012] As a further aspect of the present invention, the behavior pattern recognition result includes a behavior number index set, a task success frequency classification table, and behavior sequence clustering labels; the semantic scene configuration optimization result includes a task semantic mapping table, a node display duration adjustment item, and a node execution order reordering item; the task execution time sequence reconstruction result includes a node adaptability sorting table, a task priority matching item, and a role time period cross matrix; the permission adjudication result includes a permission level adjustment item, a role access permission leveling table, and a permission satisfaction frequency sorting table; the industry-education integration task scheduling result includes a node combination reordering sequence, a role available time period integration table, and an integration task scheduling path set.

[0013] As a further aspect of the present invention, the behavior recognition module includes:

[0014] The behavior collection submodule obtains interaction data between students and industry roles based on operation logs, records behaviors and extracts operation times according to task number and time sequence, calculates the time difference between behaviors and filters out out-of-limit behaviors, and obtains an interaction behavior interval dataset.

[0015] The sequence construction submodule clusters behavior records based on the interaction behavior interval dataset, counts the number of successful tasks for each task and calculates the frequency, filters behavior sequences with frequencies exceeding the mean, extracts the order of behavior nodes, and generates a high-frequency behavior node sorting sequence.

[0016] The behavior chain recognition submodule calls the high-frequency behavior node sorting sequence to identify the consecutively failed task stages, calculates the behavior node sorting offset and combines it with the failure count ratio, using the formula:

[0017] ;

[0018] Obtain the offset of each failure stage, establish a mapping between stages and nodes, and obtain the behavior pattern recognition results;

[0019] in, This represents the offset value of the behavior chain. Representing the The position of a node in a high-frequency sort. Represents the average order of nodes. Represents the time difference between adjacent nodes. Represents the duration of the failure phase. This represents the node latency value.

[0020] As a further aspect of the present invention, the semantic scene optimization module includes:

[0021] The node data extraction submodule calculates the average number of times the task nodes are displayed based on the total display time and number of times in the behavior pattern recognition results, obtains the median value of the number of times the nodes are displayed, determines whether the average number of times each node is displayed is lower than the median value, filters out nodes that meet the conditions, and generates a set of low-frequency nodes.

[0022] The failure distribution calculation submodule calls the number of failures of each node in the low-frequency node set, calculates its failure ratio, obtains the average failure ratio of all nodes, determines whether the node failure ratio is higher than the average, and generates a list of failure nodes with higher than average ratio.

[0023] The display duration adjustment submodule calls the above-average failure node ratio list and the original display duration, and uses the following formula based on the failure ratio, number of displays, skip delay time, and time difference between nodes:

[0024] ;

[0025] The display duration adjustment value of the node is obtained through calculation. The order of the task nodes is rearranged according to the adjusted display duration to generate semantic scene configuration optimization results.

[0026] in, Indicates the first The adjusted display duration for each task node. This is the original display duration of the node. For the first The failure rate of each node, The average failure rate of all nodes. For the number of times a node is accessed. The time interval between task nodes. This represents the delay caused by a node being skipped.

[0027] As a further aspect of the present invention, the timing scheduling module includes:

[0028] The execution efficiency extraction submodule calculates the total number of task completions and the total execution time based on the interaction records of each task node in the semantic scenario configuration optimization results, according to the task node number, and calculates the average execution time and completion ratio respectively to obtain the node average execution time value and node completion ratio value.

[0029] The adaptability calculation submodule calls the average execution time of the nodes and the node completion ratio, and combines the failure probability, execution fluctuation, and task complexity indicators to calculate the following formula:

[0030] ;

[0031] The task node fit value of the node is obtained by calculation, and then the fit value is compared with a set threshold to filter and generate a set of nodes with high fit.

[0032] wherein, represents the adaptation value of the th task node, is the node completion rate, is the failure probability, is the average execution time of all nodes, is the average execution time of the node , is the execution time variance of the node , is the relative task complexity;

[0033] The task timing generation submodule calls the high adaptation node set, and generates a task execution timing reconstruction result according to the task priority and the role available time period intersection comparison, and rearranges the node order according to the time window.

[0034] As a further scheme of the present application, the permission decision module comprises:

[0035] The role permission reading submodule obtains the role access permission level and the task node permission requirement of each task node according to the task execution timing reconstruction result, and generates a role access permission record;

[0036] The permission difference matrix construction submodule constructs a role permission difference matrix based on the role access permission record, and uses the formula:

[0037] ;

[0038] calculates the permission difference of each role at different task nodes, and generates a permission difference matrix;

[0039] wherein, represents the permission difference between the role and the task node , represents the access permission level of the role at the task node , represents the permission requirement of the task node in the task , is the total number of tasks;

[0040] The permission adjustment submodule extracts a role set whose permission satisfaction times are lower than the median value of the task total number in the permission satisfaction times matrix, obtains the completion rate of the role set in the task, adjusts the role access permission level, and generates a permission decision result.

[0041] As a further scheme of the present application, the production-education integration task scheduling module comprises:

[0042] The role information reading submodule reads the role available time period and the teaching task priority according to the adjusted role in the permission decision result, obtains the available time period information of each role in a specific task, and sorts according to the priority of the teaching task to generate a role available time period record;

[0043] The task node adaptation degree sorting submodule calculates the adaptation degree of the role node based on the role available time period record and the teaching task priority, evaluates the matching degree between the role and the task node and sorts, and generates a task node adaptation degree sorting;

[0044] The scheduling path construction submodule extracts the task node combination in which the available time period overlap degree between roles in the task node adaptation degree sorting exceeds the overlap degree threshold, rearranges according to the time sequence, and uses the formula:

[0045] ;

[0046] The operation obtains the scheduling path rearrangement, forms the production-education integration scheduling path, and generates the production-education integration task scheduling result;

[0047] Wherein, represents the total length of the overlapping period of the task node , represents the available time period length of the role in the task node , represents the task execution time of the role in the task node , is the number of roles participating in the task.

[0048] Compared with the prior art, the advantages and positive effects of the present application are:

[0049] In the present application, the high-frequency behavior chain is extracted by constructing the behavior sequence, the key node sequence of the task path is determined, and the behavior data recognition accuracy is enhanced; the node display strategy is adjusted based on the display frequency and the failure proportion, the matching degree of the task content and the scene semantics is improved; the task node order is rearranged according to the adaptation degree, and the execution path is optimized in combination with the role time period and the task priority; the access level is dynamically adjusted through the permission difference matrix, and the permission allocation is refined; the time sequence of the high-overlap node combination is rearranged, the task cooperation efficiency and the resource utilization rate are improved, and the task scheduling optimization mechanism with a closed loop structure is constructed. BRIEF DESCRIPTION OF DRAWINGS

[0050] Figure 1 is the system flowchart of the present application;

[0051] Figure 2 is the acquisition flowchart of the behavior recognition module of the present application;

[0052] Figure 3 The acquisition flowchart of the semantic scene optimization module of the present application is shown in the figure.

[0053] Figure 4 The acquisition flowchart of the timing scheduling module of the present application is shown in the figure.

[0054] Figure 5 The acquisition flowchart of the permission arbitration module of the present application is shown in the figure.

[0055] Figure 6 The acquisition flowchart of the production and education integration task scheduling module of the present application is shown in the figure. DETAILED DESCRIPTION

[0056] The technical solutions in the present application will be described below with reference to the accompanying drawings.

[0057] In the embodiments of the present application, the words such as "example", "for example" are used to represent as an example, illustration or description. Any embodiment or design scheme described as "example" in the present application should not be interpreted as more preferred or more advantageous than other embodiments or design schemes. In fact, the word "example" is intended to present the concept in a specific way. In addition, in the embodiments of the present application, the meaning expressed by "and / or" can be both, or can be one of the two options.

[0058] In the embodiments of the present application, "image" and "picture" can be used interchangeably at times. It should be pointed out that when the distinction is not emphasized, the meanings expressed are consistent. "Of", "corresponding" and "corresponding" can be used interchangeably at times. It should be pointed out that when the distinction is not emphasized, the meanings expressed are consistent.

[0059] In the embodiments of the present application, sometimes the subscript such as W1 can be written in the form of non-subscript such as W1. When the distinction is not emphasized, the meanings expressed are consistent.

[0060] In order to make the technical problems, technical solutions and advantages of the present application clearer, specific embodiments will be described in detail below with reference to the accompanying drawings.

[0061] Please refer to Figure 1 The present application provides a technical solution: a production and education integration teaching system driven by virtual simulation technology, which comprises:

[0062] The behavior recognition module obtains interaction behavior data of students and industry roles, constructs a behavior sequence according to teaching task numbers and time sequences, calculates a behavior correlation degree according to operation time differences, counts a successful number of teaching tasks according to feedback information, classifies the behavior sequence and calculates an occurrence frequency, extracts a behavior node order in a behavior chain with a frequency exceeding a sequence average, marks a teaching task stage corresponding to a continuous failure of the behavior sequence, and generates a behavior pattern recognition result;

[0063] The semantic scene optimization module maps teaching tasks and semantic labels based on semantic content of task nodes in the behavior pattern recognition result, reads a total display time and a number of times of each task node in the behavior sequence and calculates an average value, filters a node set with an average display number of times lower than a median value in the task nodes, counts a failure number of times of students in a corresponding behavior path and calculates a failure number ratio, reallocates a display time of a task node with a failure ratio exceeding an average failure ratio and adjusts an order of the task node, and generates a semantic scene configuration optimization result;

[0064] The time sequence scheduling module calls a total execution time and a total completion number of times of interaction records of task nodes in the semantic scene configuration optimization result, calculates an average execution time and a completion ratio according to node grouping, sorts the task nodes based on the calculation result, extracts nodes with an adaptation degree exceeding a set adaptation degree threshold, and cross-compares a role available time period and a task priority, and generates a task execution time sequence reconstruction result;

[0065] The permission ruling module reads an access permission level of a role corresponding to each task node and a task permission requirement of the task node in the task execution time sequence reconstruction result, constructs a role permission difference matrix and counts a permission satisfaction number of times of the role in the task, extracts a role set with a permission satisfaction number of times lower than a median value in a total number of tasks, obtains a task completion rate of the role set in all task records, sorts and adjusts the access permission level of the role according to the completion rate, and forms a permission ruling result;

[0066] The production and teaching integration task scheduling module re-reads a role available time period and a teaching task priority according to the role with an adjusted permission in the permission ruling result, sorts a task node adaptation degree in the teaching task and constructs a task priority execution sequence, extracts a task node combination with an overlapping degree exceeding an overlapping degree threshold in the role available time period in the priority sequence, rearranges the integration scheduling path according to the time sequence, and outputs a production and teaching integration task scheduling result.

[0067] The behavior pattern recognition result includes a behavior number index set, a task success frequency classification table, and a behavior sequence clustering label; the semantic scene configuration optimization result includes a task semantic mapping table, a node display duration adjustment item, and a node execution order rearrangement item; the task execution time sequence reconstruction result includes a node adaptation degree sorting table, a task priority matching item, and a role time period intersection matrix; the authority ruling result includes an authority level adjustment item, a role access authority grading table, and an authority satisfaction frequency sorting table; and the production and education integration task scheduling result includes a node combination reordering column, a role available time period integration table, and an integration task scheduling path set.

[0068] Referring to Figure 2 , the behavior recognition module includes:

[0069] The behavior collection submodule obtains student and industry role interaction data based on operation logs, records behaviors according to task numbers and time sequences, extracts operation times, calculates time differences between behaviors, and filters out behaviors exceeding the limit, to obtain an interaction behavior interval data set;

[0070] The user's entire log record generated on the interactive platform is called, and the record fields related to the task execution behavior are extracted, including user ID, task number, behavior event name, behavior event timestamp, behavior feedback identifier, event duration, terminal device identifier, operation path, and operation type. By traversing the log record, the behavior events are sorted in ascending order according to the timestamp field, and a behavior event sequence arranged in chronological order is constructed. For example, for a user with a student ID of S001, the behavior records generated in task stage A include a series of behavior events such as entering the task, clicking the resource, and submitting the result, with timestamps of 10:03:12, 10:03:25, and 10:04:40. The system sorts the behavior sequence in order and locates the operation order, then processes the behavior sequence in each group of tasks by grouping according to the task number field, extracts the start and end time points of the behavior sequence in each group of tasks, and calculates the duration to obtain the start and end time of each task behavior chain. Then, according to the "success" or "failure" in the feedback identifier field, the task feedback is classified, and the time difference is calculated in the sequence. The time difference threshold is set to 180 seconds. When the time interval between two consecutive behaviors exceeds the threshold, the system considers that the operation behavior has been interrupted, and the behavior segment is excluded from the analysis sequence. For example, the behavior interval of student S001 in task A is (10:03:12 to 10:03:25 for 13 seconds, and 10:03:25 to 10:04:40 for 75 seconds), which does not exceed 180 seconds, so the behavior chain is retained. Conversely, if the interval exceeds 180 seconds, such as 250 seconds, the behavior chain is excluded, and the remaining behavior is recombined into a new valid behavior segment. In addition, the time difference between each pair of retained behavior events is calculated and a time interval list is formed. For example, the difference between clicking the resource and submitting the result is 75 seconds, which is recorded as the time interval value of the behavior pair in the interactive behavior interval dataset. Each record in the interactive behavior interval dataset includes five fields: task number, pre-behavior time, post-behavior time, operation time difference, and behavior feedback identifier. For example, task number A001, pre-behavior 10:03:25, post-behavior 10:04:40, time interval 75 seconds, and feedback identifier "success". Finally, the interactive behavior interval dataset is formed under multiple user and multiple task dimensions.

[0071] The sequence construction submodule clusters the behavior records according to the interactive behavior interval dataset, counts the number of successful tasks and calculates the frequency, selects the behavior sequence with a frequency exceeding the average, extracts the behavior node order, and generates a high-frequency behavior node sorting sequence.

[0072] For each user task chain, the task behavior sequence is abstracted as a node sequence, and the node identifier is composed of the behavior event name and the time difference, such as "click resource (13s)-submit task (75s)". Then the system performs frequency statistics on each behavior node sequence. For each behavior path sequence, such as "enter task-view resource-submit task", the system records the number of occurrences in the data set. By traversing all user task behavior chains and matching the sequence combination frequency, the statistical threshold is set as the average frequency. Assuming that there are 120 types of behavior paths, the frequencies are as follows: sequence A appears 32 times, sequence B appears 26 times, sequence C appears 18 times, and sequence D appears 7 times. The average frequency is The system only retains the path sequences with a frequency greater than the average. In this example, sequences A, B, and C are retained. After filtering, the node sequence extraction phase is entered. For each high-frequency path, the node composition and interval sequence are extracted according to the behavior occurrence order, such as sequence A "login-browse-ask-submit", with time intervals of (12s, 34s, 51s). The node sequence is N1→N2→N3→N4, respectively represented as N1="login", N2="browse", etc. During the node sequence recording process, the same type of high-frequency sequence is standardized and the weight is converted. For example, if the frequency of the path "browse-submit-answer" is 26 times, accounting for about 21.6% of all behavior paths, the corresponding weight is 0.216. Each high-frequency behavior path is calculated according to the frequency normalization is the number of occurrences of the high-frequency sequence, is the total number of high-frequency sequences. The final high-frequency behavior node ordering sequence is a list of three-tuple containing "node sequence + behavior interval + weight value", such as sequence 1 = [N1 login, N2 browse, N3 submit], with time intervals of (12s, 27s) and weight of 0.268. This list will be used as the standard ordering reference data input system for subsequent behavior chain offset analysis in the failure phase.

[0073] The behavior chain recognition submodule calls the high-frequency behavior node ordering sequence to identify the continuous failure task phase, calculates the behavior node ordering offset, and combines the failure number ratio using the formula:

[0074] ;

[0075] Get each failure phase offset, establish a phase and node mapping, and get the behavior pattern recognition result.

[0076] where, represents the behavior chain order offset value, represents the ​the position of a node in the high-frequency sequence, representative node average order, representative adjacent node time difference, representative failed phase duration, representative node delay value;

[0077] Read the behavior node chain of each task and its execution feedback label, filter the tasks with feedback "failed" and segment them in chronological order. Assuming tasks T001 to T004 failed consecutively, the system divides them into a failure phase, and then enters the offset calculation process. For all behavior sequences in this failure phase, the system compares their behavior node order with the high-frequency behavior node sequence to calculate the offset of each node in the high-frequency sequence. Set the high-frequency sequence as [N1: login, N2: browse, N3: submit, N4: answer], and the actual execution sequence as [N2: browse, N1: login, N4: answer, N3: submit]. The system converts the position information to position index (high-frequency sequence index: 1, 2, 3, 4, actual: 2, 1, 4, 3), and calculates the offset value of each node as |2−1|=1, |1−2|=1, |4−3|=1, |3−4|=1, and the total offset as 4. The node average order is , and the node order difference part is calculated as Then the system processes the square sum of the adjacent behavior node time difference list. Set the difference values as 35 seconds, 28 seconds, and 47 seconds, and the square sum as , the square root as , the total execution duration of the failure phase as , and the response delay value of each node as , for example, the node delays are 8 seconds, 10 seconds, 6 seconds, seconds. Substitute into the formula: ;

[0078] Get the behavior chain order offset of the failure phase , then calculate the offset value for all failure phases, and establish a mapping relationship to record the node sequence, offset value, time span, and other information of each failure phase for behavior pattern structure comparison and risk identification. In this process, the system sets the offset alert threshold For 0.2, the offset degree is between 0.1 and 0.2 according to experience, which is a controllable stage, more than 0.2 is a serious offset, and less than 0.05 is a standard operation stage. According to the actual detection result, it is judged that this stage is a serious offset stage, and the behavior chain needs to be marked as an abnormal sequence for subsequent identification and training processing. Finally, the behavior pattern recognition result is obtained, which is in the form of "failure stage number T001-T004, offset degree 0.227, node offset order degree is high, action delay is concentrated between node N3 submission and N4 answer, and abnormal behavior feature extraction and correction need to be focused on".

[0079] Please refer to Figure 3 , the semantic scene optimization module comprises:

[0080] The node data extraction submodule calculates the average display times based on the total display duration and times of nodes in the behavior sequence, obtains the median value of the display times, judges whether the average display times of each node are lower than the median value, selects the nodes meeting the conditions, and generates a low-frequency node set;

[0081] The time duration data and access times data of the extraction task node in the multiple user behavior data are aggregated in the behavior log with "behavior type" as the field, the display type node with complete start and end time in the behavior event is selected, and the time duration data of each node activation is recorded. In the example, node A appears 3 times in the data of students S001, S002 and S003, and each display duration is 12 seconds, 18 seconds and 14 seconds, the total display duration is 44 seconds, the access times is 3, and the average display times is calculated as , the average display duration is seconds; the system sorts all task nodes in ascending order according to the display times, obtains the median display times, and if the system has a total of 13 nodes, the display times set is: 1, 1, 1, 2, 2, 2, 3, 3, 3, 4, 5, 6, 7, the median display times is the 7th item, that is, 3, and then it is judged whether each node is less than the median value. For example, the display times of node A is 3, which does not meet the condition, and the display times of node B is 2, which enters the selection set; the system performs the comparison operation node by node, and outputs the node set with display times less than the median value 3 as the subsequent processing item. In the actual data set, the number of selected nodes is 6, including nodes B, C, F, G, I and J, with access times 2, 1, 2, 1, 1 and 2 respectively, and total display duration 28 seconds, 12 seconds, 24 seconds, 15 seconds, 9 seconds and 18 seconds respectively. The judgment process is based on the median number determination logic, which has a clear interval division, and the display times interval can be set as: low-frequency display ≤2, normal display 3-5, and high-frequency display >5, to form a stable task execution frequency partition. Finally, the system outputs the low-frequency node set.

[0082] The failure distribution calculation submodule calls the number of failures of each node in the low-frequency node set, calculates its failure ratio, obtains the average failure ratio of all nodes, determines whether the failure ratio of a node is higher than the average, and generates a list of failure nodes with higher than average ratio.

[0083] Match the task feedback identifier field from the behavior records, filter records where the task result is failure, and count the number of times each low-frequency node appears in the failed tasks. Assuming node B appears 6 times in the failed task sequence and 12 times in the total behavior path, its failure rate is... Following this method, all low-frequency nodes are counted sequentially. Example calculation results are: nodes B (0.50), C (0.75), F (0.40), G (0.20), I (0.33), and J (0.60). The system then calculates the average failure rate of these nodes. This value is the "average failure rate benchmark". Then, the failure rate of each node is compared with 0.46. If it is greater than this average, it is included in the high-risk display node set. For example, nodes C (0.75) and J (0.60) are higher than 0.46, and the system marks them as "nodes with a high failure rate". This judgment logic does not rely on abstract weight settings, but judges by specific ratios. The failure rate range is defined as follows: failure rate > 0.46 is a high failure node, 0.30 to 0.46 is a medium-risk node, and < 0.30 is a stable node. The division has a direct range judgment logic and the values ​​have engineering significance. Finally, a list of failure node ratios above the average is generated.

[0084] The display duration adjustment submodule calls the list of failure nodes exceeding the average ratio and the original display duration, and uses the following formula based on the failure ratio, number of displays, skip delay time, and time difference between nodes:

[0085] ;

[0086] The display duration adjustment value of the node is obtained through calculation. The order of the task nodes is rearranged according to the adjusted display duration to generate semantic scene configuration optimization results.

[0087] in, Indicates the first The adjusted display duration for each task node. This is the original display duration of the node. For the first The failure rate of each node, The average failure rate of all nodes. For the number of times a node is accessed. The time interval between task nodes. The delay time caused by a node being skipped;

[0088] Assume the original display duration of node C is T.k = 24 seconds, the failure rate FR1 = 0.75, the access frequency VC1 = 3, the inter-node time difference data is: TD1 = 6 seconds, TD2 = 5 seconds, TD3 = 8 seconds, the skip delay time is SD1 = 4 seconds, SD2 = 3 seconds, SD3 = 2 seconds, and the average failure rate is calculated as , and the result is as follows:

[0089] The expansion of each item is as follows:

[0090] The numerator part is: ;

[0091] The denominator part is: ;

[0092] The overall incremental value is: ;

[0093] Therefore, the display duration of node C is adjusted to:

[0094] seconds;

[0095] The final output node display duration adjustment value is obtained, and the nodes are sequentially rearranged according to the value to generate a semantic scene configuration optimization result. The result shows that node C needs to increase its display time because of the high failure rate and frequent skipping. The adjusted display duration is 24.07 seconds, which is 0.07 seconds longer than the original display duration, and the adjustment ratio is 0.29%. This numerical result represents the visibility of the node in the semantic scene and needs to be adjusted upwards. The display phase should be extended to align with the abnormal behavior feature intensity. This value is used as a display duration adjustment factor to directly participate in the node sequence rearrangement and affect the overall task chain structure. It is further used for the derivation and output of the semantic scene configuration optimization result.

[0096] Please refer to Figure 4 , the timing scheduling module includes:

[0097] The execution efficiency extraction submodule calculates the total number of task completions and the total amount of execution time based on the interaction records of each task node in the semantic scene configuration optimization result, calculates the average execution time and completion rate, and obtains the node average execution time value and node completion rate value.

[0098] Task nodes are numbered and grouped by node dimension. The total number of task completions and total execution time for each node across all interactions are extracted. The number of task completions is obtained by counting the number of entries with the task status field set to "Completed" in the interaction records. The total execution time is obtained by summing the differences between the node's start and end times. In the example, node T001 completed 8 records out of 12 interactions, with a total execution time of 1020 seconds. Therefore, the average execution time for this node is... Seconds, node completion rate is The system executes the same process on all nodes to obtain the node number and its two corresponding key indicator parameters, and establishes an index mapping table. In this mapping table, each node number is recorded with its average execution time and completion rate. The system deletes abnormal records without duration or status fields based on field integrity verification. In the actual system scenario, students S001, S002, and S003 all execute task nodes T001, T002, and T003, with a total of 72 behavioral data entries involving 3 roles. The total number of task completion markers is 52, and the total number of activated node records is 96. Through the above statistics and calculations, the average execution time and node completion rate of all nodes are obtained.

[0099] The adaptability calculation submodule calls the average execution time of the node and the node completion ratio, combined with failure probability, execution fluctuation, and task complexity metrics, using the following formula:

[0100] ;

[0101] The task node fit value of the node is obtained by calculation, and then the fit value is compared with a set threshold to filter and generate a set of nodes with high fit.

[0102] in, Indicates the first The fit value of each task node. For node completion rate, The probability of failure. The average execution time across all nodes. For nodes Average execution time For nodes Execution time variance This refers to the relative task complexity.

[0103] The failure probability (FP) of a task node is obtained by dividing the number of failures by the number of activations. For example, if node T002 is activated 15 times and fails 5 times, then FP2 = ... The node completion rate (PC) is calculated by dividing the number of successful nodes by the number of activations. Let's say 10 successful nodes are represented by PC2 = ... , the average execution duration DT2 = 135 seconds, the statistical average duration of all nodes is 49 seconds², the complexity level code DC is set to 5, the maximum level is 10, so RC2 = , substitute the formula operation:

[0104] ;

[0105] Formula expansion calculation:

[0106] Numerator = ;

[0107] Denominator = ;

[0108] Adaptation value = ;

[0109] Referring to the average execution stability interval standard deviation and the average completion rate of the nodes, if the adaptation threshold is set to 1.0, combined with the distribution of adaptive nodes and abnormal nodes in the previous training set, it is derived that the interval is set to 0.8-1.2, so the adaptation degree higher than 1.0 is determined to be sufficient adaptation, and the node T002 meets the conditions and is marked as a schedulable task.

[0110] The results show that the adaptation value of node T002 is 1.019, which has exceeded the threshold of 1.0, indicating that it can support scheduling arrangement in terms of behavior stability, duration consistency and complexity bearing, and the value directly constitutes the task node adaptation value and enters the subsequent node screening.

[0111] The task timing generation submodule calls the high adaptation node set, compares the task priority and the available time period of the role according to the intersection, rearranges the node order according to the time window, and generates the task execution timing reconstruction result.

[0112] The high adaptation nodes are ranked in descending order according to the task priority field, and then each node belonging to a role is traversed to query the available time interval in the task execution period, for example, the role corresponding to the node T002 is R03, and the available time interval is 8:00-11:30 and 13:30-17:00 every day during the task period from June 1, 2024 to June 15, the system cross-matches the expected execution time of the task node with the idle time interval of the role, if the average execution time of the node is 135 seconds, the system inserts a time window in the role slot in units of minutes, after the same process is performed on all nodes, all nodes are sorted according to the task priority and time sequence, and a new node time arrangement table is generated, which records the node number, expected start time, expected end time, bound role and task priority, the system checks whether there is a conflict and time overlap between nodes, and when an overlap is detected, the node with a low priority is removed or delayed according to the priority, and the loop is processed until a non-overlapping time plan for all tasks is generated, and finally the task execution time sequence reconstruction result is obtained.

[0113] Please refer to Figure 5 , the authority decision module comprises:

[0114] The role access permission reading submodule obtains the role access permission level and task node permission requirement of each task node according to the task execution time sequence reconstruction result, and generates a role access permission record;

[0115] Assuming that the access permission level of role A is 2, indicating that the access permission of this role on the task node is at a medium level, and the access permission level of role B is 3, indicating that it has a higher access permission, and the node task permission requirement defines the minimum permission level required for task completion, for example, the permission requirement of a certain task node is 2, indicating that the node requires the role to have at least medium permission to operate. In this process, all role access permissions and task node permission requirements are extracted from the database or record of task execution time sequence reconstruction and used as data input for subsequent processing, and a role access permission record is generated, which contains the permission information of each role on each task node and serves as the basis for subsequent construction and adjustment of the permission difference matrix.

[0116] The permission difference matrix construction submodule constructs a role permission difference matrix based on the role access permission record, using the formula:

[0117] ;

[0118] Calculate the permission difference of each role on different task nodes to generate a permission difference matrix;

[0119] Wherein, represents the permission difference between the role and the task node , Representative roles At the access permission level of the task node, Representative task node In the mission Permission requirements, Total number of tasks;

[0120] Assume there are 3 roles (A, B, C) and 3 task nodes. The permission requirements for the task nodes are as follows: [Task 1 permission requirement = 2, Task 2 permission requirement = 3, Task 3 permission requirement = 1]. The access permission levels for each role are as follows:

[0121] Character A: Quest 1 level = 2, Quest 2 level = 3, Quest 3 level = 1;

[0122] Character B: Quest 1 level = 1, Quest 2 level = 2, Quest 3 level = 1;

[0123] Character C: Quest 1 level = 3, Quest 2 level = 2, Quest 3 level = 3;

[0124] Substitute into the formula to calculate:

[0125] Step 1: Calculate permission differences:

[0126] For character A:

[0127] Differences in Task 1: (Performs permission requirements);

[0128] Differences in Task 2: (Performs permission requirements);

[0129] Differences in Task 3: (Meets permission requirements);

[0130] For character B:

[0131] Differences in Task 1: (Permission requirements not met);

[0132] Differences in Task 2: (Permission requirements not met);

[0133] Differences in Task 3: (Meets permission requirements);

[0134] For character C:

[0135] Differences in Task 1: (Permission requirements not met);

[0136] Differences in Task 2: (Permission requirements not met);

[0137] Task 3 difference: (Does not meet the permission requirements).

[0138] Step 2: Count the number of times the role's permissions meet in the task:

[0139] Role A: Number of tasks that meet permission requirements = 3 (all tasks meet);

[0140] Role B: Number of tasks that meet permission requirements = 1 (only task 3 meets);

[0141] Role C: Number of tasks that meet permission requirements = 0 (no tasks meet).

[0142] Step 3: Compare with the median value of the total number of tasks:

[0143] The total number of tasks is 3, so the median value of the total number of tasks is 2. According to the median value, the number of times Role A's permissions meet is higher than the median value, while Role B and Role C are lower than the median value.

[0144] Step 4: Generate the number of times the permissions meet the matrix:

[0145] The generated number of times the permissions meet the matrix is as follows:

[0146] Role A: Number of times = 3;

[0147] Role B: Number of times = 1;

[0148] Role C: Number of times = 0;

[0149] The results show that Role A meets the permission requirements in all task nodes, meaning that the role has sufficient permissions to perform all tasks, so the number of times its permissions meet is 3, which exceeds the median value of the total number of tasks 2; Role B only meets the permission requirements in task 3, with a permission meeting number of 1, which is lower than the median value, so further adjustment of access permissions is needed; Role C fails to meet the permission requirements of any task, with a permission meeting number of 0, which means its permission settings are completely incompatible with the task requirements and also need to be adjusted.

[0150] The permission adjustment submodule extracts the role set whose number of times the permissions meet is lower than the median value of the total number of tasks in the number of times the permissions meet the matrix, obtains the completion rate of the role set in the task, adjusts the role access permission level, and generates the permission ruling result;

[0151] From the permission satisfaction times matrix, a set of roles with permission satisfaction times lower than the median value of the total number of tasks is extracted. These roles may not be able to meet the permission requirements on more task nodes due to insufficient or mismatched permissions. The task completion rate of these roles in all task nodes is calculated, which is the proportion of completed tasks for each role. Assuming that role A has completed tasks 1 and 3, but not task 2, the completion rate of role A is 2 / 3, i.e. 66.7%. By comparing the completion rates of all roles, the roles are sorted by completion rate. Roles with lower completion rates will be adjusted to access permission levels. Generally, roles with lower completion rates will be given higher access permissions to ensure that they can complete tasks smoothly during task execution. Finally, the access permission levels of the roles are adjusted based on the sorting results to obtain the permission ruling results.

[0152] Please refer to Figure 6 , the production and teaching integration task scheduling module includes:

[0153] The role information reading submodule reads the available time period and teaching task priority of the roles based on the adjusted permission roles in the permission ruling results, obtains the available time period information of each role in a specific task, and sorts them according to the priority of the teaching task to generate a role available time period record.

[0154] The permission adjustment of each role is obtained. The adjusted permission determines the task nodes and time periods that the role can participate in. In actual scenarios, role A may be assigned to task node 1, role B to task node 2, and role C to task node 3. The available time period information of the roles is obtained based on the actual task schedule, the free time of the roles, and the requirements of the teaching tasks. According to the available time period of each role and the priority of each task, the priority is sorted, which is usually determined by the importance, urgency, and key role in the task. For example, the priority of task 1 is higher than that of task 2, and the priority of task 2 is higher than that of task 3. Finally, appropriate roles are assigned to each task node to generate a role available time period record.

[0155] The task node adaptation degree sorting submodule calculates the adaptation degree of the role node based on the role available time period record and the teaching task priority, evaluates the matching degree between the role and the task node, and sorts them to generate a task node adaptation degree sorting.

[0156] The adaptation degree between the role and the task node is calculated by comparing the available time period of the role and the time requirement of the task node, for example, the task node 1 requires to be carried out from 8:00 to 10:00, and the available time period of the role A is from 9:00 to 10:00, so the adaptation degree of the role A on the task node is 1 hour, and the adaptation degree is quantified according to the time overlap degree of the role and the task, for each role, the adaptation degree value represents the actual participation time of the role on the task node, then, by evaluating the adaptation degree of each role on the task node, the task nodes are sorted according to the role adaptation degree from high to low, the combination of the role with high adaptation degree and the task node will be arranged in priority, and the basis for sorting is mainly the priority of the task node and the available time of the role in the task, and the task node adaptation degree sorting is generated.

[0157] The scheduling path construction submodule extracts the task node combination whose available time period overlap degree between roles in the task node adaptation degree sorting exceeds the overlap degree threshold, rearranges according to the time sequence, and adopts the formula:

[0158] ;

[0159] The operation obtains the scheduling path rearrangement, forms the production-education integration scheduling path, and generates the production-education integration task scheduling result.

[0160] Among them, represents the total length of the overlap period of the task node , represents the available time period length of the role in the task node , represents the task execution time of the role in the task node , is the number of roles participating in the task;

[0161] Suppose that the role A of the task node 1 has a task execution time of 1 hour, and the available time period of the role A is 2 hours (9:00-11:00); the role B of the task node 1 has a task execution time of 1.5 hours, and the available time period of the role B is 1 hour (9:00-10:00); the role A of the task node 2 has a task execution time of 1 hour, and the available time period of the role A is 1.5 hours (9:00-10:30); the role B of the task node 2 has a task execution time of 2 hours, and the available time period of the role B is 2 hours (9:00-11:00).

[0162] The overlap period of the task node 1 is:

[0163] For the role A: ;

[0164] For the role B: ;

[0165] ;

[0166] Overlap period of task node 2:

[0167] For role A: ;

[0168] For role B: ;

[0169] ;

[0170] The overlap period obtained by calculation is 2 hours for task node 1 and 3 hours for task node 2, and if the set overlap threshold is 0.3 hours, it can be determined that task node 1 and task node 2 meet the combination condition and are rearranged. In this case, task node 2 will be arranged first due to the longer overlap period, and finally the scheduling path is formed.

[0171] The above merely illustrates the specific embodiments of the present application, but the protection scope of the present application is not limited to this. Any person skilled in the art can easily think of changes or replacements within the technical range disclosed by the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A virtual simulation technology driven integration of production and teaching teaching system, characterized in that, The system comprises: The behavior recognition module obtains student and industry role interaction behavior data, constructs a behavior sequence according to a task number, calculates a behavior correlation degree according to an operation time difference, counts a task success number according to a behavior feedback identifier, classifies the behavior sequence, and generates a behavior pattern recognition result; The semantic scene optimization module maps a teaching task and a semantic label based on the behavior pattern recognition result, adjusts a node display duration and a sequence according to a mapping result, and generates a semantic scene configuration optimization result; The semantic scene optimization module comprises: A node data extraction submodule calculates an average display number of times based on a total display duration and a number of times of a task node in the behavior pattern recognition result, obtains a median value of the display number of times, judges whether the average display number of times of each node is lower than the median value, filters out nodes meeting the condition, and generates a low-frequency task node set; A failure distribution measurement submodule calls a failure number of each node in the low-frequency task node set, calculates a failure proportion, obtains a mean value of failure proportions of all low-frequency task nodes, judges whether a failure proportion of each node is higher than the mean value, and generates a super-mean value failure node proportion list; An display duration adjustment submodule calls the super-mean value failure node proportion list and an original display duration, obtains a display duration adjustment value of a node by using a formula according to a failure proportion, a display number of times, a delay time generated by a node being skipped, and a time interval between task nodes, and rearranges a task node sequence according to the adjusted display duration, to generate a semantic scene configuration optimization result; ； A time sequence scheduling module sorts a node adaptation degree by calling a task node average execution duration and a completion rate in the semantic scene configuration optimization result, cross-comparing a role time period and a task priority, and generating a task execution time sequence reconstruction result; wherein, denotes the adjusted display duration of the task node, denotes the original display duration of the node, denotes the failure rate of the node, denotes the average failure rate of all nodes, denotes the number of displays of the node, denotes the time interval between task nodes, denotes the delay time generated by the node being skipped. The time sequence scheduling module comprises: An execution efficiency extraction submodule calculates a total number of task completions and a total amount of execution time according to a node number based on interaction records of each task node in the semantic scene configuration optimization result, respectively obtains an average execution duration and a completion rate, and obtains a node average execution duration value and a node completion rate value; An adaptation degree calculation submodule calls the node average execution duration value and the node completion rate value, combines a failure probability, an execution fluctuation, and a task complexity index, and obtains a node task node adaptation degree value by using a formula, and then compares the adaptation degree value with a set threshold value to filter and generate a high-adaptation-degree node set; A task time sequence generation submodule calls the high-adaptation-degree node set, cross-compare a task priority and a role available time period, rearranges a node sequence according to a time window, and generates a task execution time sequence reconstruction result; ； A permission ruling module reads a node role permission in the task execution time sequence reconstruction result, sorts roles whose permission satisfaction number is lower than a median value in a total number of tasks according to a completion rate, and adjusts a role access permission level to form a permission ruling result; in, Indicates the first The fit value of each task node. For node completion rate, The probability of failure. The average execution time across all nodes. For nodes Average execution time For nodes Execution time variance This refers to the relative task complexity. The permission ruling module comprises: A role permission reading submodule obtains a role access permission level and a task node permission requirement of each task node according to the task execution time sequence reconstruction result, and generates a role access permission record; ​ ​ The role access permission record is used to construct a role permission difference matrix by a role permission difference matrix construction submodule, and a formula is used: ； The permission difference of each role at different task nodes is calculated to generate a permission difference matrix. in, Representative roles With task nodes Differences in permissions between them Representative roles At the task node Access permission levels, Representative task node In the mission Permission requirements, Total number of tasks; The permission adjustment submodule extracts a role set whose permission satisfaction frequency is lower than the median value in the task total number in the permission satisfaction frequency matrix, obtains the completion rate of the role set in the task, adjusts the role access permission level, and generates a permission ruling result. The production-education integration task scheduling module extracts a node combination whose time period overlap degree exceeds a threshold value according to the available time period of the role whose permission has been adjusted in the permission ruling result and the task priority, rearranges the time sequence to form an integration scheduling path, and outputs a production-education integration task scheduling result. The production-education integration task scheduling module comprises: The role information reading submodule re-reads the available time period of the role and the priority of the teaching task according to the role whose permission has been adjusted in the permission ruling result, obtains the available time period information of each role in a specific task, and sorts the information according to the priority of the teaching task to generate a role available time period record. The task node adaptation degree sorting submodule calculates the adaptation degree of the role node based on the role available time period record and the priority of the teaching task, evaluates the matching degree between the role and the task node and sorts them, and generates a task node adaptation degree sorting. The scheduling path construction submodule extracts a task node combination whose available time period overlap degree between roles exceeds an overlap degree threshold value in the task node adaptation degree sorting, rearranges the time sequence, and uses a formula: ； The scheduling path rearrangement is obtained by calculation to form a production-education integration scheduling path and generate a production-education integration task scheduling result. in, Representative task node The total length of the overlapping time period, Representative roles At the task node Available time period length, Representative roles At the task node Task execution time, The number of roles participating in this task.

2. The virtual simulation technology-driven production-teaching integration teaching system according to claim 1, characterized in that: The behavior pattern recognition result comprises a behavior number index set, a task success frequency classification table, and a behavior sequence clustering label; and the semantic scene configuration optimization result comprises a task semantic mapping table, a node display time length adjustment item, and a node execution order rearrangement item. The task execution time sequence reconstruction result comprises a node adaptation degree sorting table, a task priority matching item, and a role time period intersection matrix; and the permission ruling result comprises a permission level adjustment item, a role access permission level table, and a permission satisfaction frequency sorting table. The production-education integration task scheduling result comprises a node combination reordering column, a role available time period integration table, and an integration task scheduling path set. 3.The virtual simulation technology driven production-teaching integration teaching system according to claim 1, characterized in that: The behavior recognition module comprises: The behavior collection submodule obtains student and industry role interaction data based on operation logs, records behaviors and extracts operation times according to task numbers and time sequences, calculates time differences between behaviors and filters out behaviors exceeding the limit, and obtains an interaction behavior interval data set. The sequence construction submodule clusters behavior records according to the interaction behavior interval data set, counts the number of successful tasks and calculates the frequency, filters out behavior sequences whose frequency exceeds the average value, extracts node sequences, and generates a high-frequency behavior node sorting sequence. For each user task chain, the task behavior sequence is abstracted as a node sequence, and the node identifier is composed of the behavior event name and the time difference. Then, the system performs frequency statistics on each behavior node sequence. For each behavior node sequence, the system records its occurrence times in the data set. By traversing all user task behavior chains and matching the sequences, the system sets the statistical threshold as the average frequency, and only retains the path sequences with a frequency greater than the average; The behavior chain recognition submodule calls the high-frequency behavior node sorting sequence to identify the continuous failure task stage, calculates the behavior chain order offset value, and combines the failure number ratio to use the formula: ； Get the offset degree of each failure stage, establish the stage and node mapping, and get the behavior pattern recognition result; wherein, represent a behavior chain order offset value, represent a position of the first node in a high frequency order, represent a node average order, represent a neighboring node time difference, represent a failed phase duration, represent a node delay value.

Images