Teaching resource dynamic scheduling system and method based on multi-agent cooperation

Abstract

This invention relates to the field of teaching resource scheduling technology, and discloses a dynamic scheduling system and method for teaching resources based on multi-agent collaboration. The system includes: a basic modeling module for constructing teaching chain units, teaching chain graphs, resource information tables, and matching rules; a constraint generation module for generating a set of constraint information; a disturbance identification module for identifying affected units, affected ranges, and non-scheduling domain states; a candidate generation module for generating a set of effective composite resource candidates; a collaborative evaluation module for generating evaluation results for candidate combinations; a scheduling execution module for determining the optimal composite resource combination and completing local rescheduling; and a state update module for updating execution state information. This invention realizes the dynamic adjustment of teaching resources and the updating of execution state during the teaching execution process.

Description

Technical Field

[0001] This invention belongs to the field of teaching resource scheduling technology, specifically relating to a dynamic scheduling system and method for teaching resources based on multi-agent collaboration. Background Technology

[0002] With the development of information technology-based teaching, schools, training institutions, and other teaching organizations typically need to make unified arrangements for resources such as teachers, teaching content, classrooms, equipment, and teaching time slots to ensure that teaching activities are carried out as planned. Existing teaching management systems mostly focus on timetable arrangement, course scheduling, or static resource allocation, usually completing the corresponding configuration of teachers, venues, and times before the start of teaching.

[0003] However, in actual teaching, situations often arise such as teachers being temporarily unavailable, classroom occupancy conflicts, equipment malfunctions, adjustments to teaching content, changes in student groupings, and failure to complete prior teaching tasks as expected, causing the original resource allocation relationships to become invalid. Existing solutions mostly involve manual adjustments or overall rearrangement, often focusing only on replacing individual resources and rarely considering the connection between knowledge points, the suitability of teaching difficulty, differences among students, and the coordination between multiple types of resources.

[0004] Furthermore, there are often dependencies between teaching activities, and the completion status of the previous teaching segment will affect the execution conditions of the subsequent teaching segment. If adjustments are made simply based on whether resources are available, it is easy to cause disruptions in the teaching chain, excessive scope of adjustments, and confusion in subsequent arrangements. Summary of the Invention

[0005] This invention provides a dynamic scheduling system and method for teaching resources based on multi-agent collaboration, which solves the technical problems in related technologies where it is difficult to balance the continuity of the teaching chain, resource coordination and efficiency of local adjustments when disturbances occur during the teaching process, such as teacher unavailability, equipment failure, classroom conflicts, changes in student grouping or incomplete previous teaching tasks, and it is easy to rely on manual adjustment or overall rearrangement.

[0006] This invention provides a dynamic scheduling system for teaching resources based on multi-agent collaboration, comprising:

[0007] The basic modeling module is used to acquire the original teaching plan and teaching resource data, construct teaching chain units and directed teaching chain graphs based on the original teaching plan, construct teaching resource information tables based on the teaching resource data, and establish matching judgment rules.

[0008] The constraint generation module is used to obtain student status data and resource status data, and generate a set of constraint information based on teaching chain units, teaching resource information tables, directed teaching chain graphs, matching judgment rules, student status data, and resource status data.

[0009] The disturbance identification module is used to acquire real-time teaching disturbance events, determine the set of affected units based on the directed teaching chain graph and the set of constraint information, determine the affected range based on the set of affected units, and determine the status of teaching chain units not included in the affected range.

[0010] The candidate generation module is used to generate a set of valid composite resource candidates based on the teaching chain unit, the teaching resource information table, the matching judgment rules, the constraint information set, the affected range, and the status of the locked non-scheduling domain teaching chain unit.

[0011] The collaborative evaluation module is used to determine the evaluation results corresponding to the effective composite resource candidate combinations in the effective composite resource candidate set based on the effective composite resource candidate set, teaching chain units, and constraint information set.

[0012] The scheduling execution module is used to determine the unique optimal candidate combination of composite resources for each teaching chain unit within the affected area, form the optimal composite resource combination set, and perform local rescheduling based on the optimal composite resource combination set and the affected area to obtain the local rescheduling result and the updated resource occupation and release status.

[0013] The status update module is used to update the execution status information based on the local rescheduling results, the unique optimal combination of composite resources, the updated resource occupancy and release status, the affected scope, and the processing records corresponding to real-time teaching disturbance events, so as to obtain new execution status information.

[0014] This invention also provides a method for dynamic scheduling of teaching resources based on multi-agent collaboration, comprising the following steps:

[0015] Step 81: Obtain the original teaching plan and teaching resource data; construct teaching chain units and directed teaching chain graphs based on the original teaching plan; construct a teaching resource information table based on the teaching resource data; and establish matching judgment rules.

[0016] Step 82: Obtain student status data and resource status data, and generate a set of constraint information based on the teaching chain unit, teaching resource information table, directed teaching chain graph, matching judgment rules, student status data and resource status data;

[0017] Step 83: Obtain real-time teaching disturbance events, determine the set of affected units based on the directed teaching chain graph and the set of constraint information, determine the affected range based on the set of affected units, and determine the status of teaching chain units not included in the affected range;

[0018] Step 84: Generate a set of valid composite resource candidates based on the teaching chain unit, teaching resource information table, matching judgment rules, constraint information set, affected scope, and locked non-scheduling domain teaching chain unit status;

[0019] Step 85: Based on the effective composite resource candidate set, teaching chain units, and constraint information set, determine the evaluation results corresponding to the effective composite resource candidate combinations in the effective composite resource candidate set;

[0020] Step 86: For each teaching chain unit within the affected area, determine the corresponding unique optimal composite resource candidate combination to form an optimal composite resource combination set. Then, perform local rescheduling based on the optimal composite resource combination set and the affected area to obtain the local rescheduling result and the updated resource occupation and release status.

[0021] Step 87: Update the execution status information based on the local rescheduling results, the unique optimal composite resource combination, the updated resource occupancy and release status, the affected scope, and the processing records corresponding to real-time teaching disturbance events, to obtain new execution status information.

[0022] The beneficial effects of this invention are as follows: Based on teaching chain units, this invention structurally breaks down teaching tasks in the original teaching plan and combines resource information such as teachers, teaching content, venues, equipment, and time slices to form a composite resource scheduling method oriented towards teaching tasks. Compared with processing methods based solely on static scheduling or single resource replacement, this invention can combine student status, resource status, and the dependencies between preceding and subsequent teaching chains during the teaching execution process to form unified constraint information, and quickly identify the affected scope when disturbances such as teacher unavailability, equipment failure, time conflicts, or incomplete preceding teaching tasks occur. Furthermore, by generating an effective composite resource candidate set, performing multi-dimensional evaluation of candidate combinations, and completing local rescheduling, this invention ensures that resource adjustments no longer depend on overall rescheduling, but rather focus on the affected teaching tasks, thereby reducing interference with unaffected teaching segments. Simultaneously, this invention can also synchronously update execution status information and resource occupancy status, facilitating continued use in subsequent teaching processes and helping to improve the continuity, adaptability, and dynamic processing efficiency of teaching resource allocation. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the module of the teaching resource dynamic scheduling system based on multi-agent collaboration of the present invention;

[0024] Figure 2 This is a flowchart of the dynamic scheduling method for teaching resources based on multi-agent collaboration of the present invention. Detailed Implementation

[0025] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed only to enable those skilled in the art to better understand and implement the subject matter described herein, and changes may be made to the function and arrangement of the elements discussed without departing from the scope of this specification. Various processes or components may be omitted, substituted, or added as needed in the examples. Furthermore, features described in some examples may be combined in other examples.

[0026] like Figures 1-2 As shown, the teaching resource dynamic scheduling system based on multi-agent collaboration includes:

[0027] The basic modeling module is used to acquire the original teaching plan and teaching resource data, construct teaching chain units and directed teaching chain graphs based on the original teaching plan, construct teaching resource information tables based on the teaching resource data, and establish matching judgment rules.

[0028] The constraint generation module is used to obtain student status data and resource status data, and generate a set of constraint information based on teaching chain units, teaching resource information tables, directed teaching chain graphs, matching judgment rules, student status data, and resource status data.

[0029] The disturbance identification module is used to acquire real-time teaching disturbance events, determine the set of affected units based on the directed teaching chain graph and the set of constraint information, determine the affected range based on the set of affected units, and determine the status of teaching chain units not included in the affected range.

[0030] The candidate generation module is used to generate a set of valid composite resource candidates based on the teaching chain unit, the teaching resource information table, the matching judgment rules, the constraint information set, the affected range, and the status of the locked non-scheduling domain teaching chain unit.

[0031] The collaborative evaluation module is used to determine the evaluation results corresponding to the effective composite resource candidate combinations in the effective composite resource candidate set based on the effective composite resource candidate set, teaching chain units, and constraint information set.

[0032] The scheduling execution module is used to determine the unique optimal candidate combination of composite resources for each teaching chain unit within the affected area, form the optimal composite resource combination set, and perform local rescheduling based on the optimal composite resource combination set and the affected area to obtain the local rescheduling result and the updated resource occupation and release status.

[0033] The status update module is used to update the execution status information based on the local rescheduling results, the unique optimal combination of composite resources, the updated resource occupancy and release status, the affected scope, and the processing records corresponding to real-time teaching disturbance events, so as to obtain new execution status information.

[0034] In this invention, multi-agent collaboration refers to the system setting up multiple processing agents, each targeting different business objects and data dimensions. These agents collaboratively participate in resource analysis, constraint generation, candidate selection, evaluation, and scheduling execution around the same teaching task, resulting in a unified teaching resource scheduling outcome. In this invention, the processing agents may include student status agents, teacher resource agents, content resource agents, venue and equipment agents, and scheduling and coordination agents. Specifically, the student status agent processes information related to student learning status, target student groups, and knowledge point completion; the teacher resource agent processes information related to teacher teaching ability, teaching scope, and available time slots; the content resource agent processes information related to courseware, exercise packages, teaching content versions, and knowledge point coverage; the venue and equipment agent processes information related to classrooms, terminal equipment, capacity conditions, and equipment compatibility; and the scheduling and coordination agent summarizes, compares, and coordinates the results output by the aforementioned processing agents to generate a scheduling outcome oriented towards the teaching task.

[0035] In one embodiment of the present invention, the system acquires the original teaching plan and teaching resource data, constructs teaching chain units and a directed teaching chain graph based on the original teaching plan, constructs a teaching resource information table based on the teaching resource data, and establishes matching judgment rules. The original teaching plan represents basic information such as course arrangement, knowledge point sequence, teaching time period, target student group, and pre-course relationships. The original teaching plan may include a timetable, teaching progress table, knowledge point dependency table, class grouping information, class hour arrangement information, and information on the connection between courses. The teaching resource data represents the attributes and availability of resources such as teachers, teaching content, venues, equipment, and time slices. The teaching resource data may include teacher profile information, courseware resource information, classroom information, teaching terminal equipment information, time slice arrangement information, and resource occupancy record information. The teaching chain unit represents the smallest independently executable teaching task obtained from the original teaching plan, which serves as the basic object for subsequent multi-agent collaborative scheduling.

[0036] Specifically, in step 11, the system acquires the original teaching plan and teaching resource data, and divides the original teaching plan into multiple teaching chain units based on the knowledge point sequence information, teaching time period information, and target student group information in the original teaching plan. For each teaching chain unit, its knowledge point set, difficulty level, target student group, execution time window, and required resource type set are recorded. The knowledge point set represents the scope of teaching content covered by the teaching chain unit; the difficulty level represents the teaching depth or learning level corresponding to the teaching task; the execution time window represents the start and end time range allowed for the execution of the teaching task; and the required resource type set represents the types of resources needed to complete the teaching task. A minimum executable teaching task refers to the smallest unit of teaching activity that can be configured with teachers, content, venue, equipment, and time slices and executed independently while maintaining the integrity of the teaching semantics. For example, the explanation of function concepts can be divided into a teaching chain unit, representing that this unit is aimed at a specific student group, executed within a specified teaching time period, and requires corresponding teachers, courseware, and teaching equipment.

[0037] In step 12, the system determines the dependencies between teaching chain units based on the sequence of knowledge points, execution time windows, and prerequisite relationships recorded in the original teaching plan, and constructs a directed teaching chain graph. The directed teaching chain graph represents the sequential dependencies between multiple teaching chain units, where nodes represent teaching chain units, and directed edges indicate that the next teaching chain unit can only be executed after the previous one has been completed. By establishing this graph structure, the system can record the knowledge progression and execution sequence in a structured manner during the teaching process. Simultaneously, the system extracts resource item attributes from teaching resource data for teacher resources, content resources, venue resources, equipment resources, and time slice resources, generating a teaching resource information table. Resource item attributes represent whether a resource is suitable for a specific teaching task, such as the scope of knowledge points taught and available time slots for teacher resources, the knowledge point tags and content difficulty for content resources, the capacity of venue resources, the support capacity of equipment resources, and the start and end times of time slice resources.

[0038] In step 13, the system compares the knowledge point set, difficulty level, target student group, execution time window, and required resource type set of each teaching chain unit with the resource item attributes of each resource item in the teaching resource information table. When all attributes meet the corresponding requirements, the system determines that the resource item matches the corresponding teaching chain unit and establishes a matching judgment rule. The matching judgment rule is used to indicate the judgment conditions for whether a resource item can participate in the execution of a certain teaching chain unit. Its judgment basis includes at least the knowledge point coverage relationship, difficulty adaptation relationship, student group adaptation relationship, time availability relationship, and resource category satisfaction relationship. Furthermore, the system associates and records the teaching chain unit, directed teaching chain graph, teaching resource information table, and matching judgment rule, so that subsequent constraint generation, disturbance identification, and candidate combination generation can all call unified data around the same teaching chain unit.

[0039] Through the above implementation process, this embodiment can break down teaching tasks into computable teaching chain units, clarify the dependencies between tasks, and uniformly record resource attributes such as teachers, content, venues, equipment, and time slices, enabling each agent to collaboratively process relevant information around the same teaching chain unit. This provides a foundation for the subsequent reorganization of local resources under teaching disturbances, improving the accuracy, continuity, and adaptability of teaching resource allocation.

[0040] In one embodiment of the present invention, the system acquires student status data and resource status data, and generates a constraint information set based on teaching chain units, teaching resource information tables, directed teaching chain graphs, matching judgment rules, student status data, and resource status data. Student status data represents the learning status information of the target student group at the current teaching stage. Student status data may include student grouping information, knowledge point completion status, current adaptation difficulty information, classroom participation status, and current teaching execution status association information, etc. Resource status data represents the availability and occupancy status of teacher resources, content resources, venue resources, equipment resources, and time slice resources at the current moment. Resource status data may include resource occupancy status information, resource availability status information, time window status information, equipment operation status information, and content activation status information, etc. The constraint information set represents the constraints that the current teaching task needs to meet on the student side, resource side, and link side, thereby providing a unified basis for subsequent disturbance identification, candidate generation, and scheduling execution.

[0041] Specifically, in step 21, the system acquires student status data and resource status data, and maps the student status data to the teaching chain unit based on the target student group, completed knowledge points, difficulty level, and execution status, thus obtaining the student status record corresponding to each teaching chain unit. This mapping refers to extracting status information related to the teaching chain unit from the student status data according to the student object, knowledge point range, and execution stage corresponding to the teaching chain unit, and establishing a corresponding relationship. The student status record is used to represent the learning foundation, knowledge point completion status, and adaptation level of the student object currently corresponding to a certain teaching chain unit. For example, when a teaching chain unit is aimed at basic level students and requires students to have completed the preceding knowledge points, the system can extract the completion status of the relevant knowledge points for that group from the corresponding class or group records as the student status record for that teaching chain unit.

[0042] Simultaneously, the system maps resource status data to the teaching resource information table based on the resource categories and resource identifiers, obtaining resource status records corresponding to each resource item. These resource status records indicate whether a resource is currently occupied, available for use, and within its corresponding time window. For example, a teacher resource might correspond to a currently scheduled class status, a device resource to a normally functioning status, and a time-slice resource to an unoccupied status.

[0043] In step 22, the system compares the teaching chain unit with the corresponding student status record to generate target student group constraint information, completed knowledge point constraint information, difficulty level constraint information, and execution status constraint information. The target student group constraint information indicates whether the target student object required by the teaching chain unit matches the current student status record; the completed knowledge point constraint information indicates whether the current student group has completed the prerequisite knowledge points upon which the teaching chain unit depends; the difficulty level constraint information indicates whether the teaching difficulty of the teaching chain unit is suitable for the current student's learning level; and the execution status constraint information indicates whether the teaching chain unit is currently in a state of being able to continue execution, pending execution, or unadjustable. Through the above comparisons, the system can generate constraints directly corresponding to the teaching task from the student's perspective.

[0044] Simultaneously, the system compares resource item attributes, corresponding resource status records, and matching rules in the teaching resource information table to generate resource occupancy status constraints, resource availability constraints, time window constraints, and resource attribute constraints. Specifically, resource occupancy status constraints indicate whether the resource is currently occupied by other teaching tasks; resource availability constraints indicate whether the resource is in a state where it can participate in scheduling; time window constraints indicate whether the resource is available within the corresponding execution time range; and resource attribute constraints indicate whether the resource attributes meet the requirements of the teaching chain unit regarding knowledge points, difficulty, objects, or functional conditions. For example, when a classroom's capacity is insufficient to accommodate the target student group, corresponding resource attribute constraints can be generated.

[0045] In step 23, the system, based on the directed instructional chain graph, transmits the completed knowledge point constraint information and execution status constraint information corresponding to each instructional chain unit from the preceding instructional chain unit to the succeeding instructional chain unit, generating link constraint information. Link constraint information represents the connection restrictions formed by the sequential dependencies between instructional chain units. That is, when a preceding instructional chain unit is not yet completed, the relevant knowledge point has not yet met the entry conditions for the succeeding instructional chain unit, or the preceding instructional chain unit is currently in an uninterruptible state, the system will transmit this restriction to the succeeding instructional chain unit, subjecting it to corresponding constraints during subsequent processing.

[0046] In step 24, the system associates the constraint information corresponding to the teaching chain unit with the constraint information corresponding to the resource item in the teaching resource information table according to the matching judgment rules, generating matching constraint information. The matching constraint information indicates whether a teaching chain unit still meets the matching conditions with a resource item in its current state. In other words, even if a resource meets the requirements of the teaching chain unit in terms of static attributes, if its current occupancy state, callable state, or time window state does not meet the requirements, the system can still exclude it from the subsequent candidate range through the matching constraint information.

[0047] In step 25, the system merges the constraint information of the target student group, completed knowledge points, difficulty level, execution status, resource occupancy status, resource availability status, time window, resource attribute, link, and matching constraints to generate a constraint information set. This merging refers to the unified organization and recording of various constraints according to the teaching chain unit identifier, resource identifier, and constraint category. This allows the system to directly read all constraints corresponding to a specific teaching chain unit or resource item in subsequent processing. The constraint information set can represent both the execution restrictions of the current teaching task and the boundary conditions when resources participate in scheduling, thus providing a unified constraint basis for subsequent identification of the affected scope, candidate resource filtering, and local dynamic scheduling.

[0048] Through the above implementation process, this embodiment can transform student learning status, real-time resource status, and teaching chain dependencies into a unified set of constraint information, enabling multiple agents to invoke consistent constraint criteria around the same teaching chain unit during subsequent processing. This improves the completeness and consistency of constraint expression and provides reliable support for local rescheduling under subsequent teaching disturbances.

[0049] In one embodiment of the present invention, the system acquires real-time teaching disturbance events, determines the set of affected units based on a directed teaching chain graph and a set of constraint information, determines the affected range based on the set of affected units, and further determines the status of teaching chain units not included in the affected range. Real-time teaching disturbance events represent events that occur during teaching execution and may lead to changes in the original teaching arrangements or resource configurations. Real-time teaching disturbance events may include events such as temporary teacher unavailability, changes in teaching content, classroom occupancy conflicts, equipment failures, time slice adjustments, changes in student grouping, and incomplete preceding teaching chain units. The set of affected units represents the range of teaching chain units that need to be included in subsequent processing due to the real-time teaching disturbance events. The affected range represents a local teaching chain segment covered by the set of affected units in the directed teaching chain graph.

[0050] Specifically, in step 31, the system acquires real-time teaching disturbance events. When a real-time teaching disturbance event points to a resource item in the teaching resource information table, the system determines the event-associated teaching chain unit based on the teaching chain unit corresponding to the resource item; when a real-time teaching disturbance event directly points to a teaching chain unit, the system directly determines that teaching chain unit as the event-associated teaching chain unit. Here, the event-associated teaching chain unit refers to a teaching chain unit that is directly associated with the current disturbance event at the object level. For example, when a teacher resource is temporarily unavailable for a certain period, the system can determine the event-associated teaching chain unit based on the teaching chain unit currently associated with that teacher resource; or, for example, when the preceding knowledge point corresponding to a teaching chain unit has not been achieved, the system can directly determine that teaching chain unit as the event-associated teaching chain unit.

[0051] After identifying the teaching chain units associated with the event, the system reads the constraint information corresponding to each teaching chain unit based on the constraint information set and determines whether the constraint information is consistent with the real-time teaching disturbance event. If they are consistent, a set of directly affected units is generated. Consistency here means that the event object, event type, or state change content corresponding to the real-time teaching disturbance event corresponds to the constraint object, constraint category, or restriction condition recorded in the constraint information. For example, if the equipment resource corresponding to a device failure event is the same equipment resource that a teaching chain unit depends on, and the resource callable state constraint information of that teaching chain unit indicates that the device is a necessary execution condition, then that teaching chain unit can be included in the set of directly affected units. Through this step, the system can first identify the teaching chain units directly impacted by the disturbance event, providing a foundation for subsequent impact expansion.

[0052] In step 32, the system, based on the directed instructional chain graph, passes each instructional chain unit in the directly affected unit set to its successor instructional chain units along the dependencies between them. When a successor instructional chain unit depends on the completion status, knowledge point status, or resource status of an instructional chain unit in the directly affected unit set, the successor instructional chain unit is identified as an indirectly affected unit. Here, completion status dependency means that the execution of the successor instructional chain unit is predicated on the completion of the preceding instructional chain unit; knowledge point status dependency means that the knowledge base required by the successor instructional chain unit comes from the preceding instructional chain unit; and resource status dependency means that the execution of the successor instructional chain unit depends on the resources released or the execution conditions maintained by the preceding instructional chain unit. In other words, even if an instructional chain unit is not directly targeted by a disturbance event, as long as its execution conditions depend on already affected units, it will be included in the indirectly affected units.

[0053] Subsequently, the system merges the directly affected set of units with the indirectly affected set of units to generate the affected unit set. This process extends disturbance identification from a single event object to the local teaching chain region affected by it, allowing the dependencies between teaching tasks to be reflected in the disturbance analysis. For example, if a preceding instructional unit cannot be executed due to teacher absence, its subsequent practice and assessment units may also be included in the affected unit set because the knowledge prerequisites are not met. By propagating along the directed instructional chain graph, the system can maintain locality while avoiding missing subsequent teaching tasks actually affected by the disturbance.

[0054] In step 33, the system determines the smallest continuous teaching chain segment covering the affected unit set based on its position in the directed instruction chain graph, and defines this smallest continuous teaching chain segment as the affected range. The smallest continuous teaching chain segment represents the smallest local processing range determined while satisfying the coverage requirement of the affected unit set. In other words, this range covers both directly and indirectly affected units, without including irrelevant teaching chain units in subsequent rescheduling processing. By limiting the affected range to the smallest continuous teaching chain segment, the system can restrict subsequent adjustments within necessary boundaries, thereby reducing interference with unaffected teaching arrangements. This approach is suitable for responding to local events during the teaching process, such as local time-slice conflicts, local equipment unavailability, or local student group changes.

[0055] In step 34, the system identifies teaching chain units not located within the affected scope as those not included in the affected scope, based on the inclusion relationship between the teaching chain units and the affected scope. It then reads the current execution status and current resource occupancy status of these unaffected teaching chain units to determine their overall status. The current execution status indicates whether the teaching chain unit is currently pending execution, executing, completed, or unadjustable; the current resource occupancy status indicates the current resource usage of the teaching chain unit. By reading and retaining these teaching chain unit states, the system can maintain the teaching arrangements outside the affected scope unchanged in subsequent local scheduling and use them as the basis for subsequent conflict judgments and resource boundary verifications outside the scope. In other words, although teaching chain units not included in the affected scope do not participate in this round of rescheduling, their current status is still retained and invoked as boundary conditions, thus ensuring that local adjustments do not disrupt the original unaffected teaching arrangements.

[0056] Through the above implementation process, this embodiment can locally identify and define the scope of real-time disturbance events during the teaching process, uniformly including directly affected units, indirectly affected units, and their corresponding teaching chain segments into the affected scope, while retaining the current state of teaching chain units outside the affected scope as boundary conditions. This provides a clear processing scope and state basis for subsequent local resource reorganization and dynamic scheduling, reducing unnecessary overall adjustments and improving the continuity and stability of teaching arrangements under disturbance scenarios.

[0057] In one embodiment of the present invention, the system generates a valid composite resource candidate set based on the teaching chain unit, the teaching resource information table, the matching judgment rule, the constraint information set, the affected range, and the locked non-scheduling domain teaching chain unit status. The valid composite resource candidate set represents a set of composite resource combinations that, within the current affected range, satisfy both the teaching task requirements and the resource status and link constraint requirements. Here, composite resources refer to multiple resource combinations used to support the execution of the same teaching chain unit, and may include teacher resources, content resources, venue resources, equipment resources, and time slice resources.

[0058] Specifically, in step 41, the system extracts teaching chain units located within the affected area from all teaching chain units based on the inclusion relationship between the teaching chain units and the affected area. That is, candidate resources are generated only for the local teaching chain segments involved in this round of disturbance processing, without repeatedly generating candidates for teaching chain units outside the affected area. Subsequently, based on the knowledge point set, difficulty level, target student group, execution time window, and required resource type set of the teaching chain unit, and combined with the resource item attributes and matching rules in the teaching resource information table, the system filters teacher resources, content resources, venue resources, equipment resources, and time slice resources respectively to generate a subset of candidate resource items.

[0059] The subset of candidate resources here refers to the set of resource items that, at the static attribute level, meet the requirements of the corresponding teaching chain unit for a certain type of resource. For example, for teacher resources, filtering can be based on the scope of the knowledge points to be taught, the teaching level, and the available time slots; for content resources, filtering can be based on knowledge point tags and content difficulty; for venue resources, filtering can be based on capacity and venue type; for equipment resources, filtering can be based on support capabilities and installation conditions; and for time-slice resources, filtering can be based on the execution time window. For example, when a teaching chain unit needs to explain a specific knowledge point and is geared towards intermediate-level students, the system can first filter out teacher resource items from the teacher resources that have the ability to teach the corresponding knowledge point and are suitable for that level. Through this step, a range of candidate resources of different types that basically match the teaching task can be obtained.

[0060] In step 42, the system filters a subset of candidate resource items based on the constraint information set and the status of teaching chain units not included in the affected scope; resource items that are inconsistent with the resource occupancy status constraint information, resource callable status constraint information, time window constraint information, resource attribute constraint information, link constraint information, or matching constraint information corresponding to teaching chain units within the affected scope are removed. This filtering refers to dynamically excluding various candidate resource items based on the static matching formed in step 41, combined with the current resource status, link status, and boundary status.

[0061] The resource occupancy status constraint information indicates whether a resource is currently occupied by other teaching tasks; the resource availability status constraint information indicates whether the resource is currently allowed to be reallocated; the time window constraint information indicates whether the resource is available within the target execution time window; the resource attribute constraint information indicates whether the actual attributes of the resource meet the execution requirements of the current teaching chain unit; the link constraint information indicates the connection conditions between the teaching chain unit and the preceding and following elements in the teaching chain; and the matching constraint information indicates whether the current teaching chain unit and the current resource item still maintain a matching relationship. The status of teaching chain units not included in the affected scope is used as a boundary condition in this step to prevent the system from including resources that are stably occupied by teaching tasks outside the affected scope in the candidate scope again. For example, although a classroom's static attributes meet the capacity requirements, if it is currently occupied by a teaching chain unit outside the affected scope, the system can remove it from the subset of candidate resources in this step.

[0062] In step 43, the system selects one resource item from each of the filtered candidate resource items (teacher resources, content resources, venue resources, equipment resources, and time slice resources) and combines them to generate an initial composite resource candidate set. This initial composite resource candidate set represents the complete resource configuration candidate result formed through combination when candidate options are available at the categorized resource level. In other words, the system no longer judges whether a particular teacher or classroom is replaceable individually, but combines teachers, content, venues, equipment, and time slices as overall execution conditions for the same teaching chain unit. Through this method, subsequent scheduling targets the complete composite resource package, rather than a single resource object. For example, a teacher, a courseware, a classroom, a projection device, and an idle time slice can be combined into an initial composite resource candidate combination for subsequent verification.

[0063] In step 44, the system performs time conflict verification, site capacity verification, equipment compatibility verification, content adaptation verification, teaching chain prerequisite dependency verification, and out-of-scope occupancy conflict verification on each initial composite resource candidate combination in the initial composite resource candidate set. The initial composite resource candidate combinations that pass all verifications are retained to generate a valid composite resource candidate set. Time conflict verification determines whether there is duplicate occupancy or overlapping time period conflicts among resource items in the candidate combination within the same execution time window; site capacity verification determines whether the site resources can accommodate the target student group; equipment compatibility verification determines whether the equipment resources are compatible with the site conditions and content display requirements; content adaptation verification determines whether the content resources meet the knowledge point and difficulty requirements corresponding to the teaching chain unit; teaching chain prerequisite dependency verification determines whether the teaching chain unit meets the prerequisite execution conditions in the current link state; and out-of-scope occupancy conflict verification determines whether resource items in the candidate combination conflict with the current occupancy state of teaching chain units outside the affected range.

[0064] Through the above verification, the system can eliminate candidate combinations from the initial candidate set of composite resources that, while selectable at the individual resource level, are infeasible at the combined execution level, retaining only composite resource combinations that can truly support the execution of the corresponding teaching chain unit. For example, if the teacher and courseware in a candidate combination meet the teaching content requirements, but the device resources cannot support the interactive display function of the courseware, then the candidate combination will fail the device compatibility verification. Similarly, if a candidate combination has an available time slice, but the corresponding preceding teaching chain unit has not yet met its completion conditions, then the candidate combination will fail the preceding teaching chain dependency verification. By verifying each initial candidate composite resource combination, it can be ensured that the resource combinations subsequently entering the evaluation and ranking stages have a practical execution basis.

[0065] Through the above implementation process, this embodiment can generate composite resource candidate results that meet the current state conditions around the teaching chain unit within the affected range, and eliminate resource items and resource combinations that do not meet the execution conditions in advance through constraint filtering and combination verification. As a result, subsequent scheduling can be based on complete and executable candidate resource combinations, reducing invalid combinations entering subsequent processing and improving the accuracy and stability of local dynamic scheduling.

[0066] In one embodiment of the present invention, the system determines the evaluation result corresponding to each effective composite resource candidate combination in the effective composite resource candidate set based on the effective composite resource candidate set, teaching chain units, and constraint information set. To ensure that the evaluation result simultaneously reflects the requirements for continuous execution of teaching tasks and dynamic resource adjustment, this embodiment evaluates each effective composite resource candidate combination from three dimensions: continuity cost, disturbance cost, and coupling consistency. Here, continuity cost represents the degree of influence of the candidate combination on the continuity of the teaching chain, disturbance cost represents the degree of modification of the candidate combination relative to the original execution arrangement, and coupling consistency represents the degree of adaptation between various resources within the candidate combination.

[0067] Specifically, in step 51, the system determines the differences in knowledge point coverage, difficulty level, and link connection based on the content resources in each effective candidate combination of composite resources and the knowledge point set, difficulty level, and link constraint information in the corresponding teaching chain unit. The system then performs a weighted summation of these differences to obtain the continuity cost. The continuity cost can be expressed as: ,in, Indicating teaching chain unit Corresponding effective composite resource candidate combinations The continuous cost; This indicates differences in the coverage of knowledge points. Indicates differences in difficulty levels. Indicates differences in link connectivity. , , This represents the preset weighting coefficient, used to indicate the degree of influence of the above-mentioned difference items on the continuity cost.

[0068] The continuity cost represents the impact of adopting the effective composite resource candidate combination on the teaching connection relationship between the current teaching chain unit and its preceding and following teaching chain units. The larger the continuity cost, the greater the impact of the candidate combination on the continuous execution of the teaching chain. The knowledge point coverage difference is the number of uncovered knowledge points determined by comparing the knowledge points corresponding to the content resource with the knowledge point set of the corresponding teaching chain unit. That is, the system compares the knowledge point range contained in the content resource in the candidate combination with the knowledge point set required by the teaching chain unit item by item, and determines the number of knowledge points not covered by the content resource as the knowledge point coverage difference. The difficulty level difference is the level interval determined by comparing the difficulty level corresponding to the content resource with the difficulty level of the corresponding teaching chain unit. The level interval here is used to represent the degree of difference between the difficulty of the content resource and the target difficulty of the teaching chain unit. When the two are the same, the difficulty level difference can be zero; when the difficulty of the content resource is higher or lower than the requirement of the teaching chain unit, a level difference is formed. The link connection difference is the number of connection changes between the preceding and following teaching chain units determined according to the execution time window and the link constraint information in the constraint information set. In other words, the system determines whether the original connection between a teaching chain unit and its preceding and subsequent teaching chain units has changed based on the execution time schedule corresponding to the candidate combination, and identifies the number of changed connection items as the link connection difference. For example, if the time slice corresponding to a candidate combination is shifted later, disrupting the original connection between it and subsequent practice teaching chain units, a corresponding link connection difference can be formed. By weighted summing of the above three differences, the content coverage, difficulty matching, and link connection can be uniformly converted into a continuity cost, which is used to measure the degree to which the candidate combination maintains the continuity of teaching.

[0069] In step 52, the system determines the number of resource replacements, time slot offsets, and associated impacts based on the comparison results of teacher resources, content resources, venue resources, equipment resources, and time slice resources in each effective candidate combination of composite resources with the original execution resources and original execution time windows of the corresponding teaching chain units, as well as the link constraint information in the constraint information set. The system then performs a weighted summation of these factors to obtain the disturbance cost. The disturbance cost can be expressed as: ,in, Indicating teaching chain unit Corresponding effective composite resource candidate combinations The cost of disturbance; Indicates the number of resource replacements. Indicates the time period offset. Indicates the quantity of collateral impact. , , This represents the preset weighting coefficients, used to indicate the degree of influence of each disturbance factor on the disturbance cost.

[0070] The disturbance cost represents the degree of change caused by a candidate combination relative to the original execution arrangement. A larger disturbance cost indicates a greater change to the original teaching arrangement. The resource replacement quantity is the number of inconsistent resource items determined by comparing teacher resources, content resources, venue resources, equipment resources, and time slice resources with the original execution resources and execution time windows of the corresponding teaching chain unit. In other words, the system compares each type of resource item in the candidate combination with the originally arranged resource items of the teaching chain unit and counts the number of resource categories that have been replaced. The time slot offset is the time interval determined by comparing the time window corresponding to the time slice resource with the original execution time window of the corresponding teaching chain unit. This time interval can represent the advance, delay, or absolute time difference of the candidate time slice relative to the original execution time window, reflecting the degree of change in the time arrangement of teaching activities. The number of associated effects is the number of other teaching chain units that need to be adjusted synchronously, based on the link constraint information in the constraint information set. In other words, when a candidate combination is adopted, if it causes coordinated adjustments in the scheduling, knowledge connection, or resource release of preceding and following teaching chain units, the system counts the number of other teaching chain units that need to be adjusted simultaneously and determines this number as the number of associated impacts. For example, if the execution time of a teaching chain unit is postponed, causing subsequent assessment and review teaching chain units to be delayed, the number of these subsequent teaching chain units can be included in the number of associated impacts. By weighted summing the number of resource replacements, time period offsets, and the number of associated impacts, the scope and degree of changes caused by the candidate combination to the original plan can be uniformly converted into disturbance costs.

[0071] In step 53, based on the comparison results among teacher resources, content resources, venue resources, and equipment resources in each valid candidate combination of composite resources, the system determines the adaptation value of teacher resources to the target student group, the matching value of teacher resources to content resources, the compatibility value of venue resources to equipment resources, and the consistency value of content resources to equipment resources. The system then performs a weighted summation of these values ​​to obtain the coupling consistency. This coupling consistency can be expressed as: ,in, Indicating teaching chain unit Corresponding effective composite resource candidate combinations Coupling consistency; This indicates the fit between teacher resources and the target student group. This indicates the matching value between teacher resources and content resources. This indicates the compatibility value between site resources and equipment resources. This indicates that the content resources and device resources have the same capacity. , , , This represents the preset weighting coefficient, used to indicate the degree of influence of each adaptor on coupling consistency.

[0072] The coupling consistency is used to represent the degree of coordination and adaptation among the various resources constituting the candidate combination when executing the same teaching chain unit. The higher the coupling consistency, the more coordinated the resources within the candidate combination. The teacher resource and target student group adaptation value is the number of overlapping object items determined by comparing the scope of the teacher resource's corresponding teaching audience with the target student group of the corresponding teaching chain unit. In other words, the system compares the scope of the teacher's teaching audience with the student group targeted by the teaching chain unit and counts the number of matching object items. The teacher resource and content resource matching value is the number of overlapping items determined by comparing the scope of the teacher's corresponding teaching knowledge points and teaching difficulty level with the scope of the content resource's corresponding knowledge points and difficulty level. This method can determine whether the teacher resources and content resources are coordinated in terms of teaching content and teaching level. The venue resource and equipment resource compatibility value is the number of items that meet the conditions determined by comparing the capacity and installation conditions of the venue resources with the occupancy and installation conditions of the equipment resources. The content resource and equipment resource carrying capacity consistency value is the number of items that meet the requirements determined by comparing the display and interaction requirements of the content resources with the support capabilities of the equipment resources. For example, when a content resource requires interactive display, and the candidate device resource has touch interaction and display capabilities, a corresponding conformity item can be formed in the consistency value of the content resource and the device resource. By weighted summing the above adaptation and compatibility values, the degree of cooperation between different resources within the composite resource can be uniformly represented as coupling consistency, thereby reflecting the level of collaboration of the candidate combination at the overall execution level.

[0073] In step 54, the system associates and records the continuity cost, perturbation cost, and coupling consistency corresponding to each effective candidate combination of composite resources, generating evaluation results for each candidate combination. This association and recording refers to establishing a one-to-one correspondence between the results of each evaluation dimension and the corresponding candidate combination based on the teaching chain unit identifier and the candidate combination identifier, ensuring that each effective candidate combination of composite resources has a complete evaluation record. The evaluation results may include the candidate combination identifier, the corresponding teaching chain unit identifier, the continuity cost, the perturbation cost, and the coupling consistency value.

[0074] Through the above implementation process, this embodiment can form a unified evaluation result for each effective candidate combination of composite resources, based on the continuity of the teaching chain, the degree of disturbance to the original plan, and the degree of internal synergy of composite resources. This allows subsequent scheduling selection to be based on a multi-dimensional and structured comparison. Therefore, it avoids making scheduling judgments solely based on the availability of a single resource, and helps to balance teaching continuity, adjustment scope, and resource combination suitability in scenarios of teaching disturbance, thereby improving the rationality and adaptability of local dynamic scheduling results.

[0075] In one embodiment of the present invention, the system determines a unique optimal candidate combination of composite resources for each teaching chain unit within the affected area, forming an optimal composite resource combination set. Based on the optimal composite resource combination set and the affected area, local rescheduling is performed to obtain the local rescheduling result and the updated resource occupancy and release status. The optimal composite resource combination set represents the target resource configuration result for each teaching chain unit within the affected area under the current disturbance conditions. Local rescheduling indicates that adjustments are made only to the original execution resources and execution time windows within the affected area, while the original arrangement remains unchanged for teaching chain units outside the affected area. Through this process, the system can perform deterministic replacement and status update of locally disturbed teaching tasks without altering the overall teaching arrangement.

[0076] Specifically, in step 61, for each teaching chain unit within the affected area, the system associates each effective composite resource candidate combination in the effective composite resource candidate set corresponding to that teaching chain unit with its corresponding continuity cost, disturbance cost, and coupling consistency, and sorts them in the order of minimum continuity cost, minimum disturbance cost, and maximum coupling consistency. This sorting is performed separately for each effective composite resource candidate combination corresponding to the same teaching chain unit, rather than uniformly sorting candidate combinations for different teaching chain units. In other words, the system first selects candidate combinations with lower continuity costs from the perspective of maintaining teaching chain continuity; then, when continuity costs are the same, it selects candidate combinations that require less modification to the original arrangement; and finally, when the aforementioned two conditions are the same, it selects candidate combinations with a higher degree of cooperation between resources.

[0077] When continuity cost, disturbance cost, and coupling consistency are all the same, the system determines the unique optimal composite resource candidate combination corresponding to the teaching chain unit according to the order of resource item identifiers in each effective composite resource candidate combination. The order of resource item identifiers represents the pre-defined arrangement of identifiers for various resource items in the system, such as the combined arrangement order of teacher resource identifiers, content resource identifiers, venue resource identifiers, equipment resource identifiers, and time slice resource identifiers. In this way, a unique result can be obtained even when multiple candidate combinations have the same evaluation result, avoiding the coexistence of multiple equivalent candidate combinations in subsequent scheduling and execution.

[0078] In step 62, the system establishes a correspondence between each unique optimal composite resource candidate combination and its corresponding teaching chain unit based on the optimal composite resource combination set and the affected range, according to the teaching chain unit identifier. The system then replaces the original execution resources and execution time windows of the corresponding teaching chain unit with the teacher resources, content resources, venue resources, equipment resources, and time slice resources from each unique optimal composite resource candidate combination, generating a local rescheduling result. This correspondence represents a one-to-one mapping between each teaching chain unit and its final selected resource combination, ensuring that subsequent execution replacements have a clear target. The local rescheduling result represents the execution resources and execution time arrangements of each teaching chain unit within the affected range after this round of adjustments. For example, when a teaching chain unit's original teacher resources are temporarily unavailable and replaced with another teacher resource, and its execution time window is adjusted to an adjacent idle time slice, the system can record the replaced teacher resources, content resources, venue resources, equipment resources, and time slice resources together as the local rescheduling result for that teaching chain unit.

[0079] In step 63, based on the local rescheduling results, the system reads the original and replaced execution resources of the corresponding teaching chain unit. Original execution resources no longer occupied by teaching chain units within the affected range are identified as released resources, and replaced execution resources already occupied by teaching chain units within the affected range are identified as occupied resources. Resource items with identical original and replaced execution resources are kept in their original states, generating updated resource occupancy and release states. Here, released resources refer to resources that will no longer undertake the execution tasks of the corresponding teaching chain unit after this round of local rescheduling; occupied resources refer to resources newly used to undertake the execution tasks of the corresponding teaching chain unit after this round of local rescheduling. If a resource item remains unchanged before and after the adjustment, it does not need to be released or re-occupied, but rather continues to maintain its original resource state.

[0080] In step 64, the system establishes a correspondence between teaching chain units, execution resources, execution time windows, resource release states, and resource occupancy states based on the local rescheduling results and the updated resource occupancy and release states. It also verifies whether the unique optimal composite resource candidate combination originates from the valid composite resource candidate set, whether the execution resources in the local rescheduling results originate from the unique optimal composite resource candidate combination, and whether the updated resource occupancy and release states are consistent with the local rescheduling results. This verification ensures data consistency between the preceding candidate selection, execution replacement, and resource state updates. In other words, the system not only completes rescheduling and state updates but also performs consistency checks on the resource sources, replacement results, and state changes used to avoid discrepancies between candidate selection results and actual execution results, or between resource state records and actual scheduling results.

[0081] Through the above implementation process, this embodiment can form a unique and optimal candidate combination of composite resources for each teaching chain unit within the affected area, and transform the optimal candidate combination of composite resources into a local rescheduling result and the corresponding resource state change result. Therefore, the resource adjustment process after teaching disturbances has clear selection rules, execution results, and state write-back basis, which helps reduce uncertainty in the local adjustment process and improves the stability and consistency of local dynamic scheduling.

[0082] In one embodiment of the present invention, the system updates the execution status information based on the local rescheduling results, the unique optimal combination of composite resources, the updated resource occupancy and release status, the affected scope, and the processing records corresponding to real-time teaching disturbance events, thus obtaining new execution status information. The execution status information represents the execution resources, execution time window, resource status, and event processing results corresponding to each teaching chain unit in the current scheduling round. Execution status entries represent the status records corresponding to a single teaching chain unit in the execution status information.

[0083] Specifically, in step 71, the system, based on the inclusion relationship between the affected scope and the teaching chain units corresponding to each execution status entry in the execution status information, determines the execution status entries corresponding to the teaching chain units within the affected scope as execution status entries to be updated, and determines the execution status entries corresponding to the teaching chain units not within the affected scope as inherited status entries. Here, execution status entries to be updated refer to status records that need to be rewritten due to the current round of real-time teaching disturbance events and local rescheduling; inherited status entries refer to status records that are not affected by the current round of local scheduling and continue to maintain their original content in the new execution status information.

[0084] In step 72, the system, based on the local rescheduling results, the unique optimal combination of composite resources, and the updated resource occupancy and release status, maps the execution resources, execution time windows, resource occupancy status, resource release status, original plan mapping information, and new plan mapping information according to the teaching chain unit identifier, generating an updated status entry corresponding to the execution status entry to be updated. Here, the original plan mapping information represents the original execution resources and original execution time arrangement corresponding to the teaching chain unit before the adjustment, while the new plan mapping information represents the execution resources and execution time arrangement corresponding to the teaching chain unit after this round of local rescheduling.

[0085] In step 73, the system, based on the processing record corresponding to the real-time teaching disturbance event, writes the event type, processing method, and processing completion status that are consistent with the teaching chain unit corresponding to the update status entry in the processing record. The processing record is used to represent the processing process information of the real-time teaching disturbance event in the system in this round, and may include event type, event object, processing action, processing result, and completion status.

[0086] In step 74, the system deduplicates and merges inherited state entries and updated state entries written to the processing record according to the teaching chain unit identifier. State entries with the same teaching chain unit identifier are replaced with updated state entries. State entries corresponding to teaching chain units not within the affected range retain their inheritance, and new state time identifiers and state version identifiers are written to them, resulting in new execution status information. The state time identifier indicates the update time of the state entry, and the state version identifier indicates the state version to which the state entry belongs. Through deduplication, merging, and version writing, the system can form a latest execution status information covering all teaching chain units, enabling teaching chain units within the affected range to complete state replacement, while teaching chain units outside the affected range remain unchanged, thereby ensuring the integrity and consistency of the entire teaching execution status after this round of scheduling.

[0087] Through the above implementation process, this embodiment can uniformly write the results of local rescheduling, resource status changes, and disturbance event handling into the execution status information, forming a new state basis. On the one hand, it can synchronously update the execution resources, execution time, and resource occupancy after this round of scheduling; on the other hand, it can also retain the event source, processing method, and mapping relationship between previous and subsequent plans, providing a basis for subsequent teaching disturbance handling, status tracking, and scheduling result verification, thereby improving the consistency and traceability of the state in the dynamic scheduling process.

[0088] This embodiment also provides a method for dynamic scheduling of teaching resources based on multi-agent collaboration, including the following steps:

[0089] Step 81: Obtain the original teaching plan and teaching resource data; construct teaching chain units and directed teaching chain graphs based on the original teaching plan; construct a teaching resource information table based on the teaching resource data; and establish matching judgment rules.

[0090] Step 82: Obtain student status data and resource status data, and generate a set of constraint information based on the teaching chain unit, teaching resource information table, directed teaching chain graph, matching judgment rules, student status data and resource status data;

[0091] Step 83: Obtain real-time teaching disturbance events, determine the set of affected units based on the directed teaching chain graph and the set of constraint information, determine the affected range based on the set of affected units, and determine the status of teaching chain units not included in the affected range;

[0092] Step 84: Generate a set of valid composite resource candidates based on the teaching chain unit, teaching resource information table, matching judgment rules, constraint information set, affected scope, and locked non-scheduling domain teaching chain unit status;

[0093] Step 85: Based on the effective composite resource candidate set, teaching chain units, and constraint information set, determine the evaluation results corresponding to the effective composite resource candidate combinations in the effective composite resource candidate set;

[0094] Step 86: For each teaching chain unit within the affected area, determine the corresponding unique optimal composite resource candidate combination to form an optimal composite resource combination set. Then, perform local rescheduling based on the optimal composite resource combination set and the affected area to obtain the local rescheduling result and the updated resource occupation and release status.

[0095] Step 87: Update the execution status information based on the local rescheduling results, the unique optimal composite resource combination, the updated resource occupancy and release status, the affected scope, and the processing records corresponding to real-time teaching disturbance events, to obtain new execution status information.

[0096] The embodiments of the present invention have been described above, but the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms based on the guidance of the present embodiments, all of which are within the protection scope of the present embodiments.

Claims

1. A dynamic scheduling system for teaching resources based on multi-agent collaboration, characterized in that, include: The basic modeling module is used to acquire the original teaching plan and teaching resource data, construct teaching chain units and directed teaching chain graphs based on the original teaching plan, construct teaching resource information tables based on the teaching resource data, and establish matching judgment rules. The constraint generation module is used to obtain student status data and resource status data, and generate a set of constraint information based on teaching chain units, teaching resource information tables, directed teaching chain graphs, matching judgment rules, student status data, and resource status data. The disturbance identification module is used to acquire real-time teaching disturbance events, determine the set of affected units based on the directed teaching chain graph and the set of constraint information, determine the affected range based on the set of affected units, and determine the status of teaching chain units not included in the affected range. The candidate generation module is used to generate a set of valid composite resource candidates based on the teaching chain unit, the teaching resource information table, the matching judgment rules, the constraint information set, the affected range, and the status of the locked non-scheduling domain teaching chain unit. The collaborative evaluation module is used to determine the evaluation results corresponding to the effective composite resource candidate combinations in the effective composite resource candidate set based on the effective composite resource candidate set, teaching chain units, and constraint information set. The scheduling and execution module is used to determine the unique optimal candidate combination of composite resources for each teaching chain unit within the affected area, form the optimal composite resource combination set, and perform local rescheduling based on the optimal composite resource combination set and the affected area to obtain the local rescheduling result and the updated resource occupation and release status. The status update module is used to update the execution status information based on the local rescheduling results, the unique optimal combination of composite resources, the updated resource occupancy and release status, the affected scope, and the processing records corresponding to real-time teaching disturbance events, so as to obtain new execution status information.

2. The teaching resource dynamic scheduling system based on multi-agent collaboration according to claim 1, characterized in that, Obtain the original teaching plan and teaching resource data; construct teaching chain units and directed teaching chain graphs based on the original teaching plan; construct a teaching resource information table based on the teaching resource data; and establish matching judgment rules, including: Step 11: Obtain the original teaching plan and teaching resource data. Based on the knowledge point sequence information, teaching time information and target student group information in the original teaching plan, divide the original teaching plan into teaching chain units according to the minimum executable teaching tasks, and record the knowledge point set, difficulty level, target student group, execution time window and required resource type set for each teaching chain unit. Step 12: Based on the sequential relationship of knowledge points between teaching chain units, the order of execution time windows, and the preconditions recorded in the original teaching plan, determine the preconditions and preconditions between each teaching chain unit and construct a directed teaching chain graph; and extract the resource item attributes of teacher resources, content resources, venue resources, equipment resources, and time slice resources from the teaching resource data to generate a teaching resource information table; Step 13: Based on the knowledge point set, difficulty level, target student group, execution time window, and required resource type set of each teaching chain unit, compare them with the resource item attributes of each resource item in the teaching resource information table; when all attributes meet the corresponding requirements, determine that the resource item matches the corresponding teaching chain unit, establish matching judgment rules, and associate and record the teaching chain unit, directed teaching chain graph, teaching resource information table, and matching judgment rules.

3. The teaching resource dynamic scheduling system based on multi-agent collaboration according to claim 1, characterized in that, Obtain student status data and resource status data. Based on the teaching chain unit, teaching resource information table, directed teaching chain graph, matching judgment rules, student status data, and resource status data, generate a set of constraint information, including: Step 21: Obtain student status data and resource status data. Based on the target student group, completed knowledge points, difficulty level, and execution status of the teaching chain unit, map the student status data to the teaching chain unit to obtain the student status record corresponding to each teaching chain unit. Based on the resource category and resource identifier in the teaching resource information table, map the resource status data to the teaching resource information table to obtain the resource status record corresponding to each resource item. Step 22: Based on the comparison between the teaching chain unit and the corresponding student status record, generate target student group constraint information, completed knowledge point constraint information, difficulty level constraint information, and execution status constraint information; based on the comparison between the resource item attributes in the teaching resource information table, the corresponding resource status record, and the matching judgment rule, generate resource occupancy status constraint information, resource callable status constraint information, time window constraint information, and resource attribute constraint information. Step 23: Based on the directed instruction chain graph, the completed knowledge point constraint information and execution status constraint information corresponding to each instruction chain unit are transmitted from the preceding instruction chain unit to the subsequent instruction chain unit to generate link constraint information; Step 24: According to the matching judgment rules, associate the constraint information corresponding to the teaching chain unit with the constraint information corresponding to the resource item in the teaching resource information table to generate matching constraint information; Step 25: Merge the constraint information of the target student group, the constraint information of the completed knowledge points, the constraint information of the difficulty level, the constraint information of the execution status, the constraint information of the resource occupancy status, the constraint information of the resource callable status, the constraint information of the time window, the constraint information of the resource attribute, the constraint information of the link, and the constraint information of the matching to generate a constraint information set.

4. The teaching resource dynamic scheduling system based on multi-agent collaboration according to claim 1, characterized in that, Acquire real-time teaching disturbance events, determine the set of affected units based on the directed instructional chain graph and constraint information set, determine the affected scope based on the affected unit set, and determine the status of instructional chain units not included in the affected scope, including: Step 31: Obtain real-time teaching disturbance events. When a real-time teaching disturbance event points to a resource item in the teaching resource information table, determine the event-associated teaching chain unit based on the teaching chain unit corresponding to the resource item. When a real-time teaching disturbance event points to a teaching chain unit, directly determine the event-associated teaching chain unit. Read the constraint information corresponding to each event-associated teaching chain unit based on the constraint information set. When the constraint information is consistent with the real-time teaching disturbance event, generate a set of directly affected units. Step 32: Based on the directed instructional chain graph, each instructional chain unit in the directly affected unit set is passed to the subsequent instructional chain unit along the preceding and following dependencies; when the subsequent instructional chain unit depends on the completion status, knowledge point status, or resource status of the instructional chain units in the directly affected unit set, the subsequent instructional chain unit is identified as an indirectly affected unit; the directly affected unit set and the indirectly affected unit set are merged to generate the affected unit set. Step 33: Based on the position of the affected unit set in the directed instruction chain graph, determine the smallest continuous instruction chain segment covering the affected unit set, and define the smallest continuous instruction chain segment as the affected range; Step 34: Based on the inclusion relationship between the teaching chain unit and the affected scope, determine the teaching chain units that are not within the affected scope as teaching chain units not included in the affected scope; read the current execution status and current resource occupancy status of the teaching chain units not included in the affected scope to determine the status of the teaching chain units not included in the affected scope.

5. The teaching resource dynamic scheduling system based on multi-agent collaboration according to claim 1, characterized in that, Based on the teaching chain unit, teaching resource information table, matching judgment rules, constraint information set, affected scope, and locked non-scheduling domain teaching chain unit status, a valid composite resource candidate set is generated, including: Step 41: Based on the inclusion relationship between the teaching chain unit and the affected area, extract the teaching chain units located within the affected area from the teaching chain units; based on the knowledge point set, difficulty level, target student group, execution time window and required resource type set of the teaching chain unit, and combined with the resource item attributes and matching judgment rules in the teaching resource information table, filter teacher resources, content resources, venue resources, equipment resources and time slice resources respectively to generate a subset of candidate resource items; Step 42: Filter the subset of candidate resource items based on the set of constraint information and the status of teaching chain units not included in the affected scope; remove resource items that are inconsistent with the resource occupancy status constraint information, resource callable status constraint information, time window constraint information, resource attribute constraint information, link constraint information, or matching constraint information corresponding to the teaching chain units located within the affected scope. Step 43: Based on the filtered subset of candidate resource items, select one resource item from each of the teacher resources, content resources, venue resources, equipment resources, and time slice resources to combine them and generate an initial composite resource candidate set; Step 44: Perform time conflict verification, site capacity verification, equipment compatibility verification, content adaptation verification, teaching chain prerequisite dependency verification, and out-of-scope occupation conflict verification on each initial composite resource candidate combination in the initial composite resource candidate set; retain the initial composite resource candidate combinations that pass all verifications to generate a valid composite resource candidate set.

6. The teaching resource dynamic scheduling system based on multi-agent collaboration according to claim 1, characterized in that, Based on the effective candidate set of composite resources, the teaching chain units, and the constraint information set, the evaluation results corresponding to the effective composite resource candidate combinations in the effective candidate set are determined, including: Step 51: Based on the content resources in each effective composite resource candidate combination and the knowledge point set, difficulty level, and link constraint information in the corresponding teaching chain unit, determine the differences in knowledge point coverage, difficulty level, and link connection, and then perform a weighted summation of the differences in knowledge point coverage, difficulty level, and link connection to obtain the continuity cost. Step 52: Based on the comparison results of teacher resources, content resources, venue resources, equipment resources and time slice resources in each effective composite resource candidate combination with the original execution resources and original execution time windows of the corresponding teaching chain unit, as well as the link constraint information in the constraint information set, determine the number of resource replacements, time period offsets and the number of associated effects, and perform a weighted summation of the number of resource replacements, time period offsets and the number of associated effects to obtain the disturbance cost. Step 53: Based on the comparison results among teacher resources, content resources, venue resources, and equipment resources in each effective candidate combination of composite resources, determine the adaptation value of teacher resources to the target student group, the matching value of teacher resources to content resources, the compatibility value of venue resources to equipment resources, and the consistency value of content resources to equipment resources. Then, perform a weighted summation of the adaptation value of teacher resources to the target student group, the matching value of teacher resources to content resources, the compatibility value of venue resources to equipment resources, and the consistency value of content resources to equipment resources to obtain the coupling consistency. Step 54: Correlate and record the continuity cost, disturbance cost, and coupling consistency corresponding to each effective composite resource candidate combination to generate the evaluation result corresponding to each effective composite resource candidate combination.

7. The teaching resource dynamic scheduling system based on multi-agent collaboration according to claim 6, characterized in that, The methods for determining each parameter in steps 51 to 53 include: The knowledge point coverage difference is determined by comparing the knowledge points corresponding to the content resources with the knowledge point set of the corresponding teaching chain unit to determine the number of uncovered knowledge points; The difficulty level difference is the level interval determined by comparing the difficulty level of the content resource with the difficulty level of the corresponding teaching chain unit. The link connection difference is the number of connection changes between the preceding and subsequent teaching chain units determined based on the execution time window and the link constraint information in the constraint information set; The number of resource replacements is the number of inconsistent resource items determined by comparing teacher resources, content resources, venue resources, equipment resources, and time slice resources with the original execution resources and original execution time windows of the corresponding teaching chain unit item by item. The time slot offset is the time interval determined by comparing the time window corresponding to the time slice resource with the original execution time window of the corresponding teaching chain unit; The number of cascading effects is the number of teaching chain units that need to be adjusted synchronously, excluding the current teaching chain unit, based on the link constraint information in the constraint information set. The teacher resource and target student group fit value is the number of overlapping objects determined by comparing the range of teaching objects corresponding to the teacher resource with the target student group of the corresponding teaching chain unit. The matching value between teacher resources and content resources is the number of overlapping items determined by comparing the range of knowledge points and difficulty level of the teacher resources with the range of knowledge points and difficulty level of the content resources. The compatibility value between site resources and equipment resources is the number of items that meet the conditions determined by comparing the capacity and installation conditions of site resources with the occupancy and installation conditions of equipment resources. The consistency value between content resources and device resources is the number of items that meet the display and interaction requirements of the content resources and the supporting capabilities of the device resources.

8. The teaching resource dynamic scheduling system based on multi-agent collaboration according to claim 1, characterized in that, For each teaching chain unit within the affected area, a unique optimal candidate combination of composite resources is determined, forming an optimal composite resource combination set. Based on this optimal composite resource combination set and the affected area, local rescheduling is performed to obtain the local rescheduling results and the updated resource occupancy and release status, including: Step 61: For each teaching chain unit within the affected range, associate each effective composite resource candidate combination in the effective composite resource candidate set corresponding to the teaching chain unit with the corresponding continuity cost, disturbance cost, and coupling consistency, and sort them in the order of minimum continuity cost, minimum disturbance cost, and maximum coupling consistency; when the continuity cost, disturbance cost, and coupling consistency are all the same, determine the unique optimal composite resource candidate combination corresponding to the teaching chain unit according to the resource item identification order in each effective composite resource candidate combination; Step 62: Based on the set of optimal composite resource combinations and the affected range, establish the correspondence between each unique optimal composite resource candidate combination and the corresponding teaching chain unit according to the teaching chain unit identifier; replace the original execution resources and original execution time windows of the corresponding teaching chain unit with the teacher resources, content resources, venue resources, equipment resources and time slice resources in each unique optimal composite resource candidate combination, and generate local rescheduling results. Step 63: Based on the local rescheduling results, read the original execution resources and the replaced execution resources of the corresponding teaching chain unit; determine the original execution resources that are no longer occupied by teaching chain units within the affected range as released resources, determine the replaced execution resources that have been occupied by teaching chain units within the affected range as occupied resources, and keep the resource items that are the same as the original execution resources and the replaced execution resources in their original resource states, and generate the updated resource occupation and release states; Step 64: Based on the local rescheduling results and the updated resource occupancy and release status, establish the correspondence between teaching chain units, execution resources, execution time windows, resource release status, and resource occupancy status; verify whether the unique optimal composite resource combination comes from the effective composite resource candidate set, whether the execution resources in the local rescheduling results come from the unique optimal composite resource combination, and whether the updated resource occupancy and release status is consistent with the local rescheduling results.

9. The teaching resource dynamic scheduling system based on multi-agent collaboration according to claim 1, characterized in that, Based on the local rescheduling results, the unique optimal combination of composite resources, the updated resource occupancy and release status, the affected scope, and the processing records corresponding to real-time teaching disturbance events, the execution status information is updated to obtain new execution status information, including: Step 71: Based on the inclusion relationship between the affected scope and the teaching chain units corresponding to each execution status entry in the execution status information, determine the execution status entries corresponding to the teaching chain units within the affected scope as execution status entries to be updated, and determine the execution status entries corresponding to the teaching chain units not within the affected scope as inherited status entries. Step 72: Based on the local rescheduling results, the unique optimal composite resource combination, and the updated resource occupancy and release status, match the execution resources, execution time windows, resource occupancy status, resource release status, original plan mapping information, and new plan mapping information according to the teaching chain unit identifier, and generate an updated status entry corresponding to the execution status entry to be updated. Step 73: Based on the processing record corresponding to the real-time teaching disturbance event, write the event type, processing method and processing completion status that are consistent with the teaching chain unit corresponding to the update status entry in the processing record into the corresponding update status entry. Step 74: Deduplicatively merge the inherited status entries and the updated status entries written to the processing record according to the teaching chain unit identifier. Replace the original status entries with updated status entries for status entries with the same teaching chain unit identifier. Keep the inherited status entries for teaching chain units not in the affected range unchanged, and write the new status time identifier and status version identifier to obtain the new execution status information.

10. A method for dynamic scheduling of teaching resources based on multi-agent collaboration, characterized in that, The teaching resource dynamic scheduling system based on multi-agent collaboration as described in any one of claims 1-9 includes the following steps: Step 81: Obtain the original teaching plan and teaching resource data; construct teaching chain units and directed teaching chain graphs based on the original teaching plan; construct a teaching resource information table based on the teaching resource data; and establish matching judgment rules. Step 82: Obtain student status data and resource status data, and generate a set of constraint information based on the teaching chain unit, teaching resource information table, directed teaching chain graph, matching judgment rules, student status data and resource status data; Step 83: Obtain real-time teaching disturbance events, determine the set of affected units based on the directed teaching chain graph and the set of constraint information, determine the affected range based on the set of affected units, and determine the status of teaching chain units not included in the affected range; Step 84: Generate a set of valid composite resource candidates based on the teaching chain unit, teaching resource information table, matching judgment rules, constraint information set, affected scope, and locked non-scheduling domain teaching chain unit status; Step 85: Based on the effective composite resource candidate set, teaching chain units, and constraint information set, determine the evaluation results corresponding to the effective composite resource candidate combinations in the effective composite resource candidate set; Step 86: For each teaching chain unit within the affected area, determine the corresponding unique optimal composite resource candidate combination to form an optimal composite resource combination set. Then, perform local rescheduling based on the optimal composite resource combination set and the affected area to obtain the local rescheduling result and the updated resource occupation and release status. Step 87: Update the execution status information based on the local rescheduling results, the unique optimal composite resource combination, the updated resource occupancy and release status, the affected scope, and the processing records corresponding to real-time teaching disturbance events to obtain new execution status information.

Images