Intelligent course arrangement method based on multi-constraint hierarchical optimization and dynamic backtracking

Abstract

The invention provides an intelligent course arrangement method based on multi-constraint hierarchical optimization and dynamic backtracking, and is applied to the technical field of data processing. According to the method, basic data of setups, classes, courses and the like are firstly obtained, and a multi-dimensional constraint model is established through preprocessing and is quantified into a time slot state value; based on a two-level layering strategy and dynamic teacher priority scheduling, independent classes are processed firstly, then sub-classes are processed, conflicts are cleared, and a stage course arrangement result is generated. In combination with the weekly class hours of the courses, time is distributed through an intelligent hall connection and degradation strategy, conflicts are detected in real time, and a backtracking mechanism is triggered. And verifying the conflict-free persistent data, and generating an initial course arrangement scheme. And finally, according to indexes such as course arrangement success rate and the like and application feedback, a constraint model, a scheduling algorithm and a class connection strategy are optimized, a multi-dimensional coupled intelligent course arrangement optimization scheme is formed, and the course arrangement problem under complex constraints is efficiently solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to an intelligent course scheduling method based on multi-constraint hierarchical optimization and dynamic backtracking. BACKGROUND

[0002] The existing automatic course scheduling algorithm and system generally have the following technical defects: Traditional algorithms can only solve simple space-time conflicts, such as repeated course scheduling at the same time and at the same place, lack effective modeling and processing mechanisms for complex business constraints, and cannot adapt to multiple constraint scenarios such as "not allowed to teach", "priority to teach", cross-campus teaching restrictions, resource group capacity restrictions (such as the number of classes that a laboratory can accommodate at the same time), etc., resulting in course scheduling results that are difficult to meet actual teaching business rules.

[0003] In complex scenarios with multiple constraints, traditional algorithms are prone to local optimal solutions and cannot flexibly adjust the course scheduling strategy, often resulting in course scheduling failures or course scheduling results that require a lot of manual intervention to correct, which seriously affects the efficiency of course scheduling. Lack of intelligent scheduling mechanism for teacher workload, without considering the dynamic matching of teacher constraint strictness and existing course scheduling volume, easily leading to problems such as overloading of some teachers and underutilization or overloading of resources of some teachers, violating the principle of balanced allocation of teaching resources.

[0004] In the face of large-scale course scheduling needs, such as multi-grade, multi-campus, and multi-course parallel course scheduling, the time complexity of traditional exhaustive methods or simple greedy algorithms is high, and the operation time is long, making it difficult to meet the efficient course scheduling needs in actual applications. Lack of intelligent continuous class combination strategy for different weekly course hours, unable to dynamically adapt to reasonable continuous class schemes such as 4 continuous classes and 3 continuous classes according to the course weekly hour needs, resulting in scattered course arrangement and affecting the continuity of teaching effectiveness. Without designing a dedicated course scheduling mechanism for the hierarchical relationship between independent classes and sub-classes, it is easy to cause course conflicts between parent and child classes, and cannot guarantee the consistency of class-level teaching.

[0005] It should be noted that the information disclosed in the above background section is only used to strengthen the understanding of the background of the present disclosure, and therefore can include information that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0006] Other characteristics and advantages of the present application will become apparent from the following detailed description, or will be learned by practice of the application.

[0007] According to one aspect of the present application, an intelligent class scheduling method based on multi-constraint hierarchical optimization and dynamic backtracking is provided, comprising: obtaining basic data including semester, class, course, teacher, role, and resource group; preprocessing the basic data, establishing a role, campus, resource group, and class hour multi-dimensional constraint model, and converting various constraints into specific class scheduling rules and time period markers by modeling the time slot state values; based on a two-level hierarchical class scheduling strategy and a dynamic teacher priority scheduling algorithm, in combination with the multi-dimensional constraint model, the independent class scheduling is preferentially processed, and the sub-class scheduling is then processed and the conflicts are cleaned up, the teachers with strict constraints and less class scheduling are preferentially arranged, and the target stage class scheduling execution result is generated; according to the course weekly class hour requirement, intelligent course combination and degradation strategies are adopted to dynamically adapt the continuous class schemes including 4 continuous classes and 3 continuous classes, complete time allocation and real-time detection of various conflicts, trigger the intelligent backtracking mechanism to roll back resources and try alternative schemes when conflicts occur, and generate time allocation and conflict resolution execution results; the time allocation and conflict resolution execution results are checked, and after confirming that there is no conflict, the results are persisted, the teacher class scheduling period count and the course table state are updated, and the initial class scheduling scheme is generated; based on the evaluation indexes of the class scheduling success rate, resource utilization efficiency, and class scheduling quality, in combination with the verification results of the initial class scheduling scheme, the changes in business rules, and the actual application feedback, the quantization rules and state mapping mechanism of the multi-dimensional constraint model, the weight calculation logic of the dynamic teacher priority, and the adaptation and degradation strategies of the continuous class combination are optimized, and the intelligent class scheduling optimization scheme is generated through multi-dimensional parameter coupling optimization.

[0008] The intelligent class scheduling method based on multi-constraint hierarchical optimization and dynamic backtracking provided by the present application integrates multi-dimensional constraint modeling, hierarchical class scheduling strategy, dynamic priority scheduling, and intelligent backtracking mechanism, first obtains basic data such as semester, class, course, and teacher, establishes a role, campus, resource group, and class hour multi-dimensional constraint model through preprocessing, and quantifies various constraints into 1-6 level time slot state values. A two-level hierarchical class scheduling strategy is adopted, the independent class is preferentially processed, the sub-class is then processed and the conflicts are cleaned up, in combination with the dynamic teacher priority scheduling algorithm, the teachers with strict constraints and less class scheduling are preferentially arranged. According to the course weekly class hour requirement, the intelligent continuous class and degradation strategies are adopted to adapt the schemes such as 4 continuous classes and 3 continuous classes, real-time detection of various conflicts is performed, and the backtracking mechanism is triggered to roll back resources. After checking that there is no conflict, the data is persisted to generate an initial scheme, and then the model and strategy are optimized based on the evaluation indexes and application feedback to form a multi-dimensional coupled intelligent class scheduling optimization scheme.

[0009] Through hierarchical processing and intelligent backtracking mechanism, the local optimal solution is effectively avoided, and the class scheduling success rate is improved by more than 40% under complex constraints; the dynamic teacher priority scheduling ensures that the teachers with strict constraints are preferentially arranged, avoids the imbalance of teacher workload, and improves the utilization rate of teaching resources; the multi-dimensional constraint model quantifies multiple constraints, adapts to role restrictions, and solves complex business scenarios such as cross-campus class scheduling.

[0010] The intelligent class continuation strategy reduces course fragmentation, ensures teaching continuity, and improves teaching effectiveness; the tiered scheduling and conflict pre-detection mechanism reduces time complexity and meets the high-efficiency requirements of large-scale scheduling; it supports iterative optimization based on changes in business rules and application feedback, adapting to the scheduling needs of different educational scenarios.

[0011] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0012] Figure 1 The flowchart illustrates an intelligent course scheduling method based on multi-constraint hierarchical optimization and dynamic backtracking provided in an embodiment of this application. Figure 2 The diagram shows a schematic of the structure of an intelligent scheduling device based on multi-constraint hierarchical optimization and dynamic backtracking provided in an embodiment of this application. Detailed Implementation

[0013] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0014] The following is combined with Figure 1 This application describes an intelligent scheduling method based on multi-constraint hierarchical optimization and dynamic backtracking, according to exemplary embodiments thereof. It should be noted that the following application scenarios are shown only to facilitate understanding of the spirit and principles of this application, and the embodiments of this application are not limited in any way. Rather, the embodiments of this application can be applied to any applicable scenario.

[0015] In one implementation, Figure 1 The diagram illustrates a flowchart of an intelligent scheduling method based on multi-constraint hierarchical optimization and dynamic backtracking according to an embodiment of this application.

[0016] S101 retrieves basic data including semester, class, course, teacher, role, and resource group.

[0017] In one implementation, the semester data clearly defines the semester cycle corresponding to the course scheduling, including the semester start date, end date, total number of teaching weeks, holiday arrangements, and special teaching periods (such as mid-term and final review weeks, exam weeks), which serve as the basic boundary conditions for the course scheduling time dimension.

[0018] The class data covers the basic attributes of each class, including class number, grade, major, number of students, campus, class type (independent class or sub-class), and corresponding parent class number (only sub-classes need to include this), providing the class hierarchy and affiliation basis for the implementation of the tiered scheduling strategy.

[0019] Course data contains core information about the course, including course number, course name, course type (required, elective, etc.), weekly class hours required, whether consecutive classes are required, suitable resource group type (classroom, laboratory, computer room, etc.), grade level and major range, etc., which clarifies the core scheduling requirements and resource matching conditions of the course.

[0020] Teacher data records teachers' basic information and teaching-related attributes, including teacher ID, name, department, major, range of courses that can be taught, campus, and teaching time periods (excluding personal leave and teaching and research activities), providing basic data support for teacher selection and priority scheduling.

[0021] The role data clearly defines the scheduling roles corresponding to teachers. The role types include four categories: "cannot teach", "not priority teaching", "must teach", and "priority teaching". Each role has clear scheduling permission restrictions, which serve as the core basis for role constraint modeling and priority weight calculation.

[0022] Resource group data covers information related to various teaching resources, including resource group number, resource group name, resource type (regular classroom, laboratory, computer room, etc.), affiliated campus, capacity limit (number of students that can be accommodated), equipment configuration (such as whether multimedia, special instruments, etc. are provided), and available time periods, providing data support for resource group constraint modeling and course scheduling resource allocation.

[0023] S102 preprocesses the basic data, establishes a multi-dimensional constraint model for roles, campuses, resource groups, and class hours, and models all kinds of constraints as time slot state values, which are then transformed into specific scheduling rules and time period markers.

[0024] In one implementation, the basic data for semesters, classes, courses, teachers, roles, and resource groups are cleaned and standardized to remove redundant information and fill in missing data. Standardization processing of the basic data for semesters, classes, courses, teachers, roles, and resource groups aims to improve data quality to support subsequent constraint modeling. Removing redundant information specifically includes eliminating duplicate teacher information, duplicate course registration records, and other invalid data to avoid data redundancy leading to scheduling logic confusion. Filling in missing data involves supplementing missing key fields in the data, such as adding information like unclear school districts for classes, missing weekly class hour requirements for courses, and unmarked teaching scope for teachers, ensuring the integrity of the core attributes of each type of basic data and providing comprehensive and accurate data support for the formulation of subsequent constraint rules.

[0025] The constraints of teacher roles, campuses, resource groups, and class hours are transformed into quantifiable rules that can be calculated. This clarifies role definitions, the scope of cross-campus scheduling restrictions, the upper limit of shared resource capacity, and the standards for weekly class hours and consecutive class periods. The four types of core business constraints are also transformed into directly quantifiable rules that clearly define the specific criteria for each constraint, as follows: The role constraints clearly define the four roles of teachers: "cannot teach" means completely excluded from course scheduling; "not given priority" means only considered when there are no other suitable teachers; "must teach" means that the teacher is a necessary candidate to teach the designated course; and "given priority" means that the teacher is given priority in course allocation under the same conditions.

[0026] Campus constraints clearly define the scope of restrictions on cross-campus scheduling, specifying whether teachers are allowed to teach across campuses, the number of campuses they can teach at, and the specific names of the campuses, in order to avoid time conflicts and waste of resources caused by cross-campus teaching. Resource group constraints clearly define the capacity limits of shared resources such as laboratories and computer rooms, that is, the number of students or courses that can be taught simultaneously in each type of resource group, to prevent resource overload.

[0027] The time constraints clearly define the weekly time standard for courses, such as 2 or 4 class hours per week, as well as the requirement for consecutive classes, such as whether consecutive classes are mandatory and the minimum number of consecutive classes required, providing a clear basis for subsequent time allocation.

[0028] Through a state mapping mechanism, four types of constraints are uniformly modeled as time slot state values ​​of 1-6, with different state values ​​corresponding to different constraint strengths and scheduling permissions. This mechanism transforms role constraints, campus constraints, resource group constraints, and class hour constraints into time slot state values ​​of 1-6 levels, with different state values ​​corresponding to different constraint strengths and scheduling permissions, achieving unified quantitative processing of multiple types of constraints. From state value 1 to 6, the constraint strength gradually increases, and the scheduling permission gradually decreases: state value 1 represents the lowest constraint strength and the highest scheduling permission, corresponding to scheduling scenarios with no strict restrictions; state value 6 represents the highest constraint strength and the lowest scheduling permission, corresponding to scenarios where scheduling is completely prohibited or only allowed in extremely special circumstances. For example, the role constraint "cannot teach" is mapped to state value 6, and the role constraint "must teach" is mapped to state value 2. This numerical approach makes the constraint strength quantifiable and comparable.

[0029] Based on time slot status values ​​and quantification rules, rules for determining teacher scheduling eligibility, rules for feasible time slots for cross-campus scheduling, rules for resource group usage reservations, and rules for class hour allocation priority are formulated and transformed into specific time slot markers. Based on time slot status values ​​at levels 1-6 and the aforementioned quantification rules, four types of specific scheduling rules are formulated and transformed into directly applicable time slot markers, as follows: The teacher scheduling eligibility determination rules mark the teacher's scheduling eligibility in each time period based on the status value corresponding to the role constraint. For example, a teacher with status value 6 is marked as "no scheduling eligibility" in all time periods, while a teacher with status value 2 is marked as "priority scheduling eligibility" in the appropriate time period corresponding to the course.

[0030] The cross-campus scheduling feasibility time period rule marks the specific time periods during which teachers can teach across campuses based on the status value of the campus constraints. For example, if teachers are only allowed to teach across campuses on Monday and Wednesday afternoons, then the corresponding time period is marked as "cross-campus scheduling feasible", and the other time periods are marked as "infeasible".

[0031] The resource group uses reservation rules to mark the availability status of each resource group at different time periods based on the status value of the resource group constraints. For example, if the maximum capacity of a computer room is 40 people, when the number of students in the corresponding course class meets the requirements, it is marked as "resource available" during the target time period; if the capacity is exceeded, it is marked as "resource occupied".

[0032] The time allocation priority rule marks the allocation priority of courses in different time periods based on the status value of time constraints. For example, courses that need to be taught consecutively are marked as "high priority allocation" in consecutive time periods, while courses that are not taught consecutively are marked as "normal priority allocation". By marking time periods, the scheduling rules can be directly applied to the time allocation process.

[0033] S103, based on a two-level hierarchical scheduling strategy and a dynamic teacher priority scheduling algorithm, combined with a multi-dimensional constraint model, prioritizes the scheduling of independent classes, then processes the scheduling of sub-classes and clears conflicts, prioritizes the scheduling of teachers with strict constraints and fewer classes, and generates the scheduling execution results for the target stage.

[0034] In one implementation, a two-level hierarchical scheduling strategy is adopted based on a multi-dimensional constraint model. First, scheduling resources and time planning are performed for independent classes without a parent-child relationship. Using the multi-dimensional constraint model as the core basis, the first level of the two-level hierarchical scheduling strategy is executed, focusing on independent classes without a parent-child relationship. Scheduling resource allocation needs to consider the capacity limits and equipment configuration requirements in the resource group constraints to match suitable teaching resources for independent classes, such as ordinary classrooms and laboratories, ensuring that resource capacity matches the number of students and that equipment meets course requirements. Time planning, based on the teaching week range, holiday arrangements, and weekly class hour standards in the semester data, allocates reasonable time slots for class courses, avoiding conflicting periods and special teaching periods. It also refers to the class hour allocation priority in the time slot markers to ensure the rationality and feasibility of the scheduling.

[0035] After completing the scheduling of individual classes, scheduling is then carried out for sub-classes with parent-child relationships, establishing a parent-child timetable association mechanism and clearing cross-level course conflicts. Based on the completion of scheduling for all individual classes, the second level of the two-tiered scheduling strategy is initiated, targeting sub-classes with clearly defined parent-child relationships. First, a parent-child timetable association mechanism is established, binding the sub-class course schedules to the parent class timetable to ensure the coordination of time and resources between the sub-class courses and the parent class's core courses. Then, the focus is on clearing cross-level course conflicts by comparing the class time slots, resource usage, and teacher arrangements between parent and sub-classes to eliminate time conflicts, resource usage conflicts, and teacher time conflicts, ensuring the consistency of the scheduling results and avoiding teaching chaos caused by class hierarchical relationships.

[0036] A multi-dimensional constraint model is used to select qualified teachers. Priority weights are calculated based on role constraints, with teachers having higher weights for more constraints. The selection criteria include matching the range of courses a teacher can teach with the current course schedule, ensuring the campus meets campus constraints, and guaranteeing that the time slot status value corresponding to the role constraint is not "unqualified to schedule." This ensures that selected teachers meet the basic teaching requirements. Furthermore, priority weights are calculated based on the teacher's role constraints, with higher weights for teachers with more constraints—for example, a teacher with both cross-campus restrictions and a "priority teaching" role constraint has a higher weight than a teacher with only a single constraint. This highlights the scheduling priority of teachers with strict constraints, preventing the problem of unsuitable time slots in subsequent scheduling.

[0037] The system tracks and dynamically sorts teachers by priority weight and number of scheduled classes, prioritizing teachers with strict constraints and fewer current scheduled classes. The system continuously monitors the number of scheduled classes for each selected teacher, accurately recording their current allocation. Combining the previously calculated priority weight and number of scheduled classes, teachers are dynamically sorted: the core logic is "strict constraints first, fewer scheduled classes first," meaning teachers with higher priority weights and fewer current scheduled classes are ranked higher. During scheduling, classes are allocated to teachers according to the dynamic ranking, prioritizing those ranked higher to ensure teachers with strict constraints receive suitable teaching slots and avoiding an imbalance in resource allocation where some teachers have too many classes while others have too few.

[0038] The target-stage scheduling execution results are generated by comprehensively evaluating the completion rate of tiered scheduling, the effectiveness of resolving conflicts between parent-child classes, and the balance of teacher scheduling. The results are based on three core assessment indicators: 1) Tiered scheduling completion rate: This involves statistically analyzing the completion rates of independent classes and their sub-classes to confirm whether all classes have completed their basic course arrangements; 2) Effectiveness of parent-child class conflict resolution: This involves verifying whether time conflicts, resource conflicts, and teacher conflicts between parent and child classes have been completely eliminated, and whether the residual conflict rate is zero; 3) Balance of teacher scheduling: This involves analyzing the distribution of teachers' scheduled class periods to assess whether there are any serious imbalances in course load. Based on the evaluation results of these three indicators, information on the course arrangements, teacher allocation, and resource usage of all classes is integrated to form a complete target-stage scheduling execution result, providing foundational data for subsequent time allocation and conflict resolution.

[0039] S104, based on the weekly class hour requirements of the course, adopts intelligent course combination and degradation strategies to dynamically adapt to consecutive class schemes including 4 consecutive classes and 3 consecutive classes, completes time allocation and detects various conflicts in real time. When a conflict occurs, it triggers an intelligent backtracking mechanism to roll back resources, try alternative solutions, and generate time allocation and conflict resolution execution results.

[0040] In one implementation, a corresponding intelligent consecutive class combination scheme is matched based on the remaining weekly class hours of the course. When the remaining weekly class hours are greater than or equal to 4, a 4-class consecutive scheme is matched; when they are equal to 3, a 3-class consecutive scheme is matched; when they are equal to 2, a 2-class consecutive scheme is matched; and when they are equal to 1, a 1-class consecutive scheme is matched. Using the remaining weekly class hours as the core basis, the optimal consecutive class combination scheme is dynamically matched to ensure the continuity and rationality of the course arrangement. When the remaining weekly class hours are greater than or equal to 4, a 4-class consecutive scheme is matched to minimize course fragmentation; when the remaining weekly class hours are equal to 3, a 3-class consecutive scheme is matched to balance teaching continuity and time flexibility; when the remaining weekly class hours are equal to 2, a 2-class consecutive scheme is matched to meet basic consecutive class teaching needs; and when the remaining weekly class hours are equal to 1, a 1-class consecutive scheme is matched to suit single-lesson scheduling scenarios. All consecutive class scheme matching strictly adheres to the consecutive class requirement standards within the class hour constraints.

[0041] When the optimal consecutive class schedule is not feasible, a degradation strategy is implemented. If a 4-consecutive-class schedule fails, it is converted to two 2-consecutive-class schedules; if a 3-consecutive-class schedule fails, it is converted to one 2-consecutive-class schedule plus one 1-consecutive-class schedule; if a 2-consecutive-class schedule fails, it is converted to two 1-consecutive-class schedules; if a 1-consecutive-class schedule fails, the time allocation is deemed a failure. When the matched optimal consecutive class schedule is not feasible due to resource conflicts, time conflicts, or other reasons, the degradation strategy is automatically activated to ensure the feasibility of time allocation. If the 4-consecutive-class schedule is not feasible, it is downgraded to two 2-consecutive-class schedules, and the split consecutive class periods must still meet the requirements of teaching continuity; if the 3-consecutive-class schedule is not feasible, it is downgraded to a combination of one 2-consecutive-class schedule plus one 1-consecutive-class schedule, prioritizing the consecutive class periods of core classes; if the 2-consecutive-class schedule is not feasible, it is downgraded to two 1-consecutive-class schedules, flexibly adapting to the remaining available time periods; if the 1-consecutive-class schedule is still not feasible, the time allocation for that course is directly deemed a failure to avoid invalid scheduling attempts.

[0042] During the time allocation process, real-time monitoring is conducted for teacher time conflicts, cross-campus conflicts, resource group capacity conflicts, and role restriction conflicts, recording the triggering scenarios and specific types of each conflict. Throughout the entire time allocation process, these four types of conflict detection are performed in real-time to ensure that there are no violations during the scheduling process. Teacher time conflicts are detected to verify whether the same teacher has already been assigned other courses during the target time slot; cross-campus conflicts are detected to confirm whether the time slot for a teacher teaching across campuses meets the feasible time slot requirements of the campus constraints; resource group capacity conflicts are detected to verify whether the capacity of teaching resources (such as classrooms and laboratories) during the target time slot can meet the class size requirements; and role restriction conflicts are detected to verify whether the time slot status value corresponding to the teacher's role allows scheduling during the target time slot. Simultaneously, the specific triggering scenarios (such as a teacher teaching across campuses during a time slot overlapping with another course) and conflict types are recorded to provide a basis for subsequent retrospective adjustments.

[0043] If a conflict is detected, an intelligent backtracking mechanism is triggered, rolling back the scheduled time slots, releasing the occupied time and resources, and re-matching a suboptimal consecutive class schedule or adjusting the time slots to attempt alternative scheduling solutions. When any of the above conflicts is detected, the intelligent backtracking mechanism is immediately triggered to quickly adjust the scheduling plan. First, the scheduled conflicting time slots are rolled back, releasing the core resources such as teacher time and teaching resources occupied by that time slot to prevent the conflict from spreading; then, based on the conflict type and triggering scenario, a suboptimal consecutive class schedule is re-matched (e.g., if there is a conflict of 4 consecutive classes, a downgraded plan of 2 consecutive classes is directly used), or the time slots are adjusted (e.g., avoiding time slots where teachers cannot cross campuses or time slots where resources are occupied), and alternative scheduling solutions are re-attempted until the conflict is resolved or it is confirmed that there is no feasible solution.

[0044] The execution results of time allocation and conflict resolution are generated by comprehensively evaluating the success rate of the combined class schedule adaptation, the accuracy of conflict detection, and the efficiency of the backtracking mechanism. The results are based on three core evaluation indicators: 1) the success rate of the combined class schedule adaptation, which calculates the proportion of successful adaptation between the optimal and downgraded plans; 2) the accuracy of conflict detection, which assesses the accuracy and timeliness of conflict identification to avoid missed or false detections; and 3) the efficiency of the backtracking mechanism, which records the time spent from conflict detection to plan adjustment and the resulting adjustments. Based on the evaluation results of these three indicators, the final course time allocation plan, conflict resolution records, and resource usage are integrated to form a complete execution result, providing data support for subsequent course scheduling verification.

[0045] S105 verifies the execution results of time allocation and conflict resolution. After confirming that there are no conflicts, the results are persisted, the teacher's class period count and timetable status are updated, and the initial class scheduling plan is generated.

[0046] In one implementation, the execution results of time allocation and conflict resolution are verified, checking one by one whether four types of conflicts—teacher time, cross-campus, resource group capacity, and role restrictions—have been completely eliminated. The core objective of this systematic verification of time allocation and conflict resolution results is to ensure that the scheduling results are free of any violations or conflicts. The verification process checks each category one by one: first, teacher time conflicts are checked to confirm that the same teacher is not assigned multiple courses in the same time slot; then, cross-campus conflicts are checked to verify that the time slots for teachers teaching across campuses meet the feasibility requirements of the campus constraints; next, resource group capacity conflicts are checked to confirm that the number of students occupying teaching resources does not exceed their capacity limit; finally, role restriction conflicts are checked to verify whether the teacher's scheduling in the target time slot complies with the time slot status value permissions corresponding to their role, ensuring that all four types of conflicts are completely eliminated, providing conflict-free basic data for subsequent data processing.

[0047] After confirming there are no conflicts, the complete scheduling data is synchronized to the database for persistent storage, ensuring data security and traceability. The stored scheduling data covers all core information such as class, course, teacher, teaching time slot, resource usage, and constraint matching. The database storage mechanism ensures that data will not be lost due to system failures, operational errors, or other factors. It also supports data traceability and querying, facilitating subsequent scheduling adjustments, statistical analysis, and other operations, thus guaranteeing the security and integrity of the scheduling data.

[0048] The system updates the number of scheduled classes for each teacher in real time, accurately recording each teacher's workload. Based on the final confirmed scheduling results, the system updates the number of scheduled classes for each teacher in real time. The counting scope covers all assigned classes for that teacher within the current scheduling cycle, accurately reflecting each teacher's actual workload. This provides accurate data support for subsequent analysis of teacher workload balancing and scheduling optimization, avoiding resource allocation imbalances caused by inaccurate counting.

[0049] The system synchronously maintains the class and teacher timetable status, clearly defining core information such as course times and teaching resources. It ensures that the timetable information is completely consistent with the final scheduling. Each course's specific teaching time (e.g., Monday morning, periods 1-2), teaching resources (e.g., computer lab 302 in the experimental building), and instructor information are clearly marked. This ensures that class timetables clearly present all class schedules, and teacher timetables accurately display individual teacher teaching plans, providing intuitive and accurate information for teaching implementation and student / faculty inquiries.

[0050] Through data integration and status normalization, an initial scheduling plan that meets business rules, is conflict-free, and has a balanced resource allocation is generated. This initial scheduling plan is generated by integrating and normalizing information such as conflict-free scheduling data, teacher class counts, and timetable status. The integration process strictly adheres to business rules and multi-dimensional constraint models to ensure the plan meets various constraints such as roles, campuses, resource groups, and class hours. It also achieves the goals of balanced teacher workload distribution, conflict-free course scheduling, and efficient utilization of teaching resources, ultimately forming an initial scheduling plan that meets educational and teaching needs and can be directly applied.

[0051] S106, based on evaluation indicators such as scheduling success rate, resource utilization efficiency, and scheduling quality, combined with the verification results of the initial scheduling plan, changes in business rules, and feedback from actual applications, optimizes the quantitative rules and state mapping mechanism of the multi-dimensional constraint model, the weight calculation logic of dynamic teacher priority, and the adaptation and degradation strategy of consecutive class combinations. Through multi-dimensional parameter coupling optimization, an intelligent scheduling optimization plan is generated.

[0052] In one implementation, conflict verification results are filtered to retain compliant scheduling results without any conflicts and remove unqualified scheduling results with conflicts. Compliant scheduling results must meet the condition that there are no anomalies in any of the four conflict judgment criteria. The core objective of this targeted filtering of conflict verification results is to retain fully compliant scheduling results and remove unqualified results with conflicts. The filtering criteria strictly follow the four conflict judgment rules: compliant scheduling results must meet the condition that there are no anomalies in teacher time conflicts, cross-campus conflicts, resource group capacity conflicts, and role restriction conflicts. If a scheduling result has any type of conflict (such as a teacher's cross-campus teaching time not meeting constraints, or a resource group exceeding its capacity limit), it is judged as unqualified and removed, ensuring that the filtered data are all conflict-free scheduling results that comply with basic business rules.

[0053] Based on changes in business rules and feedback data from actual applications, we optimized the quantification rules and state mapping mechanism of the multi-dimensional constraint model, the weight calculation logic of dynamic teacher priority, and the adaptation and degradation strategies for consecutive class combinations. We specifically optimized the core elements of class scheduling based on changes in business rules and feedback data from actual applications. For the multi-dimensional constraint model, we optimized its quantification rules and state mapping mechanism to make the definition of role, campus, resource group, and class hour constraints more aligned with actual business scenarios. For dynamic teacher priority scheduling, we optimized the weight calculation logic to ensure more accurate priority calculation for teachers with more constraints and fewer classes. For the intelligent consecutive class combination strategy, we optimized the adaptation logic and degradation path to make the selection of consecutive class solutions more in line with teaching needs and the execution of degradation strategies more efficient. The overall optimization process was guided by feedback from actual applications, improving the rationality and adaptability of the class scheduling logic.

[0054] Through multi-dimensional parameter coupling optimization, an intelligent scheduling optimization scheme is generated that offers more accurate constraint adaptation, more balanced resource allocation, and superior scheduling quality. This optimization integrates the optimized constraint model, priority scheduling logic, and consecutive-class combination strategy. During coupling, the focus is on coordinating the relationships between parameters to ensure that the optimization of the constraint model does not affect the execution of the scheduling logic, and that adjustments to the consecutive-class combination strategy remain consistent with constraint requirements. Ultimately, this results in an intelligent scheduling optimization scheme with superior overall performance. This scheme boasts three core advantages: more accurate constraint adaptation, better meeting complex business rules; more balanced resource allocation, effectively avoiding imbalanced teacher workload and idle resources; and superior scheduling quality, reducing course fragmentation and improving teaching continuity through optimized consecutive-class combination strategies, comprehensively addressing the core pain points of traditional scheduling algorithms.

[0055] In one implementation, such as Figure 2 As shown, this application also provides an intelligent scheduling device based on multi-constraint hierarchical optimization and dynamic backtracking, comprising: The basic data acquisition module 201 is used to acquire basic data including semester, class, course, teacher, role, and resource group. It covers core information such as semester cycle, class level, course requirements, teacher attributes, role type, and resource configuration, providing data support for the entire course scheduling process. The constraint model construction module 202 is used to preprocess the basic data, clean and organize the data and fill in the missing information, establish a multi-dimensional constraint model of roles, campuses, resource groups and class hours, and model various constraints into 1-6 level time slot status values, which are then transformed into specific scheduling rules and time period markers. The hierarchical scheduling execution module 203 is used to prioritize the scheduling of independent classes based on a two-level hierarchical scheduling strategy, a dynamic teacher priority scheduling algorithm, and a multi-dimensional constraint model. It then processes the scheduling of sub-classes and clears cross-level conflicts. Teachers are dynamically sorted according to the strictness of constraints and the number of classes scheduled, and courses are allocated in a priority manner to generate the scheduling execution results for the target stage. The time allocation and conflict resolution module 204 is used to dynamically adapt to the continuous class schemes such as 4 consecutive classes and 3 consecutive classes according to the weekly class hour requirements of the courses, using intelligent course combination and degradation strategies. It completes the time allocation and detects conflicts in teacher time, cross-campus, resource group capacity, and role restriction in real time. When a conflict occurs, it triggers an intelligent backtracking mechanism to roll back resources, try alternative solutions, and generate corresponding execution results. The scheduling result solidification module 205 is used to fully verify the execution results of time allocation and conflict resolution. After confirming that there are no conflicts, the scheduling data is persistently stored, and the teacher's scheduling period count and class and teacher timetable status are updated in real time to generate a conflict-free and resource-balanced initial scheduling plan. The scheduling optimization iteration module 206 is used to optimize the constraint model quantification rules, teacher priority weight calculation logic, and consecutive class adaptation and degradation strategies based on scheduling success rate, resource utilization efficiency, scheduling quality evaluation indicators, combined with the initial scheduling scheme verification results, business rule changes and actual application feedback. Through multi-dimensional parameter coupling optimization, it generates an intelligent scheduling optimization scheme.

[0056] The computer-readable storage medium provided in the above embodiments of this application and the intelligent scheduling method based on multi-constraint hierarchical optimization and dynamic backtracking provided in the embodiments of this application are based on the same inventive concept and have the same beneficial effects as the methods adopted, run or implemented by the applications stored therein.

[0057] The various embodiments in this application are described in a related manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the embodiments for evaluating the intelligent scheduling method, system, electronic device, and readable storage medium based on multi-constraint hierarchical optimization and dynamic backtracking are basically similar to the embodiments of the intelligent scheduling method based on multi-constraint hierarchical optimization and dynamic backtracking described above, and are therefore described simply. Relevant parts can be referred to in the descriptions of the embodiments of the intelligent scheduling method based on multi-constraint hierarchical optimization and dynamic backtracking described above.

Claims

1. An intelligent course scheduling method based on multi-constraint hierarchical optimization and dynamic backtracking, characterized in that, include: Obtain basic data including semester, class, course, teacher, role, and resource group; The basic data is preprocessed to establish a multi-dimensional constraint model for roles, campuses, resource groups, and class hours. All kinds of constraints are uniformly modeled as time slot state values ​​and transformed into specific scheduling rules and time period markers. Based on a two-level hierarchical scheduling strategy and a dynamic teacher priority scheduling algorithm, combined with a multi-dimensional constraint model, the scheduling of independent classes is processed first, then the scheduling of sub-classes is processed and conflicts are cleared. Teachers with strict constraints and few classes are given priority in scheduling, and the scheduling execution results of the target stage are generated. Based on the weekly class hour requirements, an intelligent course combination and degradation strategy is adopted to dynamically adapt to consecutive class schemes including 4 consecutive classes and 3 consecutive classes, complete the time allocation and detect various conflicts in real time. When a conflict occurs, an intelligent backtracking mechanism is triggered to roll back resources, try alternative solutions, and generate time allocation and conflict resolution execution results. Verify the execution results of time allocation and conflict resolution. After confirming that there are no conflicts, persist the results, update the teacher's class period count and timetable status, and generate the initial class scheduling plan. Based on evaluation indicators such as scheduling success rate, resource utilization efficiency, and scheduling quality, combined with the verification results of the initial scheduling plan, changes in business rules, and feedback from actual applications, the quantitative rules and state mapping mechanism of the multi-dimensional constraint model, the weight calculation logic of dynamic teacher priority, and the adaptation and degradation strategies of consecutive class combinations are optimized. Through multi-dimensional parameter coupling optimization, an intelligent scheduling optimization plan is generated.

2. The method as described in claim 1, characterized in that, The basic data is preprocessed to establish a multi-dimensional constraint model for roles, campuses, resource groups, and class hours. All constraints are then uniformly modeled as time slot state values, transforming them into specific scheduling rules and time period markers, including: Clean and organize the basic data of semester, class, course, teacher, role, and resource group, remove redundant information and fill in missing data; The constraints of teacher role, campus, resource group, and class hour are transformed into calculable quantitative rules, clarifying role definitions, cross-campus scheduling restrictions, shared resource capacity limits, and weekly class hour standards and consecutive class requirements. Through the state mapping mechanism, the four types of constraints are uniformly modeled as time slot state values ​​of 1-6, with different state values ​​corresponding to different constraint strengths and scheduling permissions; Based on the time slot status value and quantification rules, rules for determining teacher scheduling eligibility, rules for feasible time slots for cross-campus scheduling, rules for resource group usage reservations, and rules for class hour allocation priority are formulated and transformed into specific time slot markers.

3. The method as described in claim 1, characterized in that, Based on a two-tiered scheduling strategy and a dynamic teacher priority scheduling algorithm, combined with a multi-dimensional constraint model, the scheduling of independent classes is processed first, followed by the scheduling of sub-classes and conflict resolution. Teachers with strict constraints and low teaching loads are given priority in scheduling, generating the target stage scheduling execution results, including: Based on a multi-dimensional constraint model, a two-level hierarchical scheduling strategy is adopted. First, scheduling resources and time planning are carried out for independent classes without parent-level relationships. After completing the scheduling of individual classes, scheduling will be carried out for sub-classes with parent-child relationships, a parent-child timetable association mechanism will be established, and cross-level course conflicts will be cleared. Teachers qualified to teach are selected by combining a multi-dimensional constraint model, and priority weights are calculated based on role restrictions, with higher weights for teachers with more constraints. The number of classes already scheduled by teachers is statistically analyzed, and dynamically sorted according to priority weight and the number of classes scheduled, giving priority to teachers with strict constraints and a small current number of classes scheduled. The overall performance of tiered scheduling, the effectiveness of resolving conflicts between father-son classes, and the balance of teacher scheduling are used to generate the execution results of the target stage of scheduling.

4. The method as described in claim 1, characterized in that, Based on the weekly class hour requirements, an intelligent course combination and degradation strategy is adopted to dynamically adapt to consecutive class schedules, including 4-class and 3-class schedules. Time allocation is completed, and various conflicts are detected in real time. When a conflict occurs, an intelligent backtracking mechanism is triggered to roll back resources, try alternative solutions, and generate time allocation and conflict resolution execution results, including: Based on the remaining weekly class hours of the course, the corresponding intelligent consecutive class combination scheme is matched. When the remaining weekly class hours are greater than or equal to 4, 4 consecutive classes are matched; when they are equal to 3, 3 consecutive classes are matched; when they are equal to 2, 2 consecutive classes are matched; and when they are equal to 1, 1 consecutive class is matched. When the optimal consecutive class plan is not feasible, a degradation strategy is implemented. If a 4-class consecutive class fails, it is converted to 2 two-class consecutive classes. If a 3-class consecutive class fails, it is converted to 1 two-class consecutive class plus 1 one-class consecutive class. If a 2-class consecutive class fails, it is converted to 2 one-class consecutive classes. If a one-class consecutive class fails, the time allocation is judged to have failed. During the time allocation process, real-time detection is performed on teacher time conflicts, cross-campus conflicts, resource group capacity conflicts, and role restriction conflicts, and the triggering scenarios and specific types of various conflicts are recorded. If a conflict is detected, the intelligent backtracking mechanism is triggered to roll back the scheduled time slots, release the occupied time and resources, rematch the suboptimal consecutive class schedule or adjust the time slots, and try to replace the scheduling scheme. The system integrates the success rate of the integrated solution adaptation, the accuracy of conflict detection, and the efficiency of the backtracking mechanism to generate execution results for time allocation and conflict resolution.

5. The method as described in claim 1, characterized in that, The execution results of time allocation and conflict resolution are verified. After confirming there are no conflicts, the results are persisted, the teacher's class period count and timetable status are updated, and an initial scheduling plan is generated, including: The execution results of time allocation and conflict resolution are verified, and each of the four types of conflicts—teacher time, cross-campus, resource group capacity, and role restrictions—is checked to see if they have been completely eliminated. After confirming there are no conflicts, the complete scheduling data will be synchronized to the database for persistent storage to ensure data security and traceability. The system updates the number of classes scheduled by teachers in real time, accurately recording the course load of each teacher. The system synchronously maintains the class and teacher timetable status, clearly defining core information such as class times and teaching resources; Through data integration and status normalization, an initial course scheduling plan is generated that meets business rules, is conflict-free, and has a balanced allocation of resources.

6. The method as described in claim 5, characterized in that, Based on evaluation indicators such as scheduling success rate, resource utilization efficiency, and scheduling quality, and combined with the verification results of the initial scheduling plan, changes in business rules, and feedback from actual applications, this study optimizes the quantitative rules and state mapping mechanism of the multi-dimensional constraint model, the weight calculation logic of dynamic teacher priority, and the adaptation and degradation strategies for consecutive class combinations. Through multi-dimensional parameter coupling optimization, an intelligent scheduling optimization plan is generated, including: The conflict verification results are screened and processed, retaining compliant scheduling results without any conflicts and removing unqualified scheduling results with conflicts. Compliant scheduling results must meet the condition that there are no abnormalities in all four conflict judgment criteria. Based on the needs of business rule changes and feedback data from actual applications, we optimized the quantitative rules and state mapping mechanism of the multi-dimensional constraint model, the weight calculation logic of dynamic teacher priority, and the adaptation and degradation strategies of consecutive class combinations. By optimizing multi-dimensional parameters, an intelligent scheduling optimization solution is generated that achieves more accurate constraint adaptation, more balanced resource allocation, and better scheduling quality.

Images