AI-driven second classroom intelligent management cloud platform

Abstract

The invention discloses an AI-driven second classroom intelligent management cloud platform, and particularly relates to the technical field of college second classroom management. Extracting student identifications, experience identifications, occurrence time periods, participation roles, achievement contents, certification contents, declaration items, declaration sequences, item combination relationships, item mutual exclusion relationships and item precedence relationships, and outputting a unified record set; the same second classroom experience is constructed into an experience sub-graph and an item sub-graph in a knowledge graph, the experience sub-graph and the item sub-graph are further spread into a common piece, an exclusive piece and a linkage piece which can be pressed by different affirmed items in sequence, and unified resolving and result writing-back are performed on the occupation relation, the mutual exclusion relation, the precedence relation and the remaining available relation among multiple items.

Description

AI-Driven Smart Management Cloud Platform for Second Classrooms Technical Field

[0001] This invention relates to the field of extracurricular management technology in higher education institutions, and more specifically, to an AI-driven intelligent management cloud platform for extracurricular activities. Background Technology

[0002] In the management of extracurricular activities in universities, existing platforms typically focus on integrating activity participation, achievement submission, credit conversion, and evaluation applications into a unified management chain. A common approach is to record, collect, verify, and write the results of experiences such as competition awards, volunteer service, club positions, practical projects, and training lectures according to activity type, recognition criteria, and review process. However, in scenarios where a student's same award might be used simultaneously for innovation and entrepreneurship credit conversion, award and honor application, comprehensive quality record archiving, and college professional development achievement statistics, the platform not only needs to operate continuously under concurrent applications from multiple colleges and across multiple lines, but also needs to meet restrictions on the use, mutual exclusion, order, and revocation / rollback of different recognition criteria. Furthermore, the recognition results must be consistent and available for student querying, administrator review, and direct access for subsequent statistics. However, in this scenario, most existing methods still treat the same extracurricular activity as a single, unrelated event. Experiences are treated as static records that can be read separately by each assessment item. This typically only allows for single-condition comparisons, frequency controls, or column restrictions. It's difficult to handle the cascading effects of an experience being occupied by one item on the availability of other items. Therefore, in practice, situations often arise where the student's end shows multiple items corresponding to the same experience meeting the application requirements, but during the final review, summary, or graduation verification stages, it's discovered that the experience has been occupied by a preceding item, cannot coexist with another item, the calling order is incorrect, or the remaining available portion is insufficient. Ultimately, manual rollback, reassessment, or repeated interpretation is necessary to complete the process. The technical problem this application aims to solve is: how to uniformly constrain the occupancy, mutual exclusion, sequential, and remaining availability relationships of the same extracurricular experience across multiple assessment items within the Extracurricular Smart Management Cloud Platform, so that the experience forms only a unique, consistent, and traceable valid assessment result during the multi-item assessment process. Summary of the Invention

[0003] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide an AI-driven intelligent management cloud platform for the second classroom. This platform constructs the same second classroom experience into an experience subgraph and an event subgraph in a knowledge graph, and further expands it into shared pieces, exclusive pieces, and linked pieces that can be occupied in sequence by different identified events. It performs unified calculation and result writing back on the occupancy relationship, mutual exclusion relationship, sequential relationship, and remaining availability relationship between multiple events, thereby solving the problems mentioned in the background art.

[0004] To achieve the above objectives, this invention provides the following technical solution: an AI-driven intelligent management cloud platform for the second classroom, comprising: a data collection terminal module, used to receive experience records, application records, and event records uploaded by students, teachers, and administrators; extract student identifiers, experience identifiers, occurrence time periods, participating roles, achievement content, proof content, application items, application order, concurrent relationship of events, mutually exclusive relationship of events, and sequential relationship of events; and output a unified record set; and a knowledge graph construction module, used to read the unified record set, using student identifiers, experience identifiers, occurrence time periods, participating roles, achievement content, proof content, and application items as... The system uses nodes with edges representing ownership, pointing, limiting, shared, mutually exclusive, and sequential relationships to construct a second-classroom knowledge graph, outputting experience subgraphs and event subgraphs. The credential expansion module reads the experience subgraph and, based on the limiting, ownership, and pointing edges between the occurrence time node, participating role node, result content node, and proof content node, expands the same second-classroom experience into an experience credential surface composed of shared, exclusive, and linked parts. The slice generation module reads the experience credential surface and event subgraph, slices the experience credential surface into shared slices, exclusive slices, and linked slices, and outputs an occupied slice graph.

[0005] In a preferred embodiment, the system further includes: a occupancy calculation module, used to read the occupancy map and declaration records, push each declaration item node into the corresponding slice according to the declaration order, perform occupancy locking on exclusive slices, perform remaining share deduction on shared slices, perform subsequent availability status rewriting on linked slices, and cut off the occupancy path connected by mutually exclusive edges, and output the item occupancy result map and the remaining availability map; and a result generation module, used to read the item occupancy result map and the remaining availability map, perform conflict cancellation, sequence verification and map rewriting, and output a uniquely corresponding valid identification result and its occupancy trajectory.

[0006] In a preferred embodiment, the data collection terminal module includes: performing source-separated reading of experience records, application records, and event records uploaded by students, teachers, and administrators; extracting experience identifiers, student identifiers, occurrence time periods, participating roles, deliverable content, supporting documentation, application items, application order, concurrent use of events, mutual exclusion of events, and sequential relationship of events according to the type of uploading terminal; merging records from different uploading terminals but with the same experience identifier into the same data collection group; and outputting terminal-separated record groups; performing same-group alignment comparison on the occurrence time periods, participating roles, deliverable content, and supporting documentation in each terminal-separated record group; and identifying records with consistent occurrence time periods. Furthermore, when the content of the results and the content of the proof are mutually referential, an experience master piece is generated. When there is a correspondence between the declared items and the items used together, the items are mutually exclusive, or the items are sequential, an item constraint piece is generated. The experience master piece and the item constraint piece are associated with the experience identifier and written into a unified record set. Missing items are filled and conflicts are marked for each experience master piece and each item constraint piece in the unified record set. When there are multiple participating roles or multiple proof contents under the same experience identifier, the original correspondence is preserved and written into the position identifier in the piece respectively. When the same declared item corresponds to multiple sequential positions, the results are generated in order of upload time. The unified record set is output for the knowledge graph construction module to read.

[0007] In a preferred embodiment, the knowledge graph construction module includes: reading the student identifier, experience identifier, occurrence time, participating role, result content, proof content, and application item of each record in a unified record set; establishing record-to-record edges according to the rules of consistent student identifiers, overlapping occurrence times, consistency between the item name appearing in the result content and the item name appearing in the proof content, participating roles not belonging to conflicting roles defined by item mutual exclusion relationships, and the unique sequential position of the same application item corresponding to one record; grouping records with record-to-record edges into the same candidate node group; and outputting candidate node groups and candidate relationship groups; based on the candidate node groups and candidate relationship groups, for each Candidate node groups are statistically analyzed for time period breaks, role conflicts, result proof mismatches, item sequence reversals, and item usage conflicts. Candidate node groups are then reconstructed in a fixed order of first splitting and then merging. The splitting rule is to separate records containing any conflict from the original candidate node group, and the merging rule is to merge candidate node groups with consistent student identifiers and no conflict. After each round of reconstruction, candidate relationship groups are regenerated based on experience identifier, occurrence time, participating role, result content, proof content, and application item, until the members of the candidate node groups obtained from two consecutive rounds of reconstruction are completely consistent and the edge items of the candidate relationship groups are completely consistent. Stable node sets and stable edge sets are then output.

[0008] In a preferred embodiment, the knowledge graph construction module further includes: based on a stable node set and a stable edge set, performing consistency checks on each edge from the perspectives of time period, role, result proof, and matter, respectively. The time period perspective checks whether the occurrence time corresponds to the limiting relationship; the role perspective checks whether the participating role corresponds to the attribution relationship; the result proof perspective checks whether the result content corresponds to the pointing relationship; and the matter perspective checks whether the declared matter corresponds to the use relationship, mutual exclusion relationship, and sequential relationship. Each edge is processed according to the following rules: if the number of supporting items is greater than the number of counter-proof items, the edge is retained; if the number of supporting items is equal to the number of counter-proof items, the edge is retained and a conflict flag is written; and if the number of supporting items is less than the number of counter-proof items, the edge is deleted. Subsequently, experience subgraphs containing attribution relationships, pointing relationships, and limiting relationships are extracted according to experience identifiers, and matter subgraphs containing use relationships, mutual exclusion relationships, and sequential relationships are extracted according to declared matters.

[0009] In a preferred embodiment, the credential expansion module includes: reading the occurrence time node, participating role node, result content node, and proof content node from the experience subgraph; generating a shared portion according to the rule that the same proof content node is simultaneously connected to two or more result content nodes via pointing edges and falls into the same occurrence time node via limiting edges; generating an exclusive portion according to the rule that the same proof content node is connected to only one result content node via pointing edges and that result content node corresponds to only one participating role node; generating a linked portion according to the rule that the result content node connected to the same proof content node via pointing edges corresponds to multiple participating role nodes and that a change in any participating role node will cause a rewriting of the affiliation relationship of the remaining result content nodes; and outputting an initial credential surface; and recording the corresponding experience identifier, occurrence time node, and participating role node for each shared portion, each exclusive portion, and each linked portion based on the initial credential surface. Role nodes, outcome content nodes, and proof content nodes are rearranged according to the following rules: the same proof content node cannot be written into both shared and exclusive parts simultaneously; the same outcome content node is only assigned to one part under the same occurrence time node; and the same participating role node retains only one attribution chain under the same proof content node. The resulting experience credential surface is then output. Based on the experience credential surface, the limiting edges, attribution edges, and pointing edges of each part are checked item by item. The limiting edges are used to check the correspondence between the proof content node and the occurrence time node, the attribution edges are used to check the correspondence between the outcome content node and the participating role node, and the pointing edges are used to check the correspondence between the proof content node and the outcome content node. Parts where all limiting edges, attribution edges, and pointing edges are completely corresponding are retained in the experience credential surface, while parts lacking any correspondence are deleted from the experience credential surface. The resulting experience credential surface is then output for the slice generation module to read.

[0010] In a preferred embodiment, the slice generation module includes: based on the shared parts, exclusive parts, and linked parts in the experience credential surface, as well as the shared edges, mutually exclusive edges, and sequential edges in the item subgraph, reading the corresponding proof content node, result content node, participating role node, and occurrence time node for each part item by item; cutting the part that has a shared edge connection relationship with two or more application item nodes and whose corresponding proof content node, result content node, participating role node, and occurrence time node are consistent into a shared slice; cutting the part that has a connection relationship with only one application item node and that application item node is separated from the other application item nodes by a mutually exclusive edge into an exclusive slice; cutting the part that has a sequential edge connection relationship with two or more application item nodes and whose availability relationship between the preceding application item node and the subsequent application item node is rewritten into a linked slice; and outputting an initial occupied slice group; and writing an experience identifier, slice identifier, and corresponding application into each shared slice, each exclusive slice, and each linked slice according to the initial occupied slice group. The process involves reordering the following nodes: Item Node, Corresponding Proof Content Node, Corresponding Deliverable Content Node, Corresponding Participating Role Node, and Corresponding Occurrence Time Node. The reordering follows these rules: for the same proof content node under the same occurrence time node, it is assigned to only one shared segment or one exclusive segment; for the same deliverable content node under the same participating role node, it is assigned to only one exclusive segment or one linked segment; and for the same application item node under the same experience identifier, the linked segments are connected end-to-end according to their sequential edge order. This process generates an Occupied Segment Table. Based on this table, the process verifies the edge correspondence between each shared segment, exclusive segment, and linked segment and the item subgraph. Specifically, for shared segments, it verifies whether there are shared edges between all corresponding application item nodes; for exclusive segments, it verifies whether there are mutually exclusive edges between the corresponding application item node and other application item nodes; and for linked segments, it verifies whether there are sequential edges between corresponding application item nodes and whether the connection order matches the segment's internal arrangement order. The complete segments corresponding to the edge items and their interconnections are then written into the Occupied Segment Graph, which is then output for the Occupied Segment Calculation Module to read.

[0011] In a preferred embodiment, the occupancy calculation module includes: establishing a matter piece matrix, piece continuation matrix, mutual exclusion matrix, and sequential matrix based on shared pieces, exclusive pieces, linked pieces, mutual exclusion edges, and sequential edges in the occupancy piece map, as well as the declaration items and declaration order in the declaration records, according to the same experience identifier; continuously multiplying the row vector of the matter piece matrix corresponding to the current declaration item with the piece continuation matrix to generate a candidate piece sequence; deleting candidate piece sequences with non-zero mutual exclusion matrix mapping values, mismatched sequential matrix positions, and duplicate exclusive pieces; and outputting a candidate path table; based on the candidate path table, establishing a path matrix, shared surplus vector, exclusive occupancy vector, and sequential edge matrix for each candidate piece sequence. The dynamic transmission vector is processed, and shared deduction, exclusive locking, and linkage transmission are sequentially performed on the path matrix. The path matrix is ​​then multiplied by the item fragment matrix to obtain the item expansion matrix. Singular value decomposition is performed on the item expansion matrix, and the column space corresponding to non-zero singular values ​​is extracted. Orthogonal projection is performed on the column space to generate standardized path codes. One candidate path with the same standardized path code and consistent shared surplus vector, exclusive occupancy vector, and linkage transmission vector is retained. The above operation is repeated after each application item is written in sequence until the standardized path code set, shared surplus vector set, exclusive occupancy vector set, and linkage transmission vector set obtained in two adjacent rounds are consistent. A stable path table is then output.

[0012] In a preferred embodiment, the occupancy calculation module further includes: constructing an occupancy verification matrix, a surplus verification matrix, and a transfer verification matrix for each stable path based on the stable path table; multiplying the occupancy verification matrix with the item piece matrix to verify the item correspondence; verifying the shared deduction result bit by bit with the shared surplus vector using the surplus verification matrix; verifying the linkage transfer result bit by bit with the linkage transfer vector using the transfer verification matrix; and writing the stable paths that have been verified to be valid into the item occupancy result graph, and writing their corresponding shared surplus vector and linkage transfer vector into the remaining available graph.

[0013] In a preferred embodiment, the result generation module includes: based on the event occupancy result map and the remaining availability map, aggregating the occupancy paths, occupancy sequence, shared reserve, and linkage transmission results corresponding to each declared event according to the same experience identifier; performing conflict cancellation on each occupancy path under the same experience identifier; deleting occupancy paths in the occupancy sequence that have mutually exclusive edge connections, duplicate deductions after writing back the shared reserve, or linkage transmission results that do not correspond to the order of edges; and outputting a pending result group; based on the pending result group, performing a sequential review on each remaining occupancy path under the same experience identifier, and resolving conflicts in each occupancy path. The order of the declared items is aligned with the order of the occupied blocks, and the occupied paths corresponding to each item are retained as valid paths. Paths with reversed order are deleted. Then, valid identification results and occupied trajectories are generated according to the occupied block sequence, shared reserve, and linkage transmission results in the valid paths. Based on the valid identification results and occupied trajectories, the valid identification results are written to the corresponding experience identifier and the result node of the declared item, the occupied trajectories are written to the connection edges between the corresponding occupied blocks, and the shared reserve and linkage transmission results are written back to the corresponding nodes and edges. The unique valid identification results and their occupied trajectories are output.

[0014] The technical effects and advantages of this invention are as follows: 1. By constructing an experience credential surface, an occupancy map, and an occupancy result map, this solution unifies the occupancy relationships, mutual exclusion relationships, sequential relationships, and remaining availability relationships of the same experience across multiple assessment items into a single solution chain, relatively reducing manual rollbacks and reversals caused by prior occupancy and subsequent conflicts during the final review stage; 2. By constructing a unified record set into a knowledge graph containing experience subgraphs and item subgraphs, and performing reconstruction and consistency checks on the corresponding edges of records, scattered records can be transformed into experience nodes and item nodes with clear relationships, relatively improving the problems of unstable cross-terminal record grouping and misaligned item attachments; 3. By expanding the same experience into shared parts, exclusive parts, and linked parts, and then further dividing it into shared parts... The system of identifying fragments, exclusive fragments, and linked fragments transforms the original method of reading entire experiences into one based on callable structures, thereby improving the clarity of call boundaries when multiple items are submitted in parallel. 4. By performing shared deduction, exclusive locking, and linked transfer on fragments according to the order of submission, and by filtering out invalid paths using mutual exclusion matrices, priority matrices, and path retention rules, the system can relatively suppress duplicate deductions, order reversals, and incompatible coexistence of the same experience in multiple item calls. 5. By performing item-specific verification, surplus verification, and transfer verification on stable paths, and then generating item occupancy result diagrams and remaining availability diagrams, the system can output the identification results and remaining availability status synchronously, thereby relatively improving the reusability of status during subsequent review, statistics, and resubmission. Attached Figure Description

[0015] Figure 1 is a schematic diagram of the system module structure of the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Referring to Figure 1 in the specification, the AI-driven intelligent management cloud platform for the second classroom of the present invention includes: a data collection terminal module, used to receive experience records, application records, and event records uploaded by students, teachers, and administrators; extract student identifiers, experience identifiers, occurrence time periods, participating roles, achievement content, proof content, application items, application order, concurrent relationship of events, mutual exclusion relationship of events, and sequential relationship of events; and output a unified record set. In this embodiment, the purpose of the data collection terminal module is to first compress the heterogeneous records uploaded by students, teachers, and administrators into the same field definition, and then organize the experience information and event constraint information corresponding to the same second classroom experience into a unified record set that can be directly read by the subsequent knowledge graph construction module. This process avoids problems such as mixed record sources, unstable experience grouping, and misaligned event constraints during subsequent node mapping. To achieve this, the original uploaded records are first split into columns and fields extracted. Then, experience information within the same collection group is compared to generate experience master pieces and event constraint pieces. Finally, the generated results are filled with missing items, conflict annotations are applied, and the order is adjusted, ensuring the output has a unified field structure, stable attribution relationships, and clear sequential relationships. The implementation process includes the following steps: First, the original records from different upload terminals are converted into terminal-separated record groups with consistent field structures that can be grouped by experience. The input consists of experience records uploaded from student terminals, experience records uploaded from teacher terminals, and experience records uploaded from management terminals. The system includes experience records, application records uploaded by students, application records uploaded by teachers, application records uploaded by administrators, and event records. The processing steps include: for each uploaded record, first writing the upload terminal type and platform receiving time; then extracting the experience identifier, student identifier, time period, participating roles, achievement content, supporting documentation, application item, application order, concurrent application relationship, mutually exclusive application relationship, and sequential application relationship according to a fixed field order. Experience records must retain at least the experience identifier, student identifier, time period, participating roles, achievement content, and supporting documentation; application records must retain at least the experience identifier, student identifier, application item, and application order; and event records must retain at least the event identifier, event name, and concurrent application relationship. The system uses relationships, mutual exclusion relationships, and sequential relationships. When the original record has an experience identifier, it reads the identifier directly. When the original record does not have an experience identifier, it concatenates the student identifier, occurrence date, event name, and result content in sequence to generate a joint experience identifier. Then, records from different upload terminals are merged into the same collection group according to the rule of consistent experience identifiers. Within the same collection group, records are stored separately according to the upload terminal category. The output is the terminal-separated record group, which is written to the collection buffer for the same group's alignment comparison step. The handling of anomalies or missing records is as follows: when a time period is missing, the record is retained but a time period missing mark is written. When the student identifier is missing or the experience identifier cannot be generated, the record is written to the supplementary record table and does not enter the collection group.Secondly, experience master pieces representing factual events and event constraint pieces representing event constraints are extracted from the same collection group, ensuring that experience information and event information are structurally separated before entering the unified record set. The input is terminal-separated record groups. Processing actions include: performing same-group alignment comparison on the occurrence time, participating roles, result content, and proof content in the same collection group. The rule for determining consistent occurrence time is that the start time and end time are the same. Records with overlapping but not completely identical occurrence times are only marked with an overlap marker and are not directly considered consistent. The rule for determining mutual relevance between result content and proof content is that the result content contains proof material number, and the proof content contains result name or experience identifier, with at least two fields corresponding to each other. Consistency is required; when the occurrence time is consistent and the results and proofs are mutually referential, the experience identifier, student identifier, occurrence time, participating role, results, and proofs in the set of records should be written into the main experience segment; when there is a corresponding relationship between the application items and the items, the mutual exclusion relationship between the items, or the sequential relationship between the items, the experience identifier, student identifier, application item, application order, and corresponding item relationship in the set of records should be written into the item constraint segment, where the corresponding judgment rule is that the application item identifier is consistent with the item identifier in the item record; then, the main experience segment and the item constraint segment are associated according to the experience identifier and written into a unified record set, the output being a unified record set containing associated segment items and written into the segment association table for the missing item completion step to read; abnormal or missing The process is as follows: when the content of the deliverable and the content of the proof do not point to each other, no experience master piece is generated and a missing pointer is written; when the declared item has no corresponding item identifier in the item record, no item constraint piece is generated and an item missing pointer is written. Furthermore, field completion, conflict disclosure, and order fixing are performed on the items in the unified record set, ensuring that the records read by the subsequent knowledge graph construction module have complete fields, clear conflicts, and a unique order. The input consists of the unified record set and the item association table. The processing actions include: performing missing item completion on each experience master piece and item constraint piece according to a fixed completion order, with priority given to completing the participating roles and proof content on the teacher's side, priority given to completing the declared item and item relationship on the management side, and priority given to completing the deliverable content on the student's side. The system will perform conflict labeling on the completed segments. Conflict types include time period conflict, role conflict, result proof mismatch conflict, and event sequence conflict. The time period conflict is determined when there are two different time periods with different start and end times under the same experience identifier. The role conflict is determined when the same result content corresponds to two different participating roles. The result proof mismatch conflict is determined when the result content and proof content no longer meet the mutual pointing condition. The event sequence conflict is determined when the same application item has two different application sequences. When there are multiple participating roles or multiple proof contents under the same experience identifier, the original correspondence between each participating role and each proof content will be retained, and they will be written into the segment position identifier according to the platform's receiving time order.When the same application item corresponds to multiple sequential positions, the results are generated in the order of platform receipt time and written back to the item constraint fragments; the output is a unified record set for the knowledge graph construction module to read and written to a unified record storage table; the handling of anomalies or missing items is as follows: for fields that cannot be filled in, a missing mark is retained and written along with the fragment item; for fragment items with conflict marks, they are not deleted, only the conflict code is retained for subsequent disassembly and verification by the knowledge graph construction module; in practical applications: for example, when the same student participates in a school-level competition, the student uploads the award certificate and application item, the teacher uploads the guidance record and participation role, and the management uploads the corresponding credit items, evaluation items, and the relationship between the two. Regarding the sequence of events, the data collection terminal module first aggregates records from the three ends into the same collection group according to experience identifiers. Then, it generates the main experience piece based on the correspondence between certificate number, competition name, and experience identifier. It also generates the event constraint piece based on the correspondence between the application items and the event identifiers in the event records. Subsequently, it supplements the missing certification content from the teacher end, supplements the missing application order from the management end, and writes conflict codes for order conflicts caused by repeated applications for the same competition. Finally, it outputs a unified record set. Through this implementation process, the subsequent knowledge graph construction module can directly build experience subgraphs and event subgraphs according to unified fields and stable grouping relationships, thereby reducing graph construction deviations caused by inconsistent record sources, missing fields, and chaotic application order.

[0018] The knowledge graph construction module reads a unified record set and uses student identifiers, experience identifiers, occurrence time periods, participating roles, outcome content, proof content, and declaration items as nodes, and uses attribution relationships, pointing relationships, limiting relationships, usage relationships, mutual exclusion relationships, and sequence relationships as edges to construct a second-classroom knowledge graph, outputting experience subgraphs and item subgraphs. In this implementation, the purpose of the knowledge graph construction module is to organize the discrete records in the unified record set into a second-classroom knowledge graph that can stably express the factual relationships of experiences and the constraint relationships of items, so that the experience subgraphs and item subgraphs read by the subsequent credential expansion module have clear attribution, clear relationships, and clear conflict states. This implementation process first establishes the corresponding edges of the records and forms candidate node groups. The process involves several steps: First, candidate node groups and candidate relationship groups are formed based on a unified record set. The inputs include student identifiers, experience identifiers, occurrence times, participating roles, outcome content, proof content, and declared items from the unified record set. The processing involves reading the unified record set record by record, establishing corresponding edges based on the following criteria: consistent student identifiers, overlapping occurrence times, consistency between the item name code corresponding to the outcome content and the item name code corresponding to the proof content, participating roles not belonging to the prohibited role set corresponding to the declared item, and only one record corresponding to the same declaration order for the same declared item. The rules stipulate that if two records overlap in their time periods, the start time of one record must be no later than the end time of another record, and the end time must be no earlier than the start time of the other record. The item name code is generated from the item name code table, and the prohibited role set is generated from the applicable role field in the item record. When two records simultaneously meet the above rules, a record-to-relationship edge is established between the two records, and records connected by this edge are grouped into the same candidate node group. All record-to-relationship edges within the candidate node group are then aggregated into a candidate relationship group. The output consists of the candidate node group and the candidate relationship group, which are written to the candidate node group table and the candidate relationship group table respectively for subsequent reconstruction. The handling of exceptions or missing records is as follows: records with missing items in the application only participate in the review process. First, records with inconsistent item names are grouped and written into a name-pending-verification table without corresponding edges. Second, conflicts within candidate node groups are resolved to form stable node sets and stable edge sets. The inputs are candidate node groups and candidate relationship groups. The processing actions are: for each candidate node group, statistics are collected on time period breaks, role conflicts, result proof mismatches, item sequence reversals, and item co-occurrence conflicts. The rule for determining time period breaks is that the occurrence times of two records are neither intersecting nor continuous, and the rule for determining continuity is that the end time of the previous record is the same as the start time of the next record. The rule for determining role conflicts is that the same result content corresponds to two different participating roles, and at least one participating role belongs to the prohibited role set of the corresponding declared item.The criteria for determining a mismatch in achievement verification are: the item name code corresponding to the achievement content does not match the item name code corresponding to the verification content, or the verification content does not simultaneously contain any two of the following: experience identifier, achievement name, and verification material number. The criteria for determining an item sequence reversal are: the two application items corresponding to the two records have a sequence relationship in the item sequence relationship table, and the application order in the records is the opposite of that sequence relationship. The criteria for determining an item concurrent conflict are: the two application items corresponding to the two records do not appear in the item concurrent relationship table. Subsequently, candidate node groups are reconstructed according to a fixed order of first splitting and then merging. During splitting, records containing any conflict item are separated from the original candidate node group. During merging, candidate nodes with consistent student identifiers and without any of the above five types of conflict items are merged. Point group merging; after each round of reconstruction, candidate relationship groups are regenerated according to experience identifier, occurrence time period, participating role, result content, proof content, and declaration item, until the candidate node group members obtained from two consecutive rounds of reconstruction are completely consistent and the candidate relationship group edge items are completely consistent; the output is a stable node set and a stable edge set, which are written to the stable node table and stable edge table respectively for consistency verification; the handling of anomalies or missing records is as follows: candidate node groups consisting of a single record are directly retained, and if multiple groups do not meet the merging conditions, the original group is retained; next, consistency verification is performed on the stable edge set and experience subgraphs and item subgraphs are extracted; the input is a stable node set and a stable edge set; the processing action is: perform a process on each edge from the perspectives of time period, role, result proof, and item. Consistency checks are performed as follows: From a time period perspective, the check verifies whether the occurrence time corresponds to the limiting relationship. The rule for establishing a limiting relationship is that the occurrence date or start and end time period corresponding to the evidence content is consistent with or completely contained within the occurrence time period in the record. From a role perspective, the check verifies whether the participating roles correspond to the attribution relationship. The rule for establishing an attribution relationship is that the executing entity corresponding to the result content is consistent with the participating role. From a result proof perspective, the check verifies whether the result content corresponds to the pointing relationship. The rule for establishing a pointing relationship is that any two of the result name, result number, and experience identifier appear simultaneously in the proof content and are consistent with the result content. From a matter perspective, the check verifies whether the declared matters correspond to the concurrent, mutually exclusive, and sequential relationships. The rule for establishing a concurrent relationship is that the matter pair appears in the matter concurrent relationship table. The rule for a mutually exclusive relationship is that the pair of items does not appear in the mutual exclusion relationship table. The rule for a sequential relationship is that the order of the items being declared is consistent with the order in the sequential relationship table. Each valid viewpoint is recorded as a supporting item, and each invalid viewpoint is recorded as a disproving item. Each edge is processed according to the following rules: if the number of supporting items is greater than the number of disproving items, the edge is retained; if the number of supporting items is equal to the number of disproving items, the edge is retained and a conflict flag is written; if the number of supporting items is less than the number of disproving items, the edge is deleted. Subsequently, experience subgraphs containing attribution, pointing, and limiting relationships are extracted according to experience identifiers, and item subgraphs containing combined, mutually exclusive, and sequential relationships are extracted according to declared items. These are written into the experience subgraph and item subgraph respectively for the voucher expansion module to read.The handling of anomalies or missing data is as follows: when an input field is missing from a certain perspective, supporting and counter-evidence items are disregarded for that perspective; only a missing field marker is written, and isolated nodes uniquely connected by the deleted edge are not included in the subgraph extraction results. In practical applications: for example, if the same student submits award certificates, guidance records, and credit application records for the same competition experience, the knowledge graph construction module first establishes corresponding edges for the records based on student identification, intersection of occurrence time periods, consistency of event name codes, and role adaptation relationships. Then, based on the inversion of event sequence, mismatch of achievement certificates, and events, and using conflict items, it performs splitting and merging of candidate node groups to obtain a stable node set and a stable edge set. Finally, it performs a four-perspective consistency check on the stable edge set and extracts the experience subgraph and event subgraph. Through this implementation process, the unified record set is transformed into a second-classroom knowledge graph with a stable structure, clear relationships, and traceable conflicts, providing direct input for the subsequent generation of experience evidence surfaces.

[0019] The credential expansion module reads the experience subgraph and, based on the time period nodes, participating role nodes, outcome content nodes, and proof content nodes, expands the same extracurricular experience into an experience credential surface composed of shared, exclusive, and linked parts. In this embodiment, the purpose of the credential expansion module is to organize the node relationships in the experience subgraph into an experience credential surface that can directly participate in the generation of subsequent slices. This allows content that can be called in parallel, can only be called individually, and will cause changes in subsequent usable relationships within the same extracurricular experience to be first layered and merged, and then its position rearranged and relationships checked. This avoids problems such as duplicate inclusion of the same proof content, confusion in the attribution of the same outcome content, or distortion of linkage relationships during subsequent slice generation. The implementation process first generates an initial credential surface composed of shared, exclusive, and linked parts based on the experience subgraph. Then, it performs field write-back and position rearrangement on the initial credential surface. Finally, it performs relational integrity verification on the rearranged result and outputs the experience credential surface. The implementation process includes the following steps: First, it identifies the shareable, exclusive, and linked credential structures within the same extracurricular experience from the experience subgraph. The input quantities are the occurrence time node, participating role node, outcome content node, proof content node, and limiting edge, belonging edge, and pointing edge in the experience subgraph. The processing actions are: reading each proof content node, retrieving the set of outcome content nodes connected to the proof content node by the pointing edge, and then retrieving the occurrence time node connected to the proof content node by the limiting edge. Retrieve the participating role nodes connected to each result content node via the home edge; when the same proof content node is simultaneously connected to two or more result content nodes via pointing edges, and the proof relationships corresponding to the two or more result content nodes all fall into the same occurrence time node via limiting edges, write the proof content node, the corresponding result content node, the corresponding participating role node, and the corresponding occurrence time node as the shared part; when the same proof content node is only connected to one result content node via a pointing edge, and the result content node is only connected to one participating role node via the home edge, write the proof content node, the result content node, the participating role node, and the corresponding occurrence time node as the exclusive part; when the same proof content node is connected to two or more result content nodes via pointing edges, and the result... If a content node corresponds to two or more participating role nodes, and deleting the belonging edge corresponding to any of the participating role nodes would cause the remaining result content nodes to lose their original belonging relationship or be transferred to other participating role nodes, then the proof content node, the corresponding result content node, the corresponding participating role node, and the corresponding occurrence time node are written as the linked part; the output is the initial credential surface and written to the initial credential table for the rearrangement step to read; the exception or missing handling is as follows: when the proof content node is not connected to any result content node, no part is generated and a missing marker is written; when the proof content node is connected to multiple occurrence time nodes and the occurrence time nodes are inconsistent, the proof content node is retained but a time conflict marker is written, and it is not directly classified into the shared part, exclusive part, or linked part;Secondly, the positions and attributions of each part in the initial voucher are fixed and organized to ensure that the voucher structure of the same extracurricular experience has a unique part attribution and a stable node arrangement before entering the subsequent slice generation; the input is the initial voucher; the processing actions are as follows: for each shared part, each exclusive part, and each linked part, the corresponding experience identifier, occurrence time node, participating role node, result content node, and proof content node are recorded respectively, and the part list is reconstructed using the experience identifier as the merge primary key; then, the reordering is performed in a fixed order, which is to first handle the conflict between shared parts and exclusive parts, then handle the attribution conflict of result content nodes, and finally handle the attribution chain conflict of participating role nodes; among them, the same proof content node cannot be written into both shared parts and exclusive parts at the same time. The processing rules for some parts are as follows: If the same proof content node appears in both the shared and exclusive parts of the initial credential table, the number of result content nodes connected to that proof content node via pointing edges is counted. If the number is one, the exclusive part is retained and the shared part is deleted; if the number is two or more, the shared part is retained and the exclusive part is deleted. The processing rules for the same result content node being assigned to only one part under the same occurrence time node are as follows: If the same result content node appears in both parts under the same occurrence time node, the part with more connecting edges to the corresponding proof content node is retained first. If the number of connecting edges is the same, the part with the earlier entry time in the initial credential table is retained. The same participating role node in the same proof content... The processing rule for retaining only one ownership chain under a node is as follows: For multiple participating role nodes corresponding to the same proof content node, generate an ownership chain sequence according to the writing order of the ownership edges, retain only the first ownership chain and delete the rest; the output is the experience credential surface and is written to the experience credential table for the verification step to read; the handling of anomalies or missing items is as follows: when the number of connecting edges is the same and the writing time is the same and cannot be directly resolved, retain the original correspondence and write a parallel mark, and wait for the verification step to continue processing; next, verify the integrity of the relationships of each part in the experience credential surface, and only retain the credential structure that can be completely supported by the experience subgraph; the inputs are the limiting edges, ownership edges, and pointing edges in the experience credential surface and experience subgraph; the processing action is: for each common part in the experience credential surface, each Each exclusive section and each linked section undergoes item-by-item verification of the limiting edge, attribution edge, and pointing edge. The limiting edge is used to verify the correspondence between the proof content node and the occurrence time node. The verification rule is that the proof content node is uniquely connected to the occurrence time node corresponding to that section through the limiting edge. The attribution edge is used to verify the correspondence between the result content node and the participating role node. The verification rule is that the result content node is uniquely connected to the participating role node corresponding to that section through the attribution edge. The pointing edge is used to verify the correspondence between the proof content node and the result content node. The verification rule is that the proof content node is directly connected to the result content node corresponding to that section through the pointing edge. When the limiting edge, attribution edge, and pointing edge of a certain section are all completely corresponding, that section is retained in the experience credential surface.When any corresponding relationship is missing in a certain part, that part is deleted from the experience credential surface. After verification, the remaining parts are re-summarized according to the experience identifier to generate an experience credential surface for the slice generation module to read and write it to the credential output table. The handling of anomalies or omissions is as follows: when only one corresponding relationship is missing in a certain part, a missing type code is written for subsequent traceability; when all parts under the same experience identifier are deleted, an empty credential marker is output and the transmission of that experience identifier to the slice generation module is stopped. In practical applications: for example, if the same student submits an award certificate and a team division of labor description in an innovation and entrepreneurship competition, and the award certificate points to both the competition award and the academic credit recognition achievement, and both occurred on the same competition date, then the credential expansion module will generate a shared credential surface corresponding to the award certificate. In this process, if another individual defense certificate only points to the student's individual presentation results, and these results correspond only to that student's participating role node, then it is generated as an exclusive part. If the team division of labor description connects both the project result node and the defense result node, and after deleting the attribution edge corresponding to the team leader role, the project result node will be rewritten as being taken over by the instructor role, then the team division of labor description is generated as a linked part. Subsequently, the module performs a rearrangement according to the certificate content node, result content node, and participating role node, and checks the limiting edge, attribution edge, and pointing edge item by item, retaining only the part with complete relationships to write into the experience credential surface. Through this implementation process, the experience subgraph is organized into an experience credential surface with a clear structure, unique attribution, and complete relationships, providing direct input for the subsequent generation of shared pieces, exclusive pieces, and linked pieces.

[0020] The slice generation module is used to read the experience voucher surface and the item sub-graph, slice the experience voucher surface into shared slices, exclusive slices, and linked slices, and output an occupancy slice diagram. In this embodiment, the purpose of the slice generation module is to further convert the shared, exclusive, and linked parts that have already been layered in the experience voucher surface into an occupancy slice diagram that can be directly read by the subsequent occupancy calculation module. This allows the item call relationships that can be used, mutually exclusive, and passed sequentially in the same second-classroom experience to be segmented first, and then returned to their positions and verified, thereby avoiding the problems of duplicate occupancy of the same proof content, mixed use of the same result content across slices, or incorrect linkage order during subsequent occupancy calculation. This implementation process first generates an initial occupancy slice group based on the experience voucher surface and the item sub-graph, and then... The initial occupied slice group is processed by writing fields and rearranging positions. Finally, the edge correspondence between each slice and the item subgraph is checked and the occupied slice diagram is output. The implementation process includes the following steps: First, based on the experience credential surface and the item subgraph, the shared slice, exclusive slice, and linked slice that can be called by the item are identified. The input quantities are the shared parts, exclusive parts, and linked parts in the experience credential surface, as well as the shared edges, mutually exclusive edges, and sequential edges in the item subgraph. The processing actions are: for each part, the corresponding proof content node, result content node, participating role node, and occurrence time node are read item by item, and then the declaration item nodes that are connected to this part are read. Among them, the connection relationship between the slice and the declaration item node is based on the item name code. The corresponding rules are established so that the item name code correspondence rule is that the item name contained in the proof content node or result content node is consistent with the item identifier of the application item node after being normalized by the item name code table; when the same part has a shared edge connection relationship with two or more application item nodes, and the proof content node, result content node, participating role node, and occurrence time node corresponding to this part are consistent under each application item node, this part is cut into a shared slice; when the same part has a connection relationship with only one application item node, and this application item node is separated from the other application item nodes by mutually exclusive edges, this part is cut into an exclusive slice; when the same part has a sequential edge connection relationship with two or more application item nodes, and When a preceding application node is invoked, causing the invocation status of the subsequent application node to change from invokeable to restricted or uninvoicable, that part is sliced ​​into a linked slice. The output is the initial occupied slice group, which is written to the initial occupied slice table for the rearrangement step to read. The exception or missing item handling is as follows: when a part does not match any application node, it is not sliced ​​and a missing item flag is written; when the same part simultaneously meets two types of slicing conditions, it is judged in a fixed order of shared slice, exclusive slice, and linked slice; if the preceding condition is met, the subsequent condition is not judged. Secondly, the fields of each slice in the initial occupied slice group are fixed and their positions are arranged to ensure that occupied slices under the same experience identifier have unique slice ownership and a stable arrangement. The input is the initial occupied slice group.The processing steps are as follows: For each shared area, each exclusive area, and each linked area, write the following: experience identifier, area identifier, corresponding application item node, corresponding proof content node, corresponding result content node, corresponding participating role node, and corresponding occurrence time node. The area identifier is generated by concatenating the experience identifier with the sequential number within the area, and the sequential number within the area is generated by incrementing the initial time of writing the area. Then, a reordering is performed in a fixed order: first, handle conflicts in the attribution of proof content nodes; then, handle conflicts in the attribution of result content nodes; finally, handle the connection order of linked areas. The rule for assigning the same proof content node to only one shared area or one exclusive area under the same occurrence time node is: if the same proof content node appears simultaneously in both a shared area and an exclusive area under the same occurrence time node... In a piece of paper, the number of application item nodes corresponding to the proof content node is counted. If there is only one node, the piece of paper is retained; if there are two or more nodes, the piece of paper is retained. The rule for classifying the same result content node into only one exclusive piece or one linked piece under the same participating role node is as follows: If the same result content node appears in both exclusive and linked pieces under the same participating role node, the number of preceding and following edges corresponding to the result content node is read first. If preceding and following edges exist, the piece of paper is retained; otherwise, the piece of paper is retained. The rule for connecting linked pieces corresponding to the same application item node under the same experience identifier according to the order of preceding and following edges is as follows: Map the application item node pairs corresponding to each linked piece to the preceding and following edge table, generating them in the order of preceding and following items. Arrange the linked pieces into a sequence, connecting the end of the previous linked piece to the beginning of the next linked piece; the output is the occupied piece table, which is written to the occupied piece storage table for the verification step to read; the abnormal or missing handling is as follows: if a duplicate piece identifier is generated, the sequence number is appended to the original piece's sequence number to regenerate the piece identifier; if there are no connectable sequential edges between linked pieces, the original linked piece is retained and a connection missing mark is written; next, verify the edge item correspondence between each piece in the occupied piece table and the item subgraph, retaining only the occupied pieces with complete relationships and valid connections and their inter-piece connections; the inputs are the shared edges, mutually exclusive edges, and sequential edges in the occupied piece table and the item subgraph; the processing action is: verify the edge item correspondence between each shared piece, each exclusive piece, and each linked piece and the item subgraph. The system is structured as follows: Shared area verification rules require that all nodes corresponding to a shared area have shared edges in pairs; Exclusive area verification rules require that nodes corresponding to an exclusive area have mutually exclusive edges with all other connected nodes; Linked area verification rules require that nodes corresponding to a linked area have sequential edges, and the order of these edges matches the order of the linked area in the occupied area table. When all corresponding edge items of an area are valid, the area is retained and written into the occupied area diagram; when an area is missing any corresponding edge item, the area is deleted from the occupied area diagram. After verification, the retained areas and their inter-area connections are summarized according to the experience identifier and written into the occupied area diagram. The occupied area diagram is then output for the pressure occupancy calculation module to read and written into the occupied area chart.The handling of anomalies or missing items is as follows: When the number of items corresponding to a shared item is less than two, the number of items corresponding to a single item is not one, or the number of items corresponding to a linked item is less than two, the item is directly deleted and an item type mismatch flag is written. When all occupied items under a certain experience identifier are deleted, an empty occupied item map flag is output and the transmission of the experience identifier to the occupancy calculation module is stopped. In practical applications: For example, a competition experience of the same student forms a shared part, a single part, and a linked part in the experience certificate. The shared part corresponds to both innovation and entrepreneurship credits and comprehensive quality items, the single part only corresponds to the evaluation item, and the linked part corresponds to two items that first participate in the college-level assessment and then enter the school-level assessment. The application item node; the slice generation module first slices the shared part into shared slices, the exclusive part into exclusive slices, and the linked part into linked slices according to the correspondence of item name codes. Then, it rearranges them according to the proof content node, result content node, and application item node. Finally, it checks whether there are shared edges between the corresponding item nodes of the shared slice, whether there are mutually exclusive edges between the corresponding item nodes of the exclusive slice and other item nodes, and whether there are sequential edges between the corresponding item nodes of the linked slice and their arrangement order is consistent. Through this implementation process, the vacancy map is transformed into an occupancy slice map with clear edge items, unique slice ownership, and stable connection order, providing direct input for the subsequent occupancy calculation module to perform occupancy locking, shared deduction, and linked transmission.

[0021] The occupancy calculation module reads the occupancy map and application records, pushes each application item node into the corresponding slice according to the application order, performs occupancy locking on exclusive slices, deducts the remaining share on shared slices, rewrites the subsequent available state of linked slices, and cuts the occupancy paths connected by mutual exclusion edges. It outputs the item occupancy result map and the remaining available map. In this embodiment, the purpose of the occupancy calculation module is to calculate a set of stable paths satisfying mutual exclusion constraints, sequence constraints, and occupancy constraints by combining the shared slices, exclusive slices, and linked slices in the occupancy map with the application items and application order in the application records. It further generates the item occupancy result map and the remaining available map, ensuring that the calling relationship of the same extracurricular experience among multiple application items first completes path filtering, and then... The process involves state transfer and final result verification to avoid issues such as item mismatch, duplicate deduction of shared pieces, or distortion of linked piece transfer during subsequent result generation. The implementation process first establishes a candidate piece sequence and filters out invalid paths, then performs state expansion and stable retention on the candidate piece sequence, and finally verifies the stable paths and writes them into the item occupancy result graph and the remaining available graph. The implementation process includes the following steps: First, a candidate piece sequence satisfying basic constraints is generated based on the occupied piece graph and declaration records; the input quantities are shared pieces, exclusive pieces, linked pieces, mutually exclusive edges, and sequential edges in the occupied piece graph, as well as the declaration items and declaration order in the declaration records; the processing action is: aggregate all declaration items and all occupied pieces corresponding to the same experience identifier, and build... The system includes a slice matrix, slice continuation matrix, mutual exclusion matrix, and sequence matrix. In the slice matrix, rows correspond to application items, columns correspond to occupied slices, and a value of 1 indicates that the application item can use the occupied slice, while a value of 0 indicates that the application item cannot use the occupied slice. In the slice continuation matrix, both rows and columns correspond to occupied slices; a value of 1 indicates that a subsequent occupied slice is allowed after a previous occupied slice, while a value of 0 indicates that continuation is not allowed. In the mutual exclusion matrix, both rows and columns correspond to application items; a value of 1 indicates that two application items are mutually exclusive, while a value of 0 indicates that two application items are not mutually exclusive. In the sequence matrix, both rows and columns correspond to application items; a value of 1 indicates that the application item in the row precedes the application item in the column, while a value of 0 indicates that the sequence relationship does not exist. Read the current application items one by one in the application order. Multiply the row vector of the item slice matrix corresponding to the current application item with the slice continuation matrix continuously. The stopping condition for continuous multiplication is that no new occupied slice position appears in the product result of the new round of multiplication, so as to obtain the candidate slice sequence corresponding to the current application item. Perform deletion processing on each candidate slice sequence. The deletion rules include: delete the candidate slice sequence when the element value of the item pair corresponding to the candidate slice sequence appears after mapping to the mutual exclusion matrix; delete the candidate slice sequence when the order of the items corresponding to the candidate slice sequence is inconsistent after mapping to the sequence matrix; delete the candidate slice sequence when the same exclusive slice appears twice or more in the same candidate slice sequence. The output is a candidate path table and is written to the candidate path storage table for reading in the state expansion step.The handling of anomalies or missing elements is as follows: When all elements in the corresponding row of a certain application item in the application item matrix are zero, a "no callable item" flag is directly written for that application item and no candidate item sequence is generated. When all candidate item sequences under the same experience identifier are deleted, an empty path flag is written and subsequent occupancy calculations for that experience identifier are stopped. Secondly, occupancy status expansion and path convergence are performed on the candidate item sequences to obtain a stable path table that can be used for result verification. The inputs are the candidate path table, application item matrix, item continuation matrix, and sequence matrix. The processing actions are: for each candidate item sequence, a path matrix, a shared surplus vector, an exclusive occupancy vector, and a linkage transmission vector are established, where each row of the path matrix corresponds to the application order. The sequence position and column correspond to the occupied slice. An element value of 1 indicates that the occupied slice is called at this sequence position, and an element value of 0 indicates that the occupied slice is not called at this sequence position. Each bit of the shared surplus vector corresponds to the remaining callable count of each shared slice. Its initial value is taken from the initial callable count field of the shared slice in the occupied slice diagram. The initial callable count is given by the event relationship constraint rules, that is, the number of event nodes that can be used together corresponding to the shared slice. Each bit of the exclusive occupancy vector corresponds to the occupancy status of each exclusive slice. The initial value is zero. An element value of 0 indicates that it is not occupied, and an element value of 1 indicates that it is occupied. Each bit of the linkage transmission vector corresponds to the status code of each linkage slice. The initial value is taken from the initial status code field of the occupied slice diagram. A prime value of zero indicates no transmission, a value of one indicates transmission, and a value of two indicates that subsequent calls are restricted after transmission. Subsequently, the path matrix is ​​processed sequentially using shared deduction, exclusive locking, and linkage transmission. Shared deduction is calculated by subtracting the shared remainder vector bit by bit based on the frequency of each shared segment in the path matrix. Exclusive locking is calculated by rewriting the corresponding bit of the exclusive occupancy vector to one when the element value corresponding to an exclusive segment in the path matrix is ​​first written as one. Linkage transmission is calculated by transmitting the linkage segment status code corresponding to the preceding and following events in the sequence matrix to the linkage segment position corresponding to the following event. After transmission, if the linkage segment is restricted from subsequent calls, the following applies. If the following conditions are met, the corresponding slice position of the subsequent item will be written as a restricted state. After the above processing is completed, the path matrix and the item slice matrix are multiplied to obtain the item expansion matrix. Singular value decomposition is performed on the item expansion matrix, and the column space corresponding to non-zero singular values ​​is extracted. The elements of the item expansion matrix are all discrete integers, and the non-zero singular values ​​are singular values ​​that are strictly greater than zero. Orthogonal projection is performed on the column space to generate a standardized path code. The generation rule of the standardized path code is to read the column positions retained after projection in the order of declaration and to concatenate the corresponding slice identifiers in order. For candidate paths with the same standardized path code and consistent shared surplus vector, exclusive occupancy vector, and linkage transmission vector, only one path is retained, and the rest are deleted.After each declaration item is written sequentially, the above operation is repeated until the standardized path code set, shared surplus vector set, exclusive occupation vector set, and linkage transmission vector set obtained in two adjacent rounds are completely consistent. The output is a stable path table, which is written to the stable path storage table for the verification step to read. The handling of anomalies or missing items is as follows: when any bit in the shared surplus vector is less than zero after deduction, the candidate path is deleted; when the same subsequent item is simultaneously in a callable state and a restricted state after linkage transmission, the restricted state is retained and a transmission conflict flag is written; when multiple candidate paths are retained under the above rules, the next round of writing continues, without premature deletion in this step; furthermore, the correspondence between stable path execution items, shared deduction results, and... The system performs a complete verification of the linkage transmission results and generates a task occupancy result graph and a remaining availability graph. The inputs are a stable path table, a task piece matrix, a shared surplus vector, and a linkage transmission vector. The processing steps are as follows: For each stable path, construct an occupancy verification matrix, a surplus verification matrix, and a transmission verification matrix. In the occupancy verification matrix, rows correspond to declared tasks, columns correspond to occupied pieces, and element values ​​are taken from the actual call status of the declared task for that occupied piece in the stable path. In the surplus verification matrix, rows correspond to declared tasks, columns correspond to shared pieces, and element values ​​are taken from the actual deduction count of the declared task for that shared piece in the stable path. In the transmission verification matrix, rows correspond to preceding declared tasks, columns correspond to subsequent declared tasks, and element values ​​are taken from the linkage piece status in the stable path. The code is passed through the following steps: First, the occupancy check matrix is ​​multiplied by the item slice matrix to verify the item correspondence. A successful verification occurs when the element value at each non-zero position in the occupancy check matrix is ​​one in the corresponding position in the item slice matrix. Then, the surplus check matrix is ​​compared bit-by-bit with the shared surplus vector to verify the shared deduction result. A successful verification occurs when the sum of the elements in the corresponding column of the surplus check matrix minus the initial callable count of a shared slice is the same as the corresponding bit in the shared surplus vector. Finally, the transmission check matrix is ​​compared bit-by-bit with the linkage transmission vector to verify the linkage transmission result. A successful verification occurs when the element values ​​in the transmission check matrix of item pairs with a sequential relationship are consistent with the status codes corresponding to the linkage transmission vector. When all three verifications are successful, the stable path is written into the item occupancy check matrix. The result graph is generated, and the shared surplus vector and linkage transmission vector corresponding to the stable path are written into the remaining available graph. The item occupancy result graph is written into the application item node, occupied piece node and inter-piece call edge according to the experience identifier. The remaining available graph is written into the shared piece surplus field and linkage piece status field according to the experience identifier. The output is the item occupancy result graph and the remaining available graph and written into the result generation module read table. The abnormal or missing handling is as follows: when any matrix in the occupancy verification matrix, surplus verification matrix or transmission verification matrix has an empty row, the stable path is deleted and a verification missing mark is written. When there are two or more stable paths that are both verified to be true under the same experience identifier, one path is retained and the rest are deleted according to the rule of the standard path code dictionary order.In practical applications: For example, if the same student applies for the same competition experience simultaneously as an innovation and entrepreneurship credit item, a comprehensive quality item, and an award item, where the innovation and entrepreneurship credit item and the comprehensive quality item share the same common segment, while the award item occupies a unique segment, and the school-level award item is also subject to the linkage segment transmission constraint of the college-level award item; the pressure calculation module first establishes the item segment matrix, segment continuation matrix, mutual exclusion matrix, and sequence matrix, then generates a candidate segment sequence according to the application order and deletes candidate segment sequences where mutual exclusion items coexist, the sequence positions are reversed, or unique segments appear repeatedly; subsequently, the shared deduction is performed sequentially on the retained candidate segment sequences. Exclusive locking and linkage transmission are implemented, and standardized path codes are generated through singular value decomposition and orthogonal projection of the event expansion matrix, merging candidate paths with the same status. Finally, the event correspondence, shared deduction results, and linkage transmission results of stable paths are checked item by item, and only the verified stable paths are written into the event occupancy result diagram, while the shared piece balance and linkage piece status are written into the remaining available diagram. Through this implementation process, the occupancy relationship of the same experience among multiple declared events is solved into a set of stable results that satisfy mutual exclusion, order, and occupancy constraints, providing a direct basis for the subsequent result generation module to output a unique and valid identification result.

[0022] The result generation module is used to read the event occupation result map and the remaining available map, perform conflict cancellation, sequential review, and map rewriting, and output a unique corresponding valid identification result and its occupation trajectory. In this embodiment, the purpose of the result generation module is to further consolidate the established occupation results in the event occupation result map and the remaining available map into a unique corresponding valid identification result and its occupation trajectory under the same experience identifier. This ensures that the multiple feasible paths obtained from the previous occupation calculation first complete conflict cancellation, then sequential review, and finally map rewriting, thereby avoiding the simultaneous retention of multiple sets of results, the retention of reversed order results, or the retention of distorted write-back results for the same second classroom experience among multiple application items. This implementation process first follows the experience identifier... The process involves: identifying and deleting conflicting paths, then performing sequential verification on the remaining paths to generate valid identification results and occupation trajectories. Finally, the valid identification results and occupation trajectories are written back to the result nodes and connecting edges. The implementation process includes the following steps: First, conflict cancellation is performed on multiple occupation paths under the same experience identifier, retaining only candidate results that meet basic consistency. The input quantities are the application item node, occupied segment node, inter-segment call edge, and occupation path in the occupation result diagram, as well as the shared reserve and linkage transmission results in the remaining available diagram. The processing action is: Based on the same experience identifier, the occupation paths, occupied segment sequences, shared reserve, and linkage transmission results corresponding to each application item under that experience identifier are aggregated to form a path aggregation table. Subsequently, conflict resolution is performed on each occupied path in the path aggregation table. The rules for determining if there are mutually exclusive edges connecting occupied slices are: if any two declared items in the same occupied path have an element value of one after mapping to the item mutual exclusion relationship table, or if there is an inter-slice relationship connected by mutually exclusive edges between any two occupied slices; the rules for determining if duplicate deductions occur after writing back shared reserves are: if the initial callable count of the same shared slice minus the total deduction count of that shared slice in the occupied path, and then minus the remaining count of that shared slice in the remaining available graph, the result is not zero; the rules for determining if the linkage transmission result does not correspond to the order of edges are: if the status code corresponding to the preceding item in the linkage transmission result is not earlier than the status code corresponding to the following item. The status code is written, or the status code of the subsequent event has been rewritten while the status code of the preceding event remains in the untransmitted state; when a certain occupied path meets any of the above deletion conditions, the occupied path is deleted; if the deletion conditions are not met, the occupied path is retained; the output is a pending result group and written to the pending result table for sequential review steps to read; the abnormal or missing handling is as follows: when no occupied path is aggregated under the same experience identifier, an empty result mark is written and the subsequent result generation of that experience identifier is stopped; when the shared margin or linkage transmission result is missing, the corresponding occupied path is written to the pending review result table instead of being directly deleted; secondly, sequential review is performed on the retained paths in the pending result group, and valid identification results and occupied trajectories are generated; the input is the pending result group;The processing steps are as follows: For each occupied path retained under the same experience identifier, read the order of the application items and the order of the occupied slices, and match the application order position with the occupied slice writing position bit by bit. The rule for determining the bit-by-bit correspondence is that the writing order position of the i-th application item in the occupied path corresponds one-to-one with the calling order position of the i-th occupied slice in the occupied slice sequence. Furthermore, if two application items have a sequential relationship in the item priority table, the occupied slice corresponding to the preceding item must be positioned before the occupied slice corresponding to the following item in the occupied slice sequence. When an occupied path satisfies the above correspondence rules, it is retained as a valid path. If any application item order in an occupied path is reversed from the occupied slice sequence order... When the time is set, the occupied path is deleted; for the retained valid paths, valid identification results and occupied trajectories are generated according to the occupied piece sequence, shared reserve, and linkage transmission results. The valid identification results include at least the experience identifier, declaration item, call establishment status, and corresponding occupied piece identifier, and the occupied trajectory includes at least the occupied piece sequence and inter-piece connection order; when there are two or more valid paths under the same experience identifier, one is retained according to the rule of the standard path code dictionary order, and the rest are deleted; the output is the valid identification results and occupied trajectory, which are written to the valid result table and trajectory table respectively for the map back-write step to read; the abnormal or missing handling is as follows: when all occupied paths under the same experience identifier are deleted, a conflict result mark is written; when there is a retained path but the standard path code dictionary order is not specified, the conflict result mark is written. When the path code is missing, a replacement code is generated by concatenating the segment identifiers of the occupied segment sequence before unique retention is performed. Next, the valid identification results and occupation trajectories are written back to the map result layer, forming a unique output result that can be directly called for querying, verification, and subsequent statistics. The input quantities are the valid identification results and occupation trajectories. The processing actions are as follows: Locate the result node set and the occupied segment connection edge set according to the experience identifier; write the valid identification results to the result nodes corresponding to the experience identifier and the declared item, where the written fields of the result nodes include the experience identifier, the declared item, the identification status, and the corresponding occupied segment identifier; write the occupation trajectory to the connection edges between the corresponding occupied segments, where the written fields of the connection edges include the previous occupied segment identifier, the next occupied segment identifier, and the connection edge identifier. The sequence position is then determined; the shared surplus is written back to the remaining number of times field of the corresponding shared piece node, and the linkage transmission result is written back to the status field of the corresponding linkage piece edge; after the write-back is completed, the old result nodes and old connection edges that already exist under the same experience identifier are overwritten, and the overwriting rule is to delete the old result nodes and old connection edges and then write the current valid identification result and the current occupation trajectory; the output is the unique corresponding valid identification result and its occupation trajectory and is written to the result output table for the platform query end, review end and statistics end to read directly; the abnormal or missing handling is as follows: when the result node fails to be located, the result node is regenerated according to the experience identifier and the node primary key is filled in; when the connection edge fails to be located, the connection edge is reconstructed according to the occupation piece sequence in the occupation trajectory and then the write-back is performed;In practical applications: For example, if the same student uses a competition experience simultaneously for innovation and entrepreneurship credits, comprehensive quality assessment, and awards, the occupancy calculation module outputs two feasible occupancy paths. One path has mutually exclusive edges connecting the comprehensive quality assessment and awards, while the other path, although lacking mutually exclusive edges, shows that the school-level awards are written before the college-level awards. The result generation module first performs conflict cancellation on the two paths, deleting the occupancy path with mutually exclusive edges. Then, it performs sequence verification on the remaining occupancy paths, confirming that the college-level awards occupy a piece of land before the school-level awards. Subsequently, it generates the valid assessment result and occupancy trajectory under this experience identifier, and writes the result nodes corresponding to innovation and entrepreneurship credits, comprehensive quality assessment, and awards into the graph result layer. Simultaneously, it writes back the remaining number of shared pieces and the status code of linked pieces to the corresponding nodes and edges. Through this implementation process, only one set of queryable, verifiable, and traceable final assessment results is retained under the same experience identifier, avoiding inconsistencies caused by multiple paths coexisting.

[0023] Working Principle: This solution first organizes the experience records, application records, and event records uploaded by students, teachers, and administrators into a unified record set. Based on this, a second-classroom knowledge graph is constructed, uniformly representing the time, role, achievements, proof, and event relationships corresponding to the same experience within a single relational structure. Furthermore, an experience is broken down into shareable, exclusive, and dynamically changing credential parts, and then further divided into occupancy pieces that can be invoked by specific application events. Subsequently, the system, according to the application order, allows each application event to sequentially invoke its corresponding occupancy piece, simultaneously processing the remaining uses of shared pieces, the locked status of exclusive pieces, and the impact of linked pieces on the availability of subsequent events. Finally, conflicting paths are deleted, and only one result path with the correct order and valid relationships is retained, generating a unique and valid recognition result and its occupancy trajectory. In other words, this solution does not simply reuse a second-classroom experience for multiple events; instead, it first determines which content within the experience can be shared, which can only be used independently, and which will affect subsequent events. Based on this, a unique and consistent recognition result is obtained. For example, a student participates in an innovation and entrepreneurship competition. This experience may be used for innovation and entrepreneurship credit recognition, comprehensive quality evaluation, and awards. After the student submits award materials, the teacher submits guidance records, and the administrator submits the rules for the use, mutual exclusion, and sequence of various items, the system first groups this information into the same experience. Then, it determines which content in the competition certificate can support both credits and comprehensive quality evaluation, which content can only be used for awards, and which content must be recognized at the college level before it can be used for the university level recognition. After that, the system calculates item by item according to the application order which items are valid, which items will conflict with each other, and which items will occupy the subsequent available parts. Finally, only one set of recognition results that is non-conflicting, correctly ordered, and traceable is retained. In this way, what the administrator sees is no longer several scattered records, but the complete process of how an experience is used among multiple recognition items, how much available content is left, and why the final result is formed.

[0024] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An AI-driven intelligent management cloud platform for extracurricular activities, characterized in that: include: The data acquisition terminal module receives experience records, application records, and event records uploaded by students, teachers, and administrators. It extracts student identifiers, experience identifiers, time periods, participating roles, achievement content, proof content, application items, application order, concurrent use relationships, mutually exclusive relationships, and sequential relationships, and outputs a unified record set. The knowledge graph construction module reads the unified record set and uses student identifiers, experience identifiers, time periods, participating roles, achievement content, proof content, and application items as nodes, and uses attribution relationships, pointing relationships, limiting relationships, concurrent use relationships, mutually exclusive relationships, and sequential relationships as edges to construct a second-classroom knowledge graph, outputting experience subgraphs and event subgraphs. The credential expansion module is used to read the experience subgraph and, based on the limiting edges, belonging edges, and pointing edges between the occurrence time node, participating role node, result content node, and proof content node, expand the same extracurricular experience into an experience credential surface composed of shared parts, exclusive parts, and linked parts. The slice generation module is used to read the experience voucher surface and the event sub-graph, slice the experience voucher surface into shared slices, exclusive slices and linked slices, and output the occupied slice graph.

2. The AI-driven intelligent management cloud platform for the second classroom according to claim 1, characterized in that: Also includes: The occupancy calculation module is used to read the occupancy map and declaration records, push each declaration item node into the corresponding slice according to the declaration order, perform occupancy locking on exclusive slices, perform remaining share deduction on shared slices, perform subsequent availability status rewriting on linked slices, and cut off the occupancy path connected by mutually exclusive edges, outputting the item occupancy result map and the remaining availability map; the result generation module is used to read the item occupancy result map and the remaining availability map, perform conflict cancellation, sequence verification and map back writing, and output the unique corresponding valid identification result and its occupancy trajectory.

3. The AI-driven intelligent management cloud platform for the second classroom according to claim 2, characterized in that: The data collection terminal module includes: reading experience records, application records, and event records uploaded by students, teachers, and administrators from different sources, extracting experience identifiers, student identifiers, time periods, participating roles, results content, supporting documentation, application items, application order, concurrent application relationships, mutually exclusive application relationships, and sequential application relationships according to the type of uploading terminal; merging records from different uploading terminals but with the same experience identifier into the same data collection group, and outputting terminal-separated record groups; performing a peer-to-peer comparison of the time periods, participating roles, results content, and supporting documentation in each terminal-separated record group, and identifying records with consistent time periods and results content... When the evidence content points to each other, an experience master piece is generated. When there is a correspondence between the declared items and the items, the items are mutually exclusive, or the items are sequential, an item constraint piece is generated. The experience master pieces and item constraint pieces are associated with each other according to the experience identifier and written into a unified record set. Missing items are filled and conflicts are marked for each experience master piece and each item constraint piece in the unified record set. When there are multiple participating roles or multiple evidence contents under the same experience identifier, the original correspondence is preserved and written into the position identifier in the piece respectively. When the same declared item corresponds to multiple sequential positions, the results are generated in order of upload time. The unified record set is output for the knowledge graph construction module to read.

4. The AI-driven intelligent management cloud platform for the second classroom according to claim 3, characterized in that: The knowledge graph construction module includes: reading the student identifier, experience identifier, occurrence time, participating role, result content, proof content, and application item of each record in the unified record set; establishing record-to-record edges according to the rules of consistent student identifiers, overlapping occurrence times, consistency between the item name appearing in the result content and the item name appearing in the proof content, participating roles not belonging to conflicting roles limited by mutual exclusion relationships, and the unique sequential position of the same application item corresponding to one record; and grouping records with corresponding record edges into the same candidate node group, outputting candidate node groups and candidate relationship groups; and processing each candidate node group step by step based on the candidate node groups and candidate relationship groups. The system identifies items with time-segment breaks, role conflicts, mismatched achievement proofs, reversed order of events, and conflicting use of events. Candidate node groups are reconstructed in a fixed order of splitting and merging. The splitting rule separates records containing any conflicting items from the original candidate node group, while the merging rule merges candidate node groups with identical student identifiers and no conflicting items. After each round of reconstruction, candidate relationship groups are regenerated based on experience identifier, time period, participating role, achievement content, proof content, and application item, until the candidate node group members and edge items obtained from two consecutive rounds of reconstruction are completely identical. A stable node set and a stable edge set are then output.

5. The AI-driven intelligent management cloud platform for the second classroom according to claim 4, characterized in that: The knowledge graph construction module further includes: based on a stable node set and a stable edge set, performing consistency checks on each edge from the perspectives of time period, role, result proof, and matter, respectively. Specifically, the time period perspective checks whether the occurrence time corresponds to the limiting relationship; the role perspective checks whether the participating role corresponds to the attribution relationship; the result proof perspective checks whether the result content corresponds to the pointing relationship; and the matter perspective checks whether the declared matter corresponds to the use relationship, mutual exclusion relationship, and sequential relationship. Each edge is processed according to the following rules: if the number of supporting items is greater than the number of counter-proof items, the edge is retained; if the number of supporting items is equal to the number of counter-proof items, the edge is retained and a conflict flag is written; and if the number of supporting items is less than the number of counter-proof items, the edge is deleted. Subsequently, experience subgraphs containing attribution, pointing, and limiting relationships are extracted according to experience identifiers, and matter subgraphs containing use relationships, mutual exclusion relationships, and sequential relationships are extracted according to declared matters.

6. The AI-driven intelligent management cloud platform for the second classroom according to claim 5, characterized in that: The credential expansion module includes: reading the occurrence time node, participating role node, result content node, and proof content node from the experience subgraph; generating a shared portion according to the rule that the same proof content node is simultaneously connected to two or more result content nodes via pointing edges and falls into the same occurrence time node via limiting edges; generating an exclusive portion according to the rule that the same proof content node is connected to only one result content node via pointing edges and that result content node corresponds to only one participating role node; generating a linked portion according to the rule that the result content node connected to the same proof content node via pointing edges corresponds to multiple participating role nodes and that a change in any participating role node will cause a rewrite of the affiliation relationship of the remaining result content nodes; and outputting the initial credential surface; based on the initial credential surface, recording the corresponding experience identifier, occurrence time node, participating role node, and result content node for each shared portion, each exclusive portion, and each linked portion. The result content nodes and proof content nodes are rearranged according to the following rules: the same proof content node cannot be written into both the shared and exclusive parts at the same time; the same result content node is only assigned to one part under the same occurrence time node; and the same participating role node retains only one attribution chain under the same proof content node. The experience credential surface is then output. Based on the experience credential surface, the limiting edge, attribution edge, and pointing edge of each part are checked item by item. The limiting edge is used to check the correspondence between the proof content node and the occurrence time node, the attribution edge is used to check the correspondence between the result content node and the participating role node, and the pointing edge is used to check the correspondence between the proof content node and the result content node. Parts with complete correspondence of limiting edge, attribution edge, and pointing edge are retained in the experience credential surface, and parts with missing correspondence are deleted from the experience credential surface. The experience credential surface is then output for the slice generation module to read.

7. The AI-driven intelligent management cloud platform for the second classroom according to claim 6, characterized in that: The slice generation module includes: based on the shared parts, exclusive parts, and linked parts in the experience credential surface, as well as the shared edges, mutually exclusive edges, and sequential edges in the item subgraph, it reads the corresponding proof content node, result content node, participating role node, and occurrence time node for each part item by item; it segments the parts that have shared edge connections with two or more application item nodes and whose corresponding proof content node, result content node, participating role node, and occurrence time node are consistent as shared slices; it segments the parts that have only one application item node connection and that application item node is separated from the other application item nodes by a mutually exclusive edge as exclusive slices; and it segments the parts that have sequential edge connections with two or more application item nodes and whose available relationships between the preceding application item node and the subsequent application item node are rewritten as linked slices, outputting an initial occupied slice group; and based on the initial occupied slice group, it writes the experience identifier, slice identifier, corresponding application item node, and corresponding... The system should prove the content nodes, corresponding result content nodes, corresponding participating role nodes, and corresponding occurrence time nodes. It should then rearrange the nodes according to the following rules: the same proof content node should be assigned to only one shared segment or one exclusive segment under the same occurrence time node; the same result content node should be assigned to only one exclusive segment or one linked segment under the same participating role node; and the linked segments corresponding to the same application item node under the same experience identifier should be connected end-to-end according to the order of their edges. The resulting occupied segment table should be output. Based on the occupied segment table, the system should verify the edge correspondence between each shared segment, each exclusive segment, and each linked segment and the item subgraph. Specifically, for shared segments, it should verify whether there are shared edges between the corresponding application item nodes; for exclusive segments, it should verify whether there are mutually exclusive edges between the corresponding application item node and other application item nodes; and for linked segments, it should verify whether there are sequential edges between the corresponding application item nodes and whether the connection order is consistent with the arrangement order within the segment. The system should then write the complete segments corresponding to the edge items and their interconnections into the occupied segment graph, and output the occupied segment graph for the pressure occupancy calculation module to read.

8. The AI-driven intelligent management cloud platform for the second classroom according to claim 7, characterized in that: The occupancy calculation module includes: based on the shared pieces, exclusive pieces, linked pieces, mutual exclusion edges, and sequential edges in the occupancy piece map, as well as the declaration items and declaration order in the declaration records, establishing a piece piece matrix, piece continuation matrix, mutual exclusion matrix, and sequential matrix according to the same experience identifier; continuously multiplying the row vector of the piece piece matrix corresponding to the current declaration item with the piece continuation matrix to generate a candidate piece sequence; deleting candidate piece sequences with non-zero mutual exclusion matrix mapping values, mismatched sequential matrix positions, and duplicate exclusive pieces; and outputting a candidate path table; based on the candidate path table, establishing a path matrix, shared surplus vector, exclusive occupancy vector, and linked transfer vector for each candidate piece sequence. The path matrix is ​​then subjected to shared deduction, exclusive locking, and linkage transmission in sequence. The path matrix is ​​multiplied by the item fragment matrix to obtain the item expansion matrix. Singular value decomposition is performed on the item expansion matrix to extract the column space corresponding to non-zero singular values. Orthogonal projection is performed on the column space to generate standardized path codes. One candidate path with the same standardized path code and consistent shared surplus vector, exclusive occupancy vector, and linkage transmission vector is retained. The above operation is repeated after each application item is written in sequence until the standardized path code set, shared surplus vector set, exclusive occupancy vector set, and linkage transmission vector set obtained in two adjacent rounds are consistent. A stable path table is then output.

9. The AI-driven intelligent management cloud platform for the second classroom according to claim 8, characterized in that: The occupancy calculation module further includes: constructing an occupancy check matrix, a surplus check matrix, and a transfer check matrix for each stable path based on the stable path table; multiplying the occupancy check matrix with the item piece matrix to check the item correspondence; checking the surplus check matrix with the shared surplus vector position by position to check the shared deduction result; checking the transfer check matrix with the linkage transfer vector position by position to check the linkage transfer result; and writing the stable paths that have been checked and found to be correct into the item occupancy result graph, and writing their corresponding shared surplus vector and linkage transfer vector into the remaining available graph.

10. The AI-driven intelligent management cloud platform for the second classroom according to claim 9, characterized in that: The result generation module includes: based on the occupancy result map and the remaining available map, it aggregates the occupancy paths, occupancy sequence, shared reserve, and linkage transmission results corresponding to each declared item according to the same experience identifier; it performs conflict cancellation on each occupancy path under the same experience identifier, deleting occupancy paths in the occupancy sequence that have mutually exclusive edge connections, duplicate deductions after writing back the shared reserve, and linkage transmission results that do not correspond to the order of edges, and outputs a pending result group; based on the pending result group, it performs a sequence review on each occupancy path retained under the same experience identifier, and sorts out the declared items in each occupancy path. The sequence of occupants is aligned position by position with the sequence of occupants, and the occupancy paths corresponding to each position are retained as valid paths. Paths with reversed order are deleted. Valid identification results and occupancy trajectories are generated based on the occupant sequence, shared reserve, and linkage transmission results in the valid paths. Based on the valid identification results and occupancy trajectories, the valid identification results are written to the result nodes of the corresponding experience identifier and declaration items, the occupancy trajectories are written to the connection edges between the corresponding occupants, and the shared reserve and linkage transmission results are written back to the corresponding nodes and edges. The unique valid identification results and their occupancy trajectories are then output.

Images