An adaptive logic resource configuration method, device and medium based on a simulation teaching system

Abstract

The application discloses a kind of based on simulation teaching system's self-adapting logic resource configuration method, equipment and medium, belong to virtual simulation teaching technical field, method includes: the teaching configuration information of collection is parsed as teaching semantic atlas;Filter out key semantic node, build the target topology parameter set containing role constraint information and sequence constraint information;Candidate practical training topology graph is generated, and it is comprehensively scored;The simulation verification process corresponding to the comprehensive score result is selected and verified, and simulation verification result is obtained;Candidate practical training topology graph is diagnosed and the configuration value in the target topology parameter set is adjusted according to the diagnostic result to generate updated teaching configuration information and store.The application can update self-adaptively to logic resource configuration, so that virtual simulation practical training topology is continuously approximated to teaching target requirement, and the matching degree between virtual simulation environment and real teaching logic is improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of virtual simulation teaching, and particularly relates to a self-adaptive logical resource configuration method, device and medium based on a simulation teaching system. BACKGROUND

[0002] At present, virtual simulation teaching has been widely applied in network security, industrial control and other fields. A related platform usually organizes multiple virtual hosts, virtual networks and various image resources into a complete practical training environment through virtualization and cloud computing technology, so that students can carry out attack and defense drills or comprehensive experiments. In such a platform, the topology structure of the practical training environment is often described by a configuration file. The configuration includes the types of virtual hosts, image sources, network connection relationships, and control information related to resource management, snapshot recovery, etc. The platform automatically pulls up virtual resources, constructs virtual networks and completes one-key deployment and release of the practical training scene according to these configurations, thereby effectively reusing underlying hardware resources and supporting large-scale virtual simulation teaching practice.

[0003] However, the above-mentioned method of generating an environment driven by a configuration file is more oriented towards technical details at the virtual resource and network level, and does not establish a clear teaching semantic layer based on the theme of a virtual simulation teaching system and method generated based on user configuration. Teachers cannot directly configure and reconstruct according to dimensions such as teaching objectives, experimental steps and difficulty levels in actual use, and can only rely on fixed pre-set topologies or indirectly modify underlying configurations by technical personnel. There is a lack of a mechanism for non-professional users to configure experimental intentions from a teaching perspective, and for the system to automatically derive a safe and usable practical training environment. This leads to obvious deficiencies in existing virtual simulation teaching platforms in supporting the generation and adjustment of simulation teaching systems based on user configurations. SUMMARY

[0004] The present application provides a self-adaptive logical resource configuration method, device and medium based on a simulation teaching system, which can adaptively update logical resource configuration, so that the virtual simulation practical training topology continuously approximates the teaching target requirements, and improves the matching degree between the virtual simulation environment and the real teaching logic.

[0005] The present application provides the following technical solutions:

[0006] In a first aspect, a self-adaptive logical resource configuration method based on a simulation teaching system is provided, comprising:

[0007] Step S1: parsing the collected teaching configuration information into a teaching semantic graph according to a preset teaching semantic rule;

[0008] Step S2: filtering out key semantic nodes from the teaching semantic graph, and constructing a target topology parameter set containing role constraint information and sequence constraint information based on the connection relationship of the key semantic nodes and the pre-defined virtual resource role label;

[0009] Step S3: generating a candidate practical training topology graph in the virtual resource library according to the target topology parameter set, and comprehensively scoring the generated candidate practical training topology graph from the teaching sequence and the operation complexity; selecting a simulation verification process corresponding to the comprehensive score result for verification to obtain a simulation verification result;

[0010] Step S4: based on the comprehensive score result and the simulation verification result, diagnosing the candidate practical training topology graph and adjusting the configuration value in the target topology parameter set according to the diagnosis result to generate updated teaching configuration information and store it.

[0011] Optionally, the step S1 specifically comprises:

[0012] Collecting teaching configuration information from the simulation teaching system and cutting it into a semantic segment set;

[0013] Retrieving the semantic segment set, classifying words or symbols containing sequence prompt implications into a sequence mark set, classifying semantic segments containing conditional words, trigger words or result words into a dependency expression set, and combining semantic segments satisfying the same object condition into semantic segment pairs and classifying them into a semantic segment pair set;

[0014] According to the richness of the sequence mark set relative to the number of semantic segments, a sequence clue density parameter is constructed, and according to the concentration of the dependency expression set in the semantic segment pair set, a dependency trigger correlation parameter is constructed;

[0015] Based on the sequence clue density parameter and the dependency trigger correlation parameter, a semantic connection balance index is constructed, and after comparing with a preset threshold, sequence connections and dependency connections are constructed between semantic nodes corresponding to the semantic segment set, and a teaching semantic graph is generated.

[0016] Optionally, the step S2 of filtering out key semantic nodes from the teaching semantic graph specifically comprises:

[0017] Reading the semantic node set and the connection relationship set in the teaching semantic graph, for each semantic node, counting the in-degree number and the out-degree number, adding the product of the in-degree number and the out-degree number to a set constant, and then performing a natural logarithm operation to obtain a structure bridge parameter;

[0018] Enumerating all simple paths between the starting nodes and the ending nodes in the teaching semantic graph, and counting the participation proportion of each semantic node in the simple paths to generate a path participation parameter;

[0019] Based on the structural bridge parameters and the path participation parameters, the key degree of each semantic node is obtained, and a key semantic node set is obtained based on a preset key degree threshold.

[0020] Optionally, in the step S2, based on the connection relationship of the key semantic nodes and the pre-defined virtual resource role label, a target topology parameter set containing role constraint information and sequence constraint information is constructed, specifically:

[0021] The operation type identifier of each key semantic node is read, different operation type identifiers are mapped to the pre-defined virtual resource role label, and key node resource role constraint information is formed;

[0022] The phase identifier of each key semantic node is read, the key semantic nodes with the same phase identifier are divided into the same phase group, and then the key semantic nodes are sorted within each phase group according to the pointing sequence of the sequence connection in the connection relationship, to obtain phased key node sequence constraint information;

[0023] The key node sequence constraint information of all phase groups and the corresponding resource role constraint information are merged to form the target topology parameter set.

[0024] Optionally, in the step S3, a candidate practical training topology graph is generated in the virtual resource library according to the target topology parameter set, and the generated candidate practical training topology graph is comprehensively scored from the teaching sequence and the operation complexity, specifically including:

[0025] The candidate practical training topology graph is generated under the constraint of the target topology parameter set;

[0026] The sequence pair set is constructed according to the key semantic node sequence constraint, the number of sequence matches and the number of sequence violations are counted in the candidate practical training topology graph, and the semantic sequence coincidence index is obtained by logical function conversion;

[0027] For each key semantic node in the corresponding virtual host node of the candidate practical training topology graph, the nearest starting node is found from the starting node set of the teaching semantic graph, to obtain the shortest path length of the two; the average path length of the key semantic nodes is obtained by averaging the shortest path lengths of all key semantic nodes;

[0028] In the key semantic node set, the target key semantic node with a successor directed connection is filtered out, and the out-degree value is counted; the average branch number of the key semantic nodes is obtained by averaging the out-degree values of all key semantic nodes;

[0029] The average path length and the average branch number are integrated to obtain the operation path complexity index;

[0030] The operation path complexity index is converted into an ideal semantic order value in the form of an exponential decay with a natural constant as the base, an offset between the semantic order fitting index and the ideal semantic order value is obtained, and the offset degree is converted into a score, which is denoted as a topology adaptation coefficient.

[0031] Optionally, in the step S3, a simulation verification process corresponding to the comprehensive score result is selected for verification to obtain a simulation verification result, and specifically:

[0032] The topology adaptation coefficient is compared with a preset topology adaptation threshold interval, if the topology adaptation coefficient is within the interval, a global simulation verification process is selected, if the topology adaptation coefficient is higher than a topology adaptation high threshold, a local critical path simulation verification process is selected, if the topology adaptation coefficient is lower than a topology adaptation low threshold, a basic connectivity simulation verification process is selected, and a simulation verification result is generated according to an execution result of the simulation verification process.

[0033] Optionally, in the step S4, specifically:

[0034] The number of failures and the number of participations of each semantic node are counted according to the simulation verification result to obtain a semantic node failure ratio parameter, the semantic node failure ratio parameter is multiplied by the criticality of the semantic node to obtain a node adjustment priority parameter, and semantic nodes with a node adjustment priority parameter not lower than a node adjustment threshold are selected to form a semantic node adjustment set;

[0035] The number of conflicts and the number of uses of each topology parameter are counted according to the simulation verification result to obtain a topology parameter conflict ratio parameter, the topology parameter conflict ratio parameter is multiplied by the comprehensive score result to obtain a parameter adjustment intensity parameter, and topology parameters with a parameter adjustment intensity parameter not lower than a parameter adjustment threshold are selected to form a topology parameter adjustment set;

[0036] Based on the semantic node adjustment set and the topology parameter adjustment set, the configuration values in the critical semantic node attributes and the target topology parameter set are bidirectionally synchronized and adjusted to generate updated teaching configuration information and store the updated teaching configuration information.

[0037] Optionally, in the step of bidirectionally synchronously adjusting the configuration values in the critical semantic node attributes and the target topology parameter set based on the semantic node adjustment set and the topology parameter adjustment set to generate updated teaching configuration information and store the updated teaching configuration information, specifically:

[0038] For each semantic node in the semantic node adjustment set, the teaching stage identifier, the operation type identifier, the predecessor semantic node set, the successor semantic node set, the predecessor connection type and the successor connection type of the semantic node are read from the teaching semantic graph, and the order constraint and the role constraint associated with the semantic node are located from the target topology parameter set;

[0039] For each semantic node in the semantic node adjustment set, all fault records in the simulation verification result with the semantic node as the fault source or fault path component are summarized to determine the dominant fault type of the semantic node; the teaching stage identifier, operation type identifier and connection type of the semantic node are adjusted according to the corresponding adjustment strategy of the dominant fault type, and the stage constraint, sequence constraint and role constraint in the target topology parameter set are synchronously modified;

[0040] For each topology parameter in the topology parameter adjustment set, the role constraint, sequence constraint and stage constraint are adjusted according to the constraint type to which the topology parameter belongs, and the connection relationship of the corresponding semantic node in the teaching semantic graph is synchronously modified.

[0041] The updated teaching configuration information set is generated according to the mapping rule, and is stored.

[0042] In a second aspect, a computer device is provided, comprising a processor and a memory; wherein the processor implements the steps of the adaptive logical resource configuration method of the simulation teaching system according to any one of the first aspect when executing the computer program stored in the memory.

[0043] In a third aspect, a computer readable storage medium is provided for storing a computer program; the computer program is executed by a processor to implement the steps of the adaptive logical resource configuration method of the simulation teaching system according to any one of the first aspect.

[0044] Compared with the prior art, the beneficial effects of the present application are:

[0045] (1) The present application parses the teaching configuration information provided by the user into a teaching semantic graph, extracts key semantic nodes at the semantic level and forms a target topology parameter set, constructs an adaptation determination result in combination with the order and complexity evaluation result of the candidate practical training topology, and writes back the teaching semantic graph and the target topology parameter set based on the simulation verification result, so as to realize a closed-loop optimization process oriented to user configuration, so that the structure generation of the virtual simulation teaching environment is no longer dependent on the static design of artificial experience, but can be adjusted based on quantitative results, so that the virtual simulation practical training topology continuously approaches the teaching target requirement, and the matching degree between the virtual simulation environment and the real teaching logic is improved.

[0046] (2) The present application introduces a semantic order and operation complexity double-dimensional evaluation mechanism, and uses a topology adaptation coefficient as a global quantitative index, so that the simulation verification result is not limited to whether the function is correct or not, but can reflect the balance state between the rationality of the teaching order and the complexity of the operation path, so as to realize reasonable selection and dynamic grading of the verification strength of the candidate topology, thereby improving the topology generation quality, shortening the iteration verification period, and reducing the resource consumption caused by manual adjustment and repeated construction.

[0047] (3) The simulation verification result is converted into an interpretable fault proportion parameter and a conflict proportion parameter according to the semantic node dimension and the topology parameter dimension, and the difference of the semantic structure and the topology constraint is driven to be adjusted, so that the optimization process is always aimed at the problem concentrated area, thereby improving the effectiveness and accuracy of the configuration adjustment under the premise of ensuring the integrity of the teaching structure, so that each round of optimization can make a clear contribution to the final practical training topology structure, and the usability and teaching implementation effect of the virtual simulation teaching scene are improved. BRIEF DESCRIPTION OF DRAWINGS

[0048] Figure 1 is a step schematic diagram of the adaptive logical resource configuration method based on the simulation teaching system. DETAILED DESCRIPTION

[0049] The application will be further described below in conjunction with the drawings. The following examples are only used to more clearly illustrate the technical solutions of the application, and cannot be used to limit the protection scope of the application. It should be noted that the terms "comprise" and any variations thereof in the specification and claims of the application and the above drawings are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device comprising a series of steps or units does not have to be limited to only those steps or units clearly listed, but can include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0050] Example 1

[0051] As shown in Figure 1 , an adaptive logical resource configuration method based on a simulation teaching system comprises:

[0052] Step S1: parsing the collected teaching configuration information into a teaching semantic atlas according to a preset teaching semantic rule;

[0053] Step S2: selecting key semantic nodes from the teaching semantic atlas, and constructing a target topology parameter set containing role constraint information and sequence constraint information based on the connection relationship of the key semantic nodes and the pre-defined virtual resource role label;

[0054] Step S3: generating a candidate practical training topology graph in the virtual resource library according to the target topology parameter set, and comprehensively scoring the generated candidate practical training topology graph from the teaching sequence and the operation complexity; selecting a simulation verification process corresponding to the comprehensive score result for verification to obtain a simulation verification result;

[0055] Step S4: based on the comprehensive score result and the simulation verification result, performing problem diagnosis on the candidate practical training topology graph and adjusting the key semantic node attributes and the configuration values in the target topology parameter set according to the diagnosis result to generate updated teaching configuration information and store it.

[0056] In the scenario of generating virtual simulation teaching environment based on user configuration, the content input by the teacher in the teaching configuration interface is mostly natural language description of teaching objectives, experimental steps and precautions. If these contents are directly used for subsequent resource scheduling and topology construction, it is difficult to reflect the sequential advancement relationship and conditional dependency relationship in the teaching activity, and it is also difficult to provide a clear structural basis for extracting key semantic nodes and constructing target topology parameter set in subsequent steps. In order to enable the subsequent steps to conduct computable topology derivation in the teaching perspective, it is necessary to convert the original teaching configuration information into a teaching semantic graph containing semantic nodes and connection relationships in step S1, and balance the effects of sequential clues and dependency clues through parameterized analysis in this process.

[0057] Step S1, specifically comprising the following sub-steps:

[0058] Step S11: Collecting teaching configuration information from the simulation teaching system and performing segmentation to form a semantic segment set.

[0059] The teaching configuration information input by the teacher from the teaching configuration interface of the simulation teaching system is collected to form a set of teaching configuration information according to the course or experimental project. Each teaching configuration information record is composed of multiple sentences, which are segmented at the sentence level by the period, semicolon and line break markers in the interface, and then segmented at the phrase level according to the semantic pause points of longer sentences. The smallest semantic unit obtained is classified into the same semantic segment set. Each semantic segment in the semantic segment set corresponds to a semantic node in the subsequent teaching semantic graph, and the overall size of the semantic segment set is used to describe the fine-grained degree of teaching semantics involved in this configuration.

[0060] Step S12: Retrieving the semantic segment set, classifying words or symbols containing sequential prompt meanings into a sequential marker set, classifying semantic segments containing condition words, trigger words or result words into a dependency expression set, and combining semantic segments that satisfy the same object condition into semantic segment pairs and classifying them into a semantic segment pair set.

[0061] All words or symbols containing sequential prompt meanings in the semantic segment set are retrieved, including sequential words representing the beginning, middle and end, and step number markers input in the configuration interface. These sequential prompts are classified into the sequential marker set. The number of elements in the sequential marker set is used to describe the absolute size of the sequential clues, and the number of elements in the semantic segment set is used to describe the absolute size of the semantic segments.

[0062] Further search for semantic expressions containing preconditions, result triggers and state change meanings in the semantic fragment set, identify semantic fragments containing conditional words, trigger words or result words, and classify these semantic fragments into the dependency expression set. The number of elements in the dependency expression set is used to describe the size of the fragments with explicit dependency semantics.

[0063] Identify semantic fragments that appear around the same operation object or the same device identifier in the semantic fragment set, combine semantic fragments that meet the same object condition into semantic fragment pairs, and all semantic fragment pairs form a semantic fragment pair set. The number of elements in the semantic fragment pair set is used to describe the pairing basis size of potential dependency relationships.

[0064] Step S13: According to the richness of the sequence mark set relative to the number of semantic fragments, construct the sequence clue density parameter, and according to the concentration of the dependency expression set in the semantic fragment pair set, construct the dependency trigger correlation degree parameter.

[0065] In order to reflect the relative richness of the sequence prompt relative to the number of semantic fragments, perform logarithmic transformation on the number of sequence marks, and perform logarithmic transformation on the number of semantic fragments, and then compare the two to obtain the sequence clue density parameter. The sequence clue density parameter is used to characterize the approximate abundance of sequence clues carried in unit semantic fragments, and is used as a basis for judging whether to connect in sequence in the subsequent.

[0066] In order to characterize the concentration of dependency expressions in the potential pairing size, the number of dependency expression sets is taken as the numerator, and the square root of the number of semantic fragment pair sets is taken as the adjustment item on the denominator side to compress, and the dependency trigger correlation degree parameter is constructed, so as to reflect the number of dependency expressions and control the influence amplitude of the parameter when the number of semantic fragment pairs is too large. The dependency trigger correlation degree parameter is a dimensionless parameter, which is used to represent the prominence of dependency semantics in the whole semantic fragment, and provides a quantitative basis for the importance of subsequent dependency connection.

[0067] Step S14: Based on the sequence clue density parameter and the dependency trigger correlation degree parameter, construct a semantic connection balance index, and after comparing with a preset threshold, build sequence connection and dependency connection between the semantic nodes corresponding to the semantic fragment set, and generate a teaching semantic graph.

[0068] On the premise that the sequential clue density parameter and the dependency trigger correlation parameter have been obtained, in order to unify the influence of the two types of clues on the connection mode in the teaching semantic graph, the semantic connection balance index is introduced. The semantic connection balance index takes the sequential clue density parameter as the base value, and constructs an exponential decay form of adjustment factor with the dependency trigger correlation parameter. Through a plus adjustment factor to form the denominator, the sequential clue density parameter is taken as the numerator to perform division operation. When the dependency trigger correlation parameter is small, the adjustment factor is large, the denominator is obviously greater than one, and the semantic connection balance index is relatively small, indicating that the sequential connection needs to be used cautiously under strong dependency constraint; when the dependency trigger correlation parameter is large, the adjustment factor tends to be small, the denominator is close to one, and the semantic connection balance index is close to the sequential clue density parameter itself, indicating that the sequential clue can play a greater role under the support of dependency semantics. The semantic connection balance index is still a dimensionless parameter, which is used to provide a nonlinear comprehensive judgment basis between sequential connection and dependency connection.

[0069] On the basis of the calculation of the semantic connection balance index, the semantic node set is constructed from the semantic segment set, and each semantic node corresponds to a semantic segment. The sequential emphasis threshold and the dependency emphasis threshold are set to divide the value interval of the semantic connection balance index. When the semantic connection balance index is higher than the sequential emphasis threshold, the sequential connection relationship is established one by one in the semantic node set according to the appearance order of the semantic segment in the teaching configuration information, and the dependency connection relationship is supplemented for the semantic nodes belonging to the dependency expression set and having semantic segment pairs. When the semantic connection balance index is lower than the dependency emphasis threshold, the dependency connection relationship is established between the semantic nodes corresponding to the semantic segment pair set, and then the sequential connection relationship is established between the semantic nodes not covered by the dependency connection according to the natural order in the configuration. When the semantic connection balance index is between the two types of thresholds, the connection relationship is constructed in an alternating manner between the sequential connection and the dependency connection, so that part of the nodes are connected according to the sequential connection, and the other part of the nodes are connected according to the dependency pairing. Through the above connection strategy, the teaching semantic graph containing semantic nodes and connection relationships is formed, and the teaching semantic graph is used as the input basis for the subsequent steps of identifying the key semantic node set and constructing the target topology parameter set.

[0070] In the embodiment, the value range of the order emphasis threshold is limited in the upper half of the whole possible value interval of the semantic connection balance index, which is obtained by offline analysis of multiple sets of real teaching configuration information. The specific method is as follows: the processing procedure of step S1 is performed in multiple courses or experimental projects respectively, the corresponding semantic connection balance index is calculated piece by piece, and the semantic connection balance index is arranged in ascending order of value. Then, the corresponding relationship between the semantic connection balance index and the order dominant degree judged by artificial is observed. When the artificial considers that the order clue is sufficient as the main connection basis, the corresponding set of semantic connection balance index values is recorded. The lower boundary of the centralized interval of these values is taken as the order emphasis threshold. The order emphasis threshold is thus in a higher position of the statistical distribution of the semantic connection balance index, which not only ensures that the order connection condition has sufficient order clue support, but also provides a clear limit for the subsequent use of the order connection priority strategy above the high threshold.

[0071] In the embodiment, the value range of the order emphasis threshold is limited in the upper half of the whole possible value interval of the semantic connection balance index, which is obtained by offline analysis of multiple sets of real teaching configuration information. The specific method is as follows: the processing procedure of step S1 is performed in multiple courses or experimental projects respectively, the corresponding semantic connection balance index is calculated piece by piece, and the semantic connection balance index is arranged in ascending order of value. Then, the corresponding relationship between the semantic connection balance index and the order dominant degree judged by artificial is observed. When the artificial considers that the order clue is sufficient as the main connection basis, the corresponding set of semantic connection balance index values is recorded. The lower boundary of the centralized interval of these values is taken as the order emphasis threshold. The order emphasis threshold is thus in a higher position of the statistical distribution of the semantic connection balance index, which not only ensures that the order connection condition has sufficient order clue support, but also provides a clear limit for the subsequent use of the order connection priority strategy above the high threshold.

[0072] Through the processing of step S1, the teaching configuration information in the form of natural language input in the teaching configuration interface is parsed into a set of semantic segments, and then the set of order marks, the set of dependency expressions and the set of semantic segment pairs are extracted. The semantic connection balance index is constructed by the order clue density parameter and the dependency trigger association degree parameter. Then, the construction proportion of the order connection and the dependency connection is determined according to the combination of the semantic connection balance index and the preset threshold, and finally the teaching semantic graph which reflects both the order advancing relationship of the teaching process and the operation condition dependency relationship is obtained. The teaching semantic graph provides structured, quantitative and semantically complete basic data for subsequent selection of the key semantic node set from the teaching perspective and generation of the target topological parameter set.

[0073] In step S1, the teaching configuration information has been parsed into a teaching semantic graph, which contains a set of semantic nodes and a set of connection relationships composed of sequential connections and dependency connections. Each semantic node corresponds to a teaching semantic segment, and the stage identifier field and the operation type identifier field recorded in the original configuration are retained. The teaching semantic graph completely presents the structure of the teaching process. However, if all semantic nodes are directly involved in subsequent virtual resource generation, the number of constraints will obviously expand, and the dominant role of key teaching links in topology generation will be diluted by a large number of secondary nodes. Therefore, in step S2, a set of key semantic nodes needs to be identified from the teaching semantic graph, and the set of key semantic nodes is converted into a set of target topology parameters, so that the subsequent candidate practical training topology generation process can accurately respond to the truly important segments in the teaching level on the resource level.

[0074] Step S2 specifically includes the following sub-steps:

[0075] Step S21: Read the set of semantic nodes and the set of connection relationships in the teaching semantic graph. For each semantic node, count the in-degree and out-degree, add the product of the in-degree and the out-degree to a set constant, and then perform a natural logarithm operation to obtain a structure bridge parameter.

[0076] The number of connections pointing to each semantic node in the set of semantic nodes is counted to obtain the in-degree, and the number of connections from the semantic node to other semantic nodes is counted to obtain the out-degree. The in-degree reflects the convergence of the semantic node in the previous teaching segment, and the out-degree reflects the branching of the semantic node in the subsequent teaching segment. To characterize the significant degree of the semantic node in the connection structure, the product of the in-degree and the out-degree is smoothed by first multiplying them and adding a fixed constant, and then performing a natural logarithm operation on the result to obtain a structure bridge parameter. The structure bridge parameter increases when both the in-degree and the out-degree increase. The natural logarithm operation makes the structure bridge parameter increase slowly with the increase in the number of connections, thereby reflecting the bridge value of the semantic nodes that simultaneously accept multiple previous paths and diverge to multiple subsequent paths in the teaching structure.

[0077] Step S22: Enumerate all simple paths between the start nodes and the end nodes in the teaching semantic graph, and count the participation proportion of each semantic node in the simple paths to generate a path participation parameter.

[0078] The semantic node set and the connection relationship set are read in the teaching semantic graph. Each semantic node in the semantic node set is derived from the semantic fragment obtained in step S1 and carries a stage identifier field and an operation type identifier field. By checking the connection relationship set, it is determined whether a predecessor connection exists for each semantic node. Semantic nodes without any predecessor connection are classified into a starting node set, and semantic nodes without any successor connection are classified into a terminating node set. The starting node set is used to describe the logical entry of the teaching process, and the terminating node set is used to describe the logical exit of the teaching process. The starting node set and the terminating node set jointly provide boundary conditions for subsequent path statistics.

[0079] In the path analysis stage, the starting node set and the terminating node set are used as boundaries, and a simple path from any starting node to any terminating node is enumerated in the teaching semantic graph by using a depth-first search method. The length of each simple path is limited to not exceeding a preset upper limit of path length, and the path length is measured according to the number of semantic nodes contained in the path. The number of all simple paths that meet the length requirement is counted as the total number of paths. For each semantic node in the semantic node set, it is counted how many simple paths the semantic node appears in, and the number of paths in which the semantic node participates is obtained. To depict the participation proportion of the semantic node in the overall teaching path structure, the number of paths in which the semantic node participates is divided by one plus the total number of paths, and the path participation parameter is obtained. The numerical range of the path participation parameter starts from zero and theoretically does not reach one. The closer the value is to one, the more paths the semantic node appears in, and the more prominent the role of the semantic node in maintaining the connectivity of the teaching process.

[0080] wherein the simple path refers to a vertex sequence from a starting vertex to a terminating vertex under the constraint of the vertex set and the edge set of the given graph, wherein each vertex appears only once and there is a connection edge between any two adjacent vertices.

[0081] Step S23: Obtain the criticality of each semantic node based on the structure bridging parameter and the path participation parameter, and obtain the key semantic node set based on a preset criticality threshold.

[0082] On the basis that the structural bridging parameter and the path participation parameter have been calculated respectively, the key degree of the node is introduced to comprehensively evaluate the importance of each semantic node. The key degree of the node takes the structural bridging parameter as the basic strength, and then amplifies the influence of the path participation parameter through a natural index transformation mode, that is, the natural index value is calculated according to the path participation parameter, and then the natural index value is multiplied by the structural bridging parameter to obtain the key degree value of the node. In this way, the semantic nodes with stronger teaching path coverage can obtain more obvious promotion in the key degree of the node, so as to pull apart the importance difference of different semantic nodes within the same teaching semantic graph. After the key degree of all semantic nodes has been calculated, the key degree threshold is introduced, and the value range of the key degree threshold is limited between the minimum value and the maximum value of the key degree of the node in the current teaching semantic graph. The acquisition logic of the key degree threshold is: execute the above calculation process on a plurality of representative teaching semantic graph samples, calculate the key degree of the node for all semantic nodes in the samples, invite personnel with relevant teaching experience to label whether these semantic nodes belong to the key teaching link, then compare the key degree value of the node with the artificial labeling, find the key degree value interval that can cover most of the labeled key nodes, and establish the lower boundary of the value interval as the key degree threshold.

[0083] In actual screening, the semantic nodes with the key degree value not lower than the key degree threshold are selected from the semantic node set to form a key semantic node set, and the key semantic node set represents those teaching fragments that have obvious contribution in the structural bridging level and the path participation level.

[0084] After the key semantic node set is determined, the stage identification field, the operation type identification field and the sequential connection and dependent connection information in the connection relationship set in the key semantic node set are converted into the target topology parameter set.

[0085] Step S24: read the operation type identification of each key semantic node, map different operation type identifications to the pre-defined virtual resource role label to form key node resource role constraint information.

[0086] Read the operation type identification field of each key semantic node, map different operation type identifications to the pre-defined virtual resource role label, such as the control role label, the data role label or the observation role label, to form the key node resource role constraint information.

[0087] Step S25: read the stage identification of each key semantic node, divide the key semantic nodes with the same stage identification into the same stage group, and then sort the key semantic nodes in each stage group according to the pointing order of the sequential connection in the connection relationship to obtain the staged key node order constraint information.

[0088] Firstly, the stage identification field is read, the key semantic nodes with the same stage identification field are divided into the same stage group, and then the key semantic nodes are sorted in each stage group according to the sequential connection in the connection relationship set, so that the key node order constraint information in the stage is obtained.

[0089] Step S26: The key node order constraint information and the corresponding resource role constraint information of all stage groups are merged to form a target topology parameter set.

[0090] The target topology parameter set contains key node execution order constraints and virtual resource role distribution constraints for different teaching stages, which are used to guide the virtual host quantity configuration, role allocation and key connection relationship setting in the candidate training topology in subsequent steps.

[0091] Through the processing of step S2, the semantic nodes in the teaching semantic graph are quantized into structure bridge parameters and path participation parameters at the structure level and the path level respectively, and further combined into key degrees of the nodes, and the key semantic node set is formed under the key degree threshold constraint. The stage identification field, operation type identification field and connection relationship information carried by the key semantic node set are arranged and converted into the target topology parameter set, so that the complex teaching semantic structure is compressed into a group of parameters that can directly drive the virtual resource instantiation and connection configuration, and the subsequent candidate training topology generation process is centered on the key teaching link for resource layout and connection design, which improves the pertinence and orderliness of topology generation under the premise of maintaining the integrity of teaching meaning.

[0092] Steps S1 and S2 analyze the teaching configuration information into a teaching semantic graph, and extract a key semantic node set and a target topology parameter set from it. The target topology parameter set clearly defines the sequence constraints of the key semantic nodes in each teaching stage and the corresponding relationship between the key semantic nodes and the virtual resource roles, and gives the skeleton of the candidate training topology from the semantic point of view. However, the target topology parameter set only describes how to connect and which roles should be allocated, and does not answer the key questions such as how much the generated topology follows the teaching order, how much operation cost the student needs to pay at the key position, and whether the sequence and complexity are reasonably balanced. Therefore, in step S3, a training topology graph needs to be generated based on the target topology parameter set, and a group of parameters with clear meaning needs to be constructed around the semantic order and operation complexity, the topology adaptation coefficient is derived through the balance relationship, and then the simulation verification process is selected according to the relationship between the topology adaptation coefficient and the threshold value, to provide reliable judgment basis for subsequent adjustment of the teaching semantic graph and the target topology parameter set.

[0093] Step S3, specifically comprising the following substeps:

[0094] Step S31: generate a candidate training topology graph under the constraints of the target topology parameter set.

[0095] In the target topology parameter set, read the set of key semantic nodes, the phase identifier field and the resource role constraint information of each key semantic node, map each key semantic node to a virtual host node, and the set of virtual host nodes constitutes the set of nodes in the training topology graph. Then, according to the key node order constraint recorded in the target topology parameter set by phase, in the same teaching phase, according to the given order, connect the virtual host nodes corresponding to adjacent key semantic nodes with a directed connection to form the operation order link within the phase. Then, according to the cross-phase dependency relationship recorded in the target topology parameter set, connect the virtual host nodes corresponding to key semantic nodes with a directed connection across phases to form the dependency link between phases. Through the above node instantiation and connection construction process, the training topology graph is obtained, which serves as a unified structure carrier for a series of subsequent parameter calculations.

[0096] Step S32: construct a sequence pair set according to the key semantic node order constraint, count the number of sequence matches and the number of sequence violations in the candidate training topology graph, and obtain the semantic sequence agreement index through logical function conversion.

[0097] In order to evaluate to what extent the training topology graph follows the teaching order given in the target topology parameter set, it is necessary to find an observation object that can directly correspond to whether the teaching order is respected. In the target topology parameter set, the order of the key semantic nodes within each teaching phase is given. By pairing the two key semantic nodes adjacent to each other in each phase, a number of sequence pairs are obtained, and the sequence pair set completely records the adjacent order relationship that the instructional designer hopes to maintain within the phase.

[0098] After the sequence pair set is constructed, the path of each sequence pair in the training topology graph is checked. Starting from the virtual host node mapped from the first key semantic node in the sequence pair, search along the directed connection to see if it can reach the virtual host node mapped from the second key semantic node in the sequence pair, and check whether the phase identifier of each intermediate node in the path is not earlier than the phase identifier of the first key semantic node. If there is at least one directed path that meets the phase order requirement, then this sequence pair is recorded as a sequence match, indicating that the training topology graph maintains the adjacent relationship set by the instructional designer. If there is no directed path from the first node to the second node, or the existing path has a phase rollback at the intermediate node, then this sequence pair is recorded as a sequence violation, indicating that the training topology graph destroys the design order at this point. Through the above process, three sets of statistical data, the number of sequence matches, the number of sequence violations, and the total number of sequence pairs, are obtained.

[0099] After the number of sequential matches and the number of sequential violations are obtained, a parameter that can reflect whether the overall phenomenon is biased towards matching or violation is needed to be constructed. The processing logic is as follows: first, subtract the number of sequential violations from the number of sequential matches to form a difference value that can be positive or negative. If the number of sequential matches in the training topology is significantly greater than the number of sequential violations, the difference value is positive and the value is large. If the number of sequential violations is dominant, the difference value is negative and the absolute value of the value is large. Then, add a fixed constant to the total number of sequences as the normalization denominator, and divide the above difference value by the denominator to obtain a sequential difference value that falls within the range of -1 to 1. Finally, in order to describe the degree of sequential compliance on a unified scale of zero to one, the sequential difference value is input into a monotonically increasing logic function. Through this function with S-shaped characteristics, the input interval of -1 to 1 is mapped to the output interval of 0 to 1. The closer the difference value is to the positive direction, the closer the output is to 1. The closer the difference value is to the negative direction, the closer the output is to 0. Thus, the semantic sequence agreement index is obtained. This dimensionless index directly reflects the overall compliance of the training topology to the target topology parameter set in terms of sequential design.

[0100] Step S33: For each key semantic node, find the nearest starting node from the set of starting nodes in the teaching semantic graph atlas corresponding to the virtual host node in the candidate practical training topology, and obtain the shortest path length between the two. Calculate the average path length of all key semantic nodes to obtain the average path length of the key semantic nodes.

[0101] When evaluating whether the training topology is suitable as a teaching environment, it is not sufficient to only examine whether the sequence is consistent. It is also necessary to examine the complexity that students experience in terms of path depth and branch number when performing operations at key positions. Therefore, a parameter that reflects the complexity of the operation path around the set of key semantic nodes is needed.

[0102] First, select any node in the starting node set as the starting point of the path and any node in the ending node set as the ending point of the path from the set of starting nodes and the set of ending nodes obtained in step S2. Enumerate all directed paths in the training topology from the starting point to the ending point without repeating nodes, while limiting the length of each path to not exceed the pre-set maximum length, with the path length measured by the number of virtual host nodes contained in the path. Through this process, the global operation path distribution from entering the experiment to ending the experiment in the candidate practical training topology can be obtained, providing a background for the next step of statistical key node position in the path.

[0103] Then for each key semantic node in the key semantic node set, find the corresponding virtual host node in the training topology graph, calculate the shortest path length between the starting node in the starting node set closest to it and the virtual host node, add the shortest path lengths of all key semantic nodes and divide by the number of key semantic nodes to obtain the key node average path length parameter. The key node average path length describes how deep students generally need to go along the operation path to reach each key teaching node. The higher the average value, the greater the vertical depth of the overall teaching process.

[0104] Step S34: In the key semantic node set, filter out target key semantic nodes with successor directed connections, and count the out-degree values thereof; average the out-degree values of all key semantic nodes to obtain the average number of branches of the key semantic nodes.

[0105] After obtaining the vertical depth, the horizontal breadth of the branches after the key nodes also needs to be examined. In the key semantic node set, filter out key semantic nodes with successor directed connections, count the respective numbers of successor connections of the virtual host nodes corresponding to these nodes, and then add these numbers and divide by the number of key semantic nodes participating in the statistics to obtain the key node average branch number parameter. The key node average branch number reflects the average number of branches in different directions after the completion of the key operation. The higher the average value, the more complex the choices and decisions students face after the key nodes.

[0106] Step S35: Integrate the average path length and the average branch number to obtain the operation path complexity index.

[0107] After obtaining the key node average path length and the key node average branch number, in order to form a single complexity index that can be used in subsequent combination with the order index, the operation path complexity index is introduced. The specific logic is as follows: first, multiply the key node average path length by the key node average branch number to make the vertical depth and the horizontal breadth of the two complexity factors jointly reflect, then add a constant to the product to prevent the result from being zero, and then use natural logarithm operation to scale and compress. Natural logarithm operation will make the index grow significantly when the complexity is low, and the growth rate will gradually slow down as the complexity continues to rise, so that the complexity change trend is preserved, and the numerical exaggeration of the extreme complexity state is limited. The operation result is the operation path complexity index, which is used as an input to measure the overall operation difficulty in the subsequent balance analysis.

[0108] Step S36: Convert the operation path complexity index into an ideal semantic order value in the form of exponential decay with a natural constant as the base, obtain the offset between the semantic order fitting index and the ideal semantic order value, and convert the offset degree into a score, denoted as the topology adaptation coefficient.

[0109] If the operation complexity is not considered, only the higher order of the sequence fitting index is pursued, and the complete order is still required when the operation is very complex, which is not reasonable for students. Conversely, if only the order requirement is reduced without considering the operation complexity, it is also difficult to ensure that the teaching goal is fully achieved. Therefore, a semantic order balance relationship is needed to associate the operation path complexity index with an expected ideal order level.

[0110] The construction logic of the semantic order balance relationship is to output an ideal semantic order value between zero and one in the form of an exponential decay with a natural constant as the base, using the operation path complexity index as input. When the operation path complexity index value is low, the output of the exponential decay function is close to one, indicating that the training topology graph is expected to follow the teaching order highly in the scenario where the operation is relatively simple. When the operation path complexity index value increases, the output of the exponential decay function gradually approaches zero, indicating that the order requirement can be relaxed moderately in the scenario where the operation is increasingly complex, allowing local adjustment while ensuring that the core order is not destroyed.

[0111] After the ideal semantic order value is given, the semantic order deviation is introduced, and the absolute value of the difference between the actual semantic order fitting index and the ideal semantic order value is obtained, which is a value directly reflecting how much the current order state deviates from the ideal state. The closer the semantic order deviation is to zero, the more the actual order performance conforms to the expectation given by the balance relationship, and the larger the deviation, the farther the combined state of the training topology graph in the order and complexity dimensions deviates from the teaching expectation.

[0112] In order to further convert this deviation into the overall adaptation level, the topology adaptation coefficient is introduced, and the semantic order deviation is input into the natural exponential decay function as the independent variable. When the semantic order deviation is zero, the output of the exponential function is close to one, and the topology adaptation coefficient is close to full score, indicating that the training topology graph has achieved the ideal balance in terms of order and complexity. When the semantic order deviation increases, the output of the exponential function approaches zero, and the topology adaptation coefficient decreases, indicating that the degree of deviation of the training topology graph from the balance relationship increases. Therefore, the topology adaptation coefficient becomes the core index that comprehensively reflects whether the teaching order and the operation complexity are coordinated.

[0113] Step S37: Determine the size relationship between the topology adaptation coefficient and the preset topology adaptation threshold interval. If it is located within the interval, select the global simulation verification process, if it is higher than the topology adaptation high threshold, select the local critical path simulation verification process, if it is lower than the topology adaptation low threshold, select the basic connectivity simulation verification process, and generate a simulation verification result according to the execution result of the simulation verification process.

[0114] After the topology adaptation coefficient has been obtained, different intensity of simulation verification processes need to be arranged according to different adaptation levels, so that the computing resources are concentrated on the topology structure that may have problems, and the well-adapted topology can quickly pass the check. Therefore, a topology adaptation high threshold and a topology adaptation low threshold are introduced, both of which are between zero and one, and the low threshold is less than the high threshold.

[0115] The acquisition logic of the topology adaptation high threshold and the topology adaptation low threshold is as follows: a large number of training topology samples are collected in multiple courses or projects, and the topology adaptation coefficient of each sample is calculated according to the above steps. At the same time, personnel with teaching experience are organized to evaluate each sample in two dimensions of teaching order rationality and operation complexity suitability, for example, divided into high, medium and low levels. Then, the topology adaptation coefficients of all samples are sorted from low to high, and the distribution positions of samples of different levels in the sorting sequence are observed. For the coefficient value interval that appears in the high-level sample set, the lower boundary of the interval is selected as the topology adaptation high threshold, so that the topology with a topology adaptation coefficient not lower than the threshold basically corresponds to the high level in teaching performance. For the coefficient interval that appears in the low-level sample set, the upper boundary of the interval is selected as the topology adaptation low threshold, so that the topology with a topology adaptation coefficient not higher than the threshold basically corresponds to the low level in teaching performance.

[0116] After the threshold has been set, when the topology adaptation coefficient of a training topology graph is higher than the topology adaptation high threshold, it is regarded as a high-adaptation topology, and only the local structure containing all key semantic nodes and adjacent nodes is executed for fast simulation verification, focusing on checking the connectivity and resource role configuration on the key path. When the topology adaptation coefficient is between the topology adaptation low threshold and the topology adaptation high threshold, it is regarded as a medium-adaptation topology, and a complete simulation verification process covering all key semantic nodes and main operation paths is executed, and whether the function is correct and whether the resource use is within an acceptable range is recorded. When the topology adaptation coefficient is lower than the topology adaptation low threshold, it is regarded as a low-adaptation topology, and only basic connectivity simulation verification is executed, and the path structure with obvious sequence violation or serious operation complexity deviation is recorded as the focus of attention in subsequent adjustment. The observation results in all simulation processes are arranged as simulation verification results, which are output together with the corresponding topology adaptation coefficient, providing specific basis for the adjustment of the teaching semantic graph and the target topology parameter set in the next step.

[0117] Step S3 takes the target topology parameter set as the starting point, and specificizes the teaching semantic constraints into the training topology graph. The semantic sequence coincidence index and the operation path complexity index are constructed through the sequential pair set and the key path structure two paving processes. Then the ideal semantic sequence value and the semantic sequence deviation are derived by using the semantic sequence balance relationship. Finally, the topology adaptation coefficient which uniformly measures the coordination degree of teaching sequence and operation complexity is formed. The simulation verification process is controlled by the topology adaptation high threshold and the topology adaptation low threshold. The simulation verification result which contains the structure performance and the teaching adaptation information is output. A complete and clear quantitative basis is established for the subsequent configuration iteration optimization.

[0118] Steps S1-S3 have converted the teaching configuration information input by the teacher into the teaching semantic graph, the key semantic node set and the target topology parameter set in turn. The training topology graph is generated under the constraint of the target topology parameter set. The semantic sequence coincidence index and the operation path complexity index are calculated. The topology adaptation coefficient is obtained through the semantic sequence balance relationship. Then the simulation verification process with corresponding intensity is executed under different adaptation levels, and the simulation verification result is obtained. The above result gives the quantitative evaluation of the candidate virtual simulation teaching environment in terms of structure rationality and running performance. However, if we only stay at the evaluation level, we cannot form a closed-loop optimization around the user configuration. Therefore, in step S4, we need to construct a set of statistical parameters with clear meaning at the semantic node level and the topology parameter level from the simulation verification result and the topology adaptation coefficient. The semantic nodes and topology parameters that need to be adjusted in this round are selected. On this basis, the teaching semantic graph and the target topology parameter set are rewritten to generate the updated teaching configuration information set, providing the configuration basis after one round of learning for the next round of virtual simulation teaching environment generation.

[0119] Step S4 specifically includes the following sub-steps:

[0120] Step S41: According to the simulation verification result, the failure times and the participation times of each semantic node are counted to obtain the semantic node failure proportion parameter. The semantic node failure proportion parameter is multiplied by the key degree of the semantic node to obtain the node adjustment priority parameter. The semantic nodes with node adjustment priority parameter not lower than the node adjustment threshold value constitute the semantic node adjustment set.

[0121] In the simulation verification phase, the virtual host nodes corresponding to the key semantic nodes in the training topology graph will be activated multiple times in multiple rounds of simulation. Some rounds run smoothly, and some rounds will have connectivity failure, teaching sequence violation or resource anomaly near a semantic node. To judge the concentration degree of the problem at the semantic node level, a unified failure statistics method needs to be established for each key semantic node.

[0122] Each semantic node in the critical semantic node set, first, count the total number of fault events directly related to this semantic node, and count all three types of events, connectivity failure, sequence violation and resource exception, into this total number; then count the number of rounds in which this semantic node actually participates in verification in all simulation rounds, that is, as long as the simulation path contains this semantic node, the corresponding round is counted into the participation number. After obtaining the total number of fault events and the number of participating rounds, in order to eliminate the amplification effect of the absolute value of the fault number caused by the increase of the participation number alone, the semantic node fault proportion parameter is introduced. The semantic node fault proportion parameter takes the total number of fault events as the numerator and one plus the number of participating rounds as the denominator for division operation, thereby obtaining a proportion value ranging from close to zero to one. When the fault proportion parameter is close to zero, it means that this semantic node rarely triggers a fault when participating in simulation; when the fault proportion parameter is close to one, it means that this semantic node almost always accompanies a fault every time it participates in simulation, thereby providing a direct basis for subsequent judgment of whether the semantic node needs to be adjusted.

[0123] Step S2 has calculated the criticality of each semantic node in the teaching semantic graph. The criticality of the node combines the structural bridging effect and the path coverage effect, and the higher the value, the more critical the position of this semantic node in the teaching structure. Step S4 needs to introduce the semantic node fault proportion parameter on the basis of the node criticality, so as to combine the two angles of structural importance and problem concentration into a quantity reflecting the adjustment urgency. For this purpose, the node adjustment priority parameter is introduced as a comprehensive evaluation index at the semantic node level. The construction logic of the node adjustment priority parameter is: first, multiply the criticality value of the node by one plus the value of the semantic node fault proportion parameter, so as to superimpose the structural importance and the fault frequency in the value. The greater the product, the more important and often the semantic node is wrong. Then take the square root of the product, and compress the value range through the square root operation, so that the abnormally high value of the few semantic nodes will not suppress the differences of other nodes. The result obtained after taking the square root is taken as the value of the node adjustment priority parameter, which is used to represent the priority of each semantic node in the current round of optimization.

[0124] After the node adjustment priority parameter of all semantic nodes is calculated, a threshold is needed to determine a part of semantic nodes with the highest priority as the actual adjustment object in this round. For this purpose, the concept of node adjustment threshold is introduced. The value range of node adjustment threshold is limited between the minimum and maximum of the node adjustment priority parameters calculated in this round. The logic of obtaining node adjustment threshold can be determined by combining historical data and expert experience, for example, collect the training topology graph and the corresponding simulation record of several rounds of optimization in multiple course projects, calculate the node adjustment priority parameter of all semantic nodes in each round, and let the teaching experts mark the semantic nodes that are actually adjusted at that time, then sort the semantic nodes according to the node adjustment priority parameter from small to large, observe the concentrated position of the semantic nodes marked by experts as actually adjusted in the sorting sequence, and select the lower boundary of the value interval covering most of the marked nodes as the node adjustment threshold. In actual operation, if the node adjustment priority parameter of a semantic node is not lower than the node adjustment threshold, the semantic node is put into the semantic node adjustment set, thereby obtaining a set of semantic nodes that are important in structure and in the problem set in simulation, providing target objects for subsequent targeted semantic layer adjustment.

[0125] Step S42: According to the simulation verification result, the conflict frequency and the use frequency of each topology parameter are counted to obtain a topology parameter conflict ratio parameter. The topology parameter conflict ratio parameter is multiplied by the comprehensive score result to obtain a parameter adjustment intensity parameter. The topology parameters with a parameter adjustment intensity parameter not lower than a parameter adjustment threshold value are selected to form a topology parameter adjustment set.

[0126] Each topology parameter in the target topology parameter set is used to describe a certain type of constraint content, which can be the order constraint of a key semantic node within a certain stage, the dependency connection constraint between two stages, or the binding constraint between a certain role and a virtual resource capability. When recording fault events, the simulation verification result records the topology parameters directly associated with the events. Through these records, the problem concentration degree can be counted at the parameter level. For each topology parameter in the target topology parameter set, first, the total number of events caused by the topology parameter in all simulation rounds is counted, and the events that can be traced back to the topology parameter among connectivity disruption, order violation, and resource anomaly are all counted as parameter conflict frequency. Second, the number of times the topology parameter is actually called or participates in resource instantiation in all simulation rounds is counted, and each round of participation in configuration calculation or resource allocation is counted as parameter use frequency. To obtain the relative conflict frequency at the parameter level, the topology parameter conflict ratio parameter is introduced. The calculation method of the topology parameter conflict ratio parameter is as follows: the parameter conflict frequency is used as the numerator, and the sum of the parameter use frequency is used as the denominator for division operation, thereby obtaining a ratio value ranging from close to zero to one. The closer the ratio value is to one, the higher the probability of conflict of the topology parameter in participating in simulation.

[0127] The topology parameter conflict proportion parameter alone cannot reflect the influence of the overall adaptation level, because when the topology adaptation coefficient is already high, the conflict proportion of individual parameters can be adjusted mildly, and when the topology adaptation coefficient is low, the parameters with a significant conflict proportion need to be adjusted with strong intensity. Therefore, a parameter adjustment intensity parameter is introduced. The construction logic of the parameter adjustment intensity parameter is as follows: first, a magnification factor is calculated according to the topology adaptation coefficient obtained in step three. The magnification factor is constructed by a natural exponential function, and the exponential part of the natural exponential function takes the value of one minus the topology adaptation coefficient. The lower the topology adaptation coefficient, the higher the value of one minus the topology adaptation coefficient, and the larger the magnification factor. The closer the topology adaptation coefficient is to one, the smaller the value of one minus the topology adaptation coefficient, and the closer the magnification factor is to the initial value of the natural exponential function. Then, the value of the topology parameter conflict proportion parameter is multiplied by the magnification factor, and the result is taken as the value of the parameter adjustment intensity parameter. In this way, when the topology adaptation coefficient is low, the magnification factor is increased, which significantly magnifies the conflict proportion value, making the parameter adjustment intensity parameter more sensitive to the problem; when the topology adaptation coefficient is close to one, the magnification factor is close to a constant, and the parameter adjustment intensity parameter mainly reflects the local conflict proportion.

[0128] After the parameter adjustment intensity parameters of all topology parameters are calculated, the set of topology parameters that actually need to be adjusted in this round needs to be divided, so a parameter adjustment threshold is introduced. The value range of the parameter adjustment threshold is limited between the minimum and maximum values of all parameter adjustment intensity parameters in this round. The parameter adjustment threshold can be obtained based on historical optimization records, for example, in multiple completed deployment and optimization processes, the parameter adjustment intensity parameter is calculated back for each topology parameter participating in adjustment, all parameter adjustment intensity parameters are sorted in ascending order of value, the values corresponding to the topology parameters that were actually modified in history are marked, the centralized position of these marked values in the sorting sequence is observed, and the lower boundary of the interval covering most of the marked values is selected as the parameter adjustment threshold. In actual operation, as long as the parameter adjustment intensity parameter of a topology parameter is not lower than the parameter adjustment threshold, this topology parameter is included in the topology parameter adjustment set, so that a set of constraint items that need to be adjusted preferentially under the current overall adaptation level is obtained.

[0129] Step S43: Based on the semantic node adjustment set and the topology parameter adjustment set, the configuration values in the key semantic node attributes and the target topology parameter set are adjusted bidirectionally to generate updated teaching configuration information and store it.

[0130] After the semantic node adjustment set and the topology parameter adjustment set are determined, the semantic node adjustment set is processed one by one. For each semantic node in the semantic node adjustment set, the corresponding teaching stage identifier, operation type identifier, predecessor semantic node set, successor semantic node set, and predecessor connection type and successor connection type are read from the teaching semantic graph, and the stage constraint, order constraint, and role constraint associated with the semantic node are located from the target topology parameter set. In the simulation verification result, all fault records in which the semantic node is the fault source or a component of the fault path are summarized, and the number of times of each of the three categories of order violation, resource exception, and dependency exception is counted. The fault category with the highest number of times is determined as the dominant fault type of the semantic node. When the dominant fault type is order violation, according to the simulation rounds in which there is no order violation in the simulation verification result, the teaching stage identifiers in which the semantic node is actually located are counted, the teaching stage identifier with the highest number of occurrences is selected as the new stage identifier, and the new stage identifier is written into the teaching semantic graph. The order constraint in the stage in which the semantic node is located in the target topology parameter set is updated to the order arrangement based on the new stage identifier, and the stage attribution information of the predecessor connection and the successor connection is adjusted to be consistent with the new stage identifier. When the dominant fault type is resource exception, according to the resource occupation recorded in the simulation verification result, the number of resource exceptions of the preset resource role associated with the semantic node in all simulation rounds is counted, the preset resource role with the lowest number of resource exceptions is selected as the new resource role, the operation type identifier of the semantic node is mapped to the new resource role label, and the new resource role is written into the teaching semantic graph. The original role constraint is replaced by the new role constraint in the target topology parameter set, and the stage constraint and the order constraint remain unchanged. When the dominant fault type is dependency exception, the original long-chain dependency connection between the semantic node and the dependency-related semantic node is deleted in the teaching semantic graph, a new buffer semantic node is inserted in the middle of the original dependency path, the original long-chain dependency is split into two short-chain dependencies, the stage identifier of the new buffer semantic node is set to a transition stage between the stage identifiers of the two ends of the original dependency, and two new dependency connection constraints are added in the target topology parameter set.

[0131] For each topology parameter in the topology parameter adjustment set, first determine its type in the target topology parameter set, determine one of the three types of stage internal order constraints, cross-stage dependency constraints or role binding constraints. When the topology parameter type is the stage internal order constraint, extract the actual execution order of the key semantic nodes in the stage in different simulation rounds from the simulation verification results, count the occurrence frequency of each type of execution order in the simulation rounds verified by simulation, and take the execution order with the highest occurrence frequency as the new stage internal order. Replace the original order constraint with the execution order in the target topology parameter set, and reconnect the order connection relationship between the semantic nodes in the stage according to the new order in the teaching semantic graph. When the topology parameter type is the role binding constraint, count the number of resource shortage events and resource overload events of the virtual host node involved in the role binding constraint in all simulation rounds from the simulation verification results. If the number of resource shortage events is higher than the preset proportion threshold, increase the resource capability level of the role corresponding to the target topology parameter set and add a set of resource configuration parameters. If the number of resource overload events is higher than the preset proportion threshold, introduce multiple resource instance identifiers in the target topology parameter set, split the original single role binding relationship into several role binding relationships with the same role but different resource instances, and add corresponding semantic nodes for the new resource instances and equivalent order connection relationships with the original semantic nodes in the teaching semantic graph. When the topology parameter type is the cross-stage dependency constraint, count the dependency abnormal events related to the dependency constraint from the simulation verification results. If the dependency constraint is marked as a dependency link abnormal source in continuous multiple simulation rounds, shorten the original cross-stage dependency span to adjacent stage dependency in the target topology parameter set, and add a supplementary dependency constraint from the intermediate stage to the target stage, and insert the semantic nodes of the corresponding intermediate stage and the new dependency connection relationship between the semantic nodes of the previous and next stages in the teaching semantic graph.

[0132] The preset proportion threshold is a dimensionless proportion value between zero and one, which is used to represent the proportion of a role binding constraint triggering resource shortage or resource overload events in the simulation rounds. The acquisition logic is as follows: based on multiple historical simulation samples, count the ratio of resource shortage event number to simulation number and the ratio of resource overload event number to simulation number for each role binding constraint, sort the above ratios from small to large, combine the evaluation results of the teaching and operation personnel on the resource state of each sample, determine the distribution interval of the sample set evaluated as unacceptable resource state in the sorting sequence, and determine the lower boundary of the proportion interval covering the unacceptable sample set as the preset proportion threshold, which is used as the judgment basis for subsequent judgment of whether resource shortage or resource overload reaches the need to trigger role binding constraint adjustment.

[0133] After the semantic node adjustment set and the topology parameter adjustment set are all processed, the updated teaching semantic graph and the updated target topology parameter set are obtained. Then, according to the mapping relationship in step one and step two, the semantic node order and the stage structure in the updated teaching semantic graph are reorganized into a course configuration paragraph editable by a teacher, and the stage constraint, the order constraint, the dependency constraint and the role constraint in the updated target topology parameter set are converted into a resource configuration field and a topology constraint field, combined to form an updated teaching configuration information set, written into a configuration repository as an input configuration source for generating a virtual simulation teaching environment based on the teaching configuration information in the next round.

[0134] Step S4 surrounds the simulation verification result and the topology adaptation coefficient, and constructs a semantic node failure ratio parameter through a failure event count and a participation round number at a semantic node level, and then derives a node adjustment priority parameter in combination with the key degree of the node, and screens out a semantic node adjustment set through a node adjustment threshold value. At a topology parameter level, a topology parameter conflict ratio parameter is constructed through a conflict event count and a use round number, and then a parameter adjustment strength parameter is constructed in combination with the topology adaptation coefficient, and a topology parameter adjustment set is screened out through a parameter adjustment threshold value. Then, the stage identifier, the operation type identifier and the connection type of the teaching semantic graph are adjusted by the semantic node adjustment set, the stage constraint, the order constraint and the role constraint of the target topology parameter set are adjusted by the topology parameter adjustment set, and the updated semantic structure and parameter structure are re-generated to form an updated teaching configuration information set, thereby completing the closed loop from configuration generation, simulation evaluation to configuration correction, and providing an accurate and operable technical basis for continuous optimization of the virtual simulation teaching environment.

[0135] Embodiment 2

[0136] The present application provides a computer device, comprising a processor and a memory; wherein the processor implements the steps of the adaptive logical resource configuration method based on the simulation teaching system when executing the computer program saved in the memory.

[0137] The more specific process of the above method can refer to the corresponding content disclosed in the foregoing embodiments, which will not be repeated here.

[0138] Embodiment 3

[0139] The present application provides a computer readable storage medium for storing a computer program; the computer program is executed by a processor to implement the steps of the adaptive logical resource configuration method based on the simulation teaching system.

[0140] The more specific process of the above method can refer to the corresponding content disclosed in the foregoing embodiments, which will not be repeated here.

[0141] The various embodiments are described in the specification in a progressive manner, each embodiment focusing on different aspects of the other embodiments, and the same or similar parts between the various embodiments can be mutually referred to. For the devices and storage media disclosed by the embodiments, since they correspond to the methods disclosed by the embodiments, the description is relatively simple, and the relevant parts can be referred to the method part.

[0142] Those skilled in the art can clearly understand that the technology in the embodiments of the present application can be realized by means of software and necessary general hardware platforms. Based on such understanding, the technical solutions in the embodiments of the present application can be embodied in the form of a software product, which can be stored in a storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the methods described in the various embodiments or some parts of the embodiments of the present application.

[0143] The above is only the preferred embodiment of the present application, and the protection scope of the present application is not limited to the above-mentioned embodiments. Any technical solution falling within the concept of the present application shall fall within the protection scope of the present application. It should be noted that, for those skilled in the art, some improvements and refinements without departing from the principles of the present application shall be considered as falling within the protection scope of the present application.

Claims

1. An adaptive logic resource configuration method based on a simulation teaching system, characterized in that, Comprise: Step S1: according to the preset teaching semantic rules collected teaching configuration information is parsed into teaching semantic atlas; Step S2: from the teaching semantic atlas, and based on the connection between the key semantic nodes and the pre-defined virtual resource role label, the target topology parameter set containing role constraint information and sequence constraint information is constructed; Step S3: according to the target topology parameter set in the virtual resource library generates candidate training topology graph, and the generated candidate training topology graph is comprehensively scored from the teaching sequence and operation complexity; Select the simulation verification process corresponding to the comprehensive score result for verification, and obtain the simulation verification result; Step S4: based on the comprehensive score result and the simulation verification result, the candidate training topology graph is diagnosed and the configuration value in the target topology parameter set is adjusted according to the diagnosis result to generate updated teaching configuration information and store; The step S1, specifically includes: Collecting teaching configuration information from the simulation teaching system and cutting to form a semantic fragment set; The semantic fragment set is searched, and the words or symbols containing sequence prompt meaning are classified into sequence mark set; The semantic fragments containing condition words, trigger words or result words are classified into dependency expression set, and the semantic fragments satisfying the same object condition are combined into semantic fragment pair set; According to the richness of the sequence mark set relative to the number of semantic fragments, the sequence clue density parameter is constructed, and according to the concentration degree of the dependency expression set in the semantic fragment pair set, the dependency trigger correlation degree parameter is constructed; Based on the sequence clue density parameter and the dependency trigger correlation degree parameter, the semantic connection balance index is constructed, and after comparing with the preset threshold, the sequence connection and dependency connection are constructed between the semantic nodes corresponding to the semantic fragment set, and the teaching semantic atlas is generated; The step S2, from the teaching semantic atlas, specifically for: Read the semantic node set and connection relationship set in the teaching semantic atlas, and for each semantic node, count the in-degree and out-degree, add the product of the in-degree and out-degree to the set constant, and then perform natural logarithm operation to obtain the structure bridge parameter; Enumerate all simple paths between the starting nodes and the ending nodes in the teaching semantic atlas, and count the participation proportion of each semantic node in the simple path to generate the path participation parameter; Based on the structure bridge parameter and the path participation parameter, the key degree of each semantic node is obtained, and based on the preset key degree threshold, the key semantic node set is obtained.

2. The method for adaptive logical resource configuration based on simulation teaching system according to claim 1, characterized in that, The step S2, based on the connection between the key semantic nodes and the pre-defined virtual resource role label, the target topology parameter set containing role constraint information and sequence constraint information is constructed, specifically: Read the operation type identifier of each key semantic node, map different operation type identifiers to pre-defined virtual resource role labels to form key node resource role constraint information; Reading the phase identifier of each key semantic node, dividing the key semantic nodes with the same phase identifier into the same phase group, and then sorting the key semantic nodes in each phase group according to the sequential connection in the connection relationship to obtain the key node sequence constraint information in stages; Merging the key node sequence constraint information of all phase groups and the corresponding resource role constraint information to form the target topology parameter set.

3. The method for adaptive logical resource configuration based on simulation teaching system as claimed in claim 1, wherein, The step S3 comprises: Generating the candidate practical training topology graph in the virtual resource library according to the target topology parameter set, and comprehensively scoring the generated candidate practical training topology graph from the teaching sequence and the operation complexity, specifically including: Generating the candidate practical training topology graph under the constraint of the target topology parameter set; According to the key semantic node sequence constraint, constructing a sequence pair set, and counting the number of sequence matches and the number of sequence violations in the candidate practical training topology graph, and then converting the semantic sequence matching index through a logical function; For each key semantic node in the corresponding virtual host node of the candidate practical training topology graph, finding the nearest starting node from the starting node set of the teaching semantic graph, and obtaining the shortest path length of the two; and averaging the shortest path lengths of all key semantic nodes to obtain the average path length of the key semantic nodes; In the key semantic node set, filtering out the target key semantic node with a successor directed connection, and counting the out-degree value; and averaging the out-degree values of all key semantic nodes to obtain the average branch number of the key semantic nodes; Comprehensively averaging the average path length and the average branch number to obtain the operation path complexity index; 4. The method for adaptive logical resource configuration based on simulation teaching system according to claim 3, characterized in that, Converting the operation path complexity index into an ideal semantic sequence value in the form of exponential decay with a natural constant as the base, obtaining the offset between the semantic sequence matching index and the ideal semantic sequence value, and converting the offset degree into a score, which is denoted as the topology adaptation coefficient. The step S3 comprises:

5. The method for adaptive logical resource configuration based on simulation teaching system as claimed in claim 1, wherein, Judging the size relationship between the topology adaptation coefficient and the preset topology adaptation threshold interval, if it is located in the interval, selecting the global simulation verification process, if it is higher than the topology adaptation high threshold, selecting the local key path simulation verification process, if it is lower than the topology adaptation low threshold, selecting the basic connectivity simulation verification process, and generating the simulation verification result according to the execution result of the simulation verification process. The step S4 comprises: According to the simulation verification result, counting the failure times and participation times of each semantic node to obtain the semantic node failure proportion parameter, multiplying the semantic node failure proportion parameter with the key degree of the semantic node to obtain the node adjustment priority parameter, and selecting the semantic nodes with the node adjustment priority parameter not lower than the node adjustment threshold to form the semantic node adjustment set; According to the simulation verification result, counting the conflict times and use times of each topology parameter to obtain the topology parameter conflict proportion parameter, multiplying the topology parameter conflict proportion parameter with the comprehensive score to obtain the parameter adjustment intensity parameter, and selecting the topology parameters with the parameter adjustment intensity parameter not lower than the parameter adjustment threshold to form the topology parameter adjustment set; Based on the semantic node adjustment set and the topology parameter adjustment set, the configuration values in the key semantic node attributes and the target topology parameter set are adjusted in a bidirectional synchronization manner to generate updated teaching configuration information and store it.

6. The method for adaptive logical resource configuration based on simulation teaching system according to claim 5, characterized in that, The configuration values in the key semantic node attributes and the target topology parameter set are adjusted in a bidirectional synchronization manner based on the semantic node adjustment set and the topology parameter adjustment set to generate updated teaching configuration information and store it, specifically: For each semantic node in the semantic node adjustment set, read its teaching stage identifier, operation type identifier, predecessor semantic node set, successor semantic node set, and predecessor connection type and successor connection type from the teaching semantic graph, and locate the order constraint and role constraint associated with the semantic node from the target topology parameter set; For each semantic node in the semantic node adjustment set, aggregate all fault records in the simulation verification results that have the semantic node as the fault source or fault path component to determine the dominant fault type of the semantic node; According to the dominant fault type, adjust the teaching stage identifier, operation type identifier, and connection type of the semantic node, and synchronously modify the stage constraint, order constraint, and role constraint in the target topology parameter set; For each topology parameter in the topology parameter adjustment set, adjust the role constraint, order constraint, and stage constraint according to the constraint type to which the topology parameter belongs, and synchronously modify the connection relationship of the corresponding semantic node in the teaching semantic graph; According to the mapping rule, generate an updated teaching configuration information set, and store it.

7. A computer device, comprising: A computer program product comprising a computer readable medium storing a computer program, the computer program comprising program instructions configured to cause a processor to perform the steps of the adaptive logical resource configuration method for a simulation-based teaching system according to any one of claims 1-6 when the computer program is executed by the processor.

8. A computer-readable storage medium, characterized in that, A computer program product comprising a computer readable medium storing a computer program, the computer program comprising program instructions configured to cause a processor to perform the steps of the adaptive logical resource configuration method for a simulation-based teaching system according to any one of claims 1-6 when the computer program is executed by the processor.

Images