After-class service business management method and system based on SaaS mode

Abstract

The invention discloses an after-class service business management method and system based on a SaaS mode, and relates to the technical field of after-class services, and the method comprises the steps: distributing an independent database fragment for each education end based on a SaaS multi-tenant isolation strategy; dynamically generating an electronic protocol template according to the course type and the real-time site capacity, filling parent information by using a placeholder replacement rule, and generating a to-be-signed document; asynchronously pushing the to-be-signed document according to the priority of the tenant through the message middleware and the elastic cloud server cluster; receiving the electronic signature completed by the parent end, and fusing the electronic signature with the to-be-signed document to generate an electronic protocol; and when the agreement signing completion rate exceeds a preset threshold value, generating a scheduling scheme by using a resource scheduling model in combination with the teacher qualification matching degree, the classroom capacity limitation and the course time continuity constraint.

Description

Technical Field

[0001] This application relates to the technical field of after-school service, and more particularly, to a method and system for managing after-school service business based on the SaaS model. Background Art

[0002] With the deepening of education reform, after-school service has gradually become an important part of school education. Traditional after-school service management mainly relies on manual operations, such as using paper forms to record student information, parent contact information, course arrangements, etc. This management method is not only inefficient but also prone to information errors and omissions.

[0003] In recent years, although some schools have started to introduce information management systems, these systems are often limited to local deployment, with problems such as high costs, poor scalability, and difficulties in data sharing. For example, the after-school service management system of a school may not be able to connect with the supervision platform for data, resulting in a serious information island phenomenon, leading to significant technical shortcomings in aspects such as resource allocation optimization and being unable to meet the high-efficiency and intelligent needs of modern education services.

[0004] In response to the above problems, no effective solutions have been proposed yet. Summary of the Invention

[0005] Embodiments of this application provide a method and system for managing after-school service business based on the SaaS model to solve the above technical problems.

[0006] This application provides a method for managing after-school service business based on the SaaS model, including:

[0007] Based on the SaaS multi-tenant isolation strategy, allocate independent database shards for each educational end;

[0008] According to the course type and real-time venue capacity, dynamically generate an electronic protocol template, and use placeholder replacement rules to fill in the parent information to generate a document to be signed;

[0009] Asynchronously push the document to be signed according to the tenant priority through the message middleware and the elastic cloud server cluster;

[0010] Receive the electronic signature completed by the parent end, integrate the electronic signature with the document to be signed, and generate an electronic protocol;

[0011] When the protocol signing completion rate exceeds the preset threshold, combine the teacher qualification matching degree, classroom capacity limit, and course time continuity constraint, and use the resource scheduling model to generate a scheduling plan.

[0012] Further, after generating the scheduling plan using the resource scheduling model, the method further includes:

[0013] Based on the classroom video stream data, extract the student interaction features, and combine with the evaluation model to determine the trend of the course effect; among them, the parameters of the evaluation model are determined according to different course types.

[0014] Based on the microservice architecture gateway, feedback the trend of the course effect to the resource scheduling model to dynamically adjust the scheduling plan.

[0015] Further, the allocation method of the SaaS multi-tenant isolation strategy includes: assigning a tenant ID to each educational end, and storing the course data of each educational end in different cloud database nodes in a sharded manner based on the consistent hashing algorithm; among them, the course data includes course information, teacher configuration, and parent data.

[0016] Dynamically generate an electronic protocol template according to the course type and the real-time venue capacity, including: setting a protocol push priority queue according to the tenant service level agreement, and realizing traffic hierarchical control through the message middleware; based on the tenant's historical course data, using a causal inference model to analyze the causal relationship between the course type, the real-time venue capacity, and the protocol terms, and dynamically generate an electronic protocol template corresponding to the tenant.

[0017] Further, the resource scheduling model includes a mixed integer programming model; when the protocol signing completion rate exceeds the preset threshold, combine the teacher qualification matching degree, classroom capacity limit, and course time continuity constraint, and use the resource scheduling model to generate a scheduling plan, including:

[0018] Take the teacher qualification, classroom capacity, and course time continuity as hard constraints, and verify the relevance between the teacher and the course through the knowledge graph.

[0019] Introduce a cross-tenant resource sharing conflict cost rule to quantify the competition relationship between different educational ends for shared teachers or venues.

[0020] Real-time adjust the weights of teacher idle rate, venue conflict rate, and cross-tenant conflict cost to adapt to the priority needs of different tenants.

[0021] Model teachers, courses, and venues as hypergraph nodes, and decompose them into multiple subgraph problems for parallel processing.

[0022] For the subgraph combination, adopt the quantum annealing heuristic algorithm to quickly approximate the global optimal solution.

[0023] When it is detected that the resource utilization rate of a certain tenant is lower than the utilization rate threshold, recycle the idle resources through Kubernetes and re-allocate them to the high-demand tenants; among them, the high-demand tenants are those whose course scheduling request volume exceeds the resource supply capacity within the preset time period.

[0024] Further, based on the classroom video stream data, extracting students' interaction features and combining with an evaluation model to determine the trend of course effectiveness, including:

[0025] Identifying the key interaction behavior chains in the video through a spatio-temporal causal convolutional network;

[0026] Eliminating the differences in classroom scenarios of different tenants according to the adversarial domain adaptation strategy;

[0027] Performing speech emotion analysis and facial expression recognition on the scenarios in the key interaction behavior chains to determine the comprehensive index of students' participation;

[0028] Using a meta-reinforcement learning framework to extract evaluation strategy features from high-scoring tenant courses and migrate them to the initial model of new tenants; where the comprehensive index of students' participation in high-scoring tenant courses is greater than the index threshold;

[0029] When outputting the trend of course effectiveness, annotating the policy reusability label; where the policy reusability label supports cross-tenant policy sharing.

[0030] Further, based on the microservice architecture gateway, feedback the trend of course effectiveness to the resource scheduling model to dynamically adjust the scheduling plan, including:

[0031] Aggregating the evaluation results of multiple tenants through a federated reinforcement learning framework and updating the global resource scheduling policy; where the evaluation results of multiple tenants include course completion rate, trend of course effectiveness, policy reusability label, and comprehensive index of students' participation;

[0032] Embedding an interpretable rule engine in the mixed integer programming model to dynamically adjust the constraint weights;

[0033] When the course effectiveness of a certain tenant continues to be lower than the threshold, trigger the resource recovery mechanism, and dynamically allocate idle teachers or venues to high-demand tenants through Kubernetes;

[0034] Based on the meta-reinforcement learning framework, migrate the scheduling strategy features of high-scoring tenants to the new tenant model.

[0035] Further, the SaaS-based after-school service business management method is implemented based on a microservice architecture, and the microservice architecture includes:

[0036] Pay-per-use service, which is used to dynamically charge according to the computing resources used by tenants, and realize usage statistics and bill generation through the API gateway;

[0037] Security authentication microservice, which is used to transmit data through the OAuth2.0 protocol and SSL / TLS encryption;

[0038] A deployment service for accelerating protocol file distribution through the CDN acceleration protocol, combined with edge computing nodes to reduce access latency.

[0039] Furthermore, based on the Kubernetes container orchestration technology, it monitors the resource utilization rate of each tenant in real time and dynamically scales the cloud server instances up and down;

[0040] Among them, elastic expansion is achieved in the following way: when the concurrent request volume of the tenant increases to a preset number, new cloud server instances are started, and traffic is allocated through a load balancer; according to the complexity of the resource scheduling model, the specifications of the cloud server instances are dynamically adjusted.

[0041] Furthermore, after the electronic protocol is generated, the method further includes:

[0042] Using an Ethereum smart contract to encrypt and store the signing records of the document to be signed, and implementing decentralized archiving of the electronic protocol through IPFS to ensure cross-tenant data isolation and immutability.

[0043] This application provides a SaaS-based after-school service business management system, including: a data storage module for allocating independent database shards for each educational end based on the SaaS multi-tenant isolation strategy; a protocol filling module for dynamically generating an electronic protocol template according to the course type and real-time venue capacity, and filling in the parent information using placeholder replacement rules to generate a document to be signed; a protocol generation module that asynchronously pushes the document to be signed to the elastic cloud server cluster according to the tenant priority through a message middleware; receiving the electronic signature completed by the parent end, fusing the electronic signature with the document to be signed to generate an electronic protocol; a scheduling plan generation module for generating a scheduling plan using a resource scheduling model when the protocol signing completion rate exceeds a preset threshold, in combination with teacher qualification matching degree, classroom capacity limit, and course time continuity constraints.

[0044] Based on the embodiments provided in this application, by adopting the SaaS multi-tenant isolation strategy and allocating independent database shards for each educational end, efficient data isolation and management are achieved. This not only ensures the security and independence of the data of each educational end, avoiding the risks of data chaos and information leakage, but also reduces the data management cost and improves the scalability and flexibility of data storage compared with the traditional on-premises deployment system. Each educational end can flexibly expand the data storage capacity according to its own needs without affecting other tenants. According to the course type and real-time venue capacity, an electronic protocol template is dynamically generated, and the parent information is filled using the placeholder replacement rule to automatically generate the document to be signed. This process avoids information errors and omissions caused by manual filling, improving the accuracy and efficiency of protocol generation. At the same time, through the message middleware and the elastic cloud server cluster, the document to be signed is asynchronously pushed according to the tenant priority, ensuring the timeliness and reliability of the document push, optimizing the parent signing experience, and further improving the efficiency and quality of after-school service management. When the protocol signing completion rate exceeds the preset threshold, combined with the teacher qualification matching degree, classroom capacity limit, and course time continuity constraint, a scheduling plan is generated using the resource scheduling model. This resource scheduling method based on comprehensive consideration of multiple factors can achieve intelligent optimization and allocation of after-school service resources, improve the utilization rate of resources such as teachers and classrooms, ensure the rationality and coherence of course arrangements, effectively solve the technical shortcomings of the traditional management method in resource allocation optimization, and meet the needs of modern education services for high efficiency and intelligence.

[0045] This application is built based on the SaaS model and naturally has good system integration and data sharing capabilities. Different from the traditional on-premises deployment system, this application can achieve data docking between the after-school service management system and other related systems such as the supervision platform, break the information silos, promote the circulation and sharing of educational data, provide more comprehensive and accurate data support for educational decision-making, and promote the coordinated development of after-school services and even the entire education system.

[0046] In summary, this application shows significant optimization effects in multiple key aspects such as data management, protocol processing, resource scheduling, and system integration, effectively overcoming many problems existing in the background technology, providing a more efficient, intelligent, and reliable solution for after-school service management, strongly promoting the innovation and development of the after-school service management model, and having high application value and promotion significance. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, form a part of this application, and the schematic embodiments and descriptions thereof are used to explain this application without unduly limiting this application. In the drawings:

[0048] Figure 1Flowchart of an alternative SaaS - based after - school service business management method according to an embodiment of the present application;

[0049] Figure 2 Flowchart of another alternative SaaS - based after - school service business management method according to an embodiment of the present application;

[0050] Figure 3 Structure diagram of an alternative SaaS - based after - school service business management system according to an embodiment of the present application.

[0051] The realization of the object of the present invention, functional features and advantages will be further described with reference to the embodiments and the accompanying drawings. Detailed implementation manners

[0052] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.

[0053] Optionally, as Figure 1 shown, the present application provides a SaaS - based after - school service business management method, including:

[0054] S101, allocate independent database shards for each educational end based on the SaaS multi - tenant isolation strategy;

[0055] S102, dynamically generate an electronic agreement template according to the course type and the real - time venue capacity, and use the placeholder replacement rule to fill in the parent information to generate a document to be signed;

[0056] In this embodiment, the electronic agreement template can be dynamically generated through a cloud natural language processing service according to the course type and the real - time venue capacity;

[0057] S103, asynchronously push the document to be signed to the tenants according to the tenant priority through a message middleware and an elastic cloud server cluster;

[0058] Among them, the message middleware is Kafka; the elastic cloud server cluster is AWS EC2. Asynchronously pushing the document to be signed according to the tenant priority supports load balancing and automatic disaster tolerance in high - concurrency scenarios.

[0059] S104, receive the electronic signature completed by the parent end, fuse the electronic signature and the document to be signed to generate an electronic agreement;

[0060] Among them, the electronic signature can be a fingerprint or a face recognized through biometric identification.

[0061] S105. When the protocol signing completion rate exceeds the preset threshold, a scheduling plan is generated using a resource scheduling model in combination with teacher qualification matching degree, classroom capacity limit, and course time continuity constraint.

[0062] In the embodiments of the present application, the resource scheduling model may include, but is not limited to, a mixed integer programming model, a dynamic programming model based on deep reinforcement learning, and a multi-tenant cooperation model based on game theory. Among them, the dynamic programming model based on deep reinforcement learning adjusts the scheduling strategy in real time through a reinforcement learning agent to adapt to sudden changes in course requirements. The multi-tenant cooperation model based on game theory generates an equilibrium allocation plan by simulating resource competition and cooperation among tenants.

[0063] Based on the embodiments provided in the present application, by adopting the SaaS multi-tenant isolation strategy, independent database shards are allocated to each educational end, realizing efficient isolation and management of data. This not only ensures the security and independence of data for each educational end, avoiding the risks of data chaos and information leakage, but also reduces the data management cost compared with traditional on-premises deployment systems, improving the scalability and flexibility of data storage. Each educational end can flexibly expand the data storage capacity according to its own needs without affecting other tenants. According to the course type and real-time venue capacity, an electronic protocol template is dynamically generated, and the parent information is filled using placeholder replacement rules to automatically generate a document to be signed. This process avoids information errors and omissions caused by manual filling, improving the accuracy and efficiency of protocol generation. At the same time, through the message middleware and the elastic cloud server cluster, the document to be signed is asynchronously pushed according to the tenant priority, ensuring the timeliness and reliability of document pushing, optimizing the parent signing experience, and further improving the efficiency and quality of after-school service management. When the protocol signing completion rate exceeds the preset threshold, a scheduling plan is generated using a resource scheduling model in combination with teacher qualification matching degree, classroom capacity limit, and course time continuity constraint. This resource scheduling method based on comprehensive consideration of multiple factors can achieve intelligent optimization configuration of after-school service resources, improve the utilization rate of resources such as teachers and classrooms, ensure the rationality and coherence of course arrangements, effectively solve the technical shortcomings of traditional management methods in resource allocation optimization, and meet the requirements of modern education services for high efficiency and intelligence.

[0064] The present application is built based on the SaaS model and naturally has good system integration and data sharing capabilities. Different from traditional on-premises deployment systems, the present application can realize data docking between the after-school service management system and other related systems such as the supervision platform, break information silos, promote the circulation and sharing of educational data, provide more comprehensive and accurate data support for educational decision-making, and promote the coordinated development of after-school services and even the entire education system.

[0065] In summary, this application demonstrates significant optimization effects in multiple key aspects such as data management, protocol processing, resource scheduling, and system integration. It effectively overcomes many problems existing in the background technology, provides a more efficient, intelligent, and reliable solution for after-school service management, strongly promotes the innovation and development of the after-school service management model, and has high application value and promotional significance.

[0066] Further, after generating the scheduling plan using the resource scheduling model, the method further includes:

[0067] Based on the classroom video stream data, extract the student interaction features through the AI service (such as AWS SageMaker) deployed in the cloud, and determine the course effect trend in combination with the evaluation model; wherein, the parameters of the evaluation model (such as the interaction frequency weight) are determined according to different course types;

[0068] Based on the microservice architecture gateway (such as the interaction frequency weight), feedback the course effect trend to the resource scheduling model to dynamically adjust the scheduling plan. Form a closed-loop link of "protocol signing → resource allocation → effect evaluation → strategy iteration".

[0069] Further, the allocation method of the SaaS multi-tenant isolation strategy includes: assigning a tenant ID to each educational end, and storing the course data of each educational end in different cloud database nodes (such as AWS Aurora sharded cluster) based on the consistent hashing algorithm to ensure cross-tenant data physical isolation; wherein, the course data includes course information, teacher configuration, and parent data;

[0070] In the embodiment of this application, the multi-tenant data sharding weight allocation model is as follows:

[0071]

[0072] where W i is the data sharding weight of the i-th tenant, which determines the allocation priority of the cloud node for storing its course data; H(T i ) is the hash value generated by the consistent hashing algorithm for the unique ID of tenant T i ; S {Ti} is the service level agreement (SLA) priority score (from level 1 to level 5) of tenant T i , for example, VIP institutions are set to 5; C priority is the dynamic priority correction coefficient, which is automatically adjusted according to the real-time request traffic (range 0.8 to 1.2); when high concurrency is detected (such as the request volume > 1000 times / second), it is automatically increased to 1.2, otherwise it is decreased to 0.8; j: the index for traversing all tenants in the system, j ∈ [1, N]; N is the total number of registered educational institutions (tenants).

[0073] Based on the embodiments provided in this application, the shard weights are calculated by weighted calculation of the hash value and the SLA priority, ensuring that the data of high-priority tenants is stored nearby and accessed quickly, meeting the personalized needs of different tenants, and improving the refinement degree of data management. The shard hotspot problem is effectively avoided, making the data storage more balanced and reasonable, improving the overall performance and stability of the system, and ensuring the smooth operation of the after-school service business. Based on the dynamic priority correction coefficient, it is automatically adjusted according to the real-time request traffic, enhancing the system's ability to handle complex scenarios such as high concurrency, and ensuring stable and efficient operation under different business loads.

[0074] According to the course type and the real-time venue capacity, an electronic protocol template is dynamically generated, including:

[0075] According to the tenant service level agreement, a protocol push priority queue is set, and traffic hierarchical control is realized through the message middleware;

[0076] Based on the tenant's historical course data, a causal inference model is used to analyze the causal relationship between the course type, the real-time venue capacity, and the protocol terms, and an electronic protocol template corresponding to the tenant is dynamically generated. For example, art courses need to attach equipment usage terms;

[0077] Noise can be injected during the template generation process through differential privacy technology (DP-Adam) to prevent cross-tenant data leakage.

[0078] Furthermore, taking the resource scheduling model including the mixed integer programming model as an example; as Figure 2 shown, when the protocol signing completion rate exceeds the preset threshold, combined with the teacher qualification matching degree, classroom capacity limit, and course time continuity constraint, a scheduling plan is generated using the resource scheduling model, including:

[0079] S201, taking the teacher qualification, classroom capacity, and course time continuity as hard constraints, and verifying the relevance between the teacher and the course through the knowledge graph;

[0080] S202, introducing a cross-tenant resource sharing conflict cost rule to quantify the competition relationship of different educational parties for shared teachers or venues;

[0081] S203, adjusting the weights of the teacher idle rate, venue conflict rate, and cross-tenant conflict cost in real time to adapt to the priority needs of different tenants;

[0082] S204, modeling teachers, courses, and venues as hypergraph nodes and decomposing them into multiple subgraph problems for parallel processing;

[0083] S205, for the subgraph combination, using the quantum annealing heuristic algorithm to quickly approximate the global optimal solution;

[0084] S206, when it is detected that the resource utilization rate of a certain tenant is lower than the utilization rate threshold, Kubernetes is used to recycle the idle resources and reallocate them to high-demand tenants; among them, high-demand tenants are those whose course scheduling request volume exceeds the resource supply capacity (such as the teacher or venue occupancy rate ≥ 90%) within a preset time period.

[0085] In the embodiment of the present application, the objective function of the mixed-integer programming model is:

[0086]

[0087] Among them, Z is the total cost of resource scheduling; I t is the proportion of idle time slices of teacher t; C t is the deviation between the qualification of teacher t and the course matching degree, which is calculated through the knowledge graph; E r is the conflict event count of venue r;

[0088] Conflict(r) is the cross-tenant resource sharing conflict coefficient (0 to 1) of venue r; T is the total number of registered teachers; R is the total number of available venues; λ t , μ t , ν r are dynamic weights, and the initial values are set to λ t = 0.5, μ t = 0.3, ν r = 0.2; λ t , μ t , ν r can be dynamically optimized through the multi-armed bandit algorithm. Specifically, each weight can be regarded as an "arm", and the reward value is calculated according to the historical scheduling effect (such as teacher utilization rate); the arm with the largest upper confidence bound (UCB) is selected as the current optimal weight.

[0089] Based on the embodiment provided by the present application, on the premise of meeting the hard constraints, by minimizing teacher waste and venue conflicts, quantifying the cross-tenant resource competition cost, realizing the optimal allocation of resources, improving the utilization rate of resources such as teachers and classrooms, and reducing the operation cost. Using the multi-armed bandit algorithm to dynamically optimize the weights enables the model to automatically adjust the strategy according to the historical scheduling effect and better adapt to the changes in business requirements. On the premise of meeting the hard constraints, flexibly responding to the priority needs of different tenants, supporting cross-tenant resource sharing, providing a more flexible operation mode for the after-school service business, and enhancing the adaptability and competitiveness of the system.

[0090] Furthermore, based on the classroom video stream data, extract the student interaction features, and combine with the evaluation model to determine the course effect trend, including:

[0091] Identify the key interaction behavior chain in the video through a spatio-temporal causal convolutional network; the key interaction behavior chain can be, for example, teacher's question → student's answer → group discussion;

[0092] According to the adversarial domain adaptation strategy, eliminate the differences in classroom scenarios of different tenants; among them, the classroom scenarios of different tenants are such as class size, equipment type;

[0093] Perform speech emotion analysis (which can be based on BERT+BiLSTM) and facial expression recognition (which can be based on ResNet) on the scenarios in the key interaction behavior chain to determine the comprehensive student engagement index;

[0094] Use the meta-reinforcement learning framework to extract the evaluation strategy features from the courses of high-scoring tenants and transfer them to the initial model of new tenants; among them, the comprehensive student engagement index of high-scoring tenants is greater than the index threshold;

[0095] When outputting the trend of course effects, label the policy reusability label (such as "applicable to art courses with less than 10 people"); among them, the policy reusability label supports cross-tenant policy sharing.

[0096] In the embodiments of this application, the comprehensive student engagement index is determined based on the following formula:

[0097]

[0098] Among them, P s is the comprehensive student engagement index (from 0 to 100) of student s; V s is the intensity of video interaction features extracted by the spatio-temporal causal convolutional network; A s is the emotional positivity score output by the speech emotion analysis; F s is the focus score output by the facial expression recognition; ω v and ω a and ω f are modal weights, dynamically allocated through meta-reinforcement learning, and the initial weights can be set as ω v =0.4, ω a =0.3, ω f =0.3, which can be adjusted according to the course type. For example, in art courses, ω a (speech emotion) can have its weight increased to 0.5; Interaction(s) is the number of effective interactions of student s; Duration(s) is the total duration of the course; ADA(s) is the scene difference correction factor (from 0.5 to 1.5) output by the adversarial domain adaptation technology.

[0099] Based on the embodiments provided in this application, by integrating multi-modal data with the time dimension and comprehensively considering factors such as students' video interactions, voice emotions, and facial expressions, the learning status of students can be more accurately evaluated, providing a strong basis for the evaluation of course effects. Through meta-reinforcement learning, the modal weights are dynamically allocated, and the weights of various factors are adjusted according to different course types, making the evaluation results more in line with the actual teaching needs and providing accurate guidance for the optimization of subsequent teaching strategies. By introducing the scenario difference correction factor output by the adversarial domain adaptation technology, the differences in classroom scenarios of different tenants are eliminated, ensuring the fairness and comparability of the evaluation results and promoting the cross-tenant analysis of course effects and experience sharing.

[0100] Furthermore, based on the microservices architecture gateway, the course effect trend is fed back to the resource scheduling model to dynamically adjust the scheduling plan, including:

[0101] Through the federated reinforcement learning framework, the evaluation results of multiple tenants are aggregated to update the global resource scheduling strategy; among them, the evaluation results of multiple tenants include the course completion rate, the course effect trend, the policy reusability label, and the comprehensive index of student participation;

[0102] An interpretable rule engine (such as SHAP value analysis) is embedded in the mixed integer programming model to dynamically adjust the constraint weights (such as enhancing the priority of VIP institution venues);

[0103] When the course effect of a certain tenant continuously falls below the threshold, a resource recovery mechanism is triggered, and idle teaching staff or venues are dynamically allocated to high-demand tenants through Kubernetes;

[0104] Based on the meta-reinforcement learning framework, the scheduling strategy features of high-scoring tenants are migrated to the new tenant model.

[0105] Furthermore, the after-school service business management method based on the SaaS model is implemented based on the microservices architecture, and the microservices architecture includes:

[0106] Pay-as-you-go service, which is used to dynamically charge according to the computing resources used by the tenant, and realize usage statistics and bill generation through the API gateway; among them, the computing resources include CPU, storage, and bandwidth;

[0107] Security authentication microservice, which is used to transmit data through the OAuth2.0 protocol and SSL / TLS encryption;

[0108] Deployment service, which is used to distribute protocol files through the CDN acceleration protocol and combine edge computing nodes () to reduce access latency.

[0109] Furthermore, based on the Kubernetes container orchestration technology, the resource utilization rate of each tenant is monitored in real time, and the cloud server instances are dynamically scaled to ensure service stability in the multi-tenant scenario;

[0110] Among them, elastic expansion is achieved in the following way: when the concurrent request volume of tenants increases to a preset quantity, new cloud server instances are started, and traffic is allocated through a load balancer; according to the complexity of the resource scheduling model, the specifications of cloud server instances (such as the number of CPU cores and the size of memory) are dynamically adjusted.

[0111] Further, after the electronic protocol is generated, the method further includes:

[0112] Using Ethereum smart contracts to encrypt and store the signing records of the documents to be signed, and implementing decentralized archiving of the electronic protocol through IPFS to ensure cross-tenant data isolation and immutability.

[0113] Optionally, as Figure 3 shown, the present application provides a SaaS-based after-school service business management system, including:

[0114] A data storage module 301, configured to allocate independent database shards for each educational end based on the SaaS multi-tenant isolation strategy;

[0115] A protocol filling module 302, configured to dynamically generate an electronic protocol template according to the course type and the real-time venue capacity, and use the placeholder replacement rule to fill in the parent information to generate a document to be signed;

[0116] A protocol generation module 303, which asynchronously pushes the document to be signed to the elastic cloud server cluster according to the tenant priority through a message middleware; receives the electronic signature completed by the parent end, and fuses the electronic signature and the document to be signed to generate an electronic protocol;

[0117] A scheduling plan generation module 304, configured to generate a scheduling plan by using a resource scheduling model in combination with the teacher qualification matching degree, the classroom capacity limit, and the course time continuity constraint when the protocol signing completion rate exceeds a preset threshold.

[0118] It should be noted that in the present application, the embodiments implemented on the side of the SaaS-based after-school service business management system can be mutually referred to the embodiments implemented on the side of the SaaS-based after-school service business management method, and the present application will not elaborate one by one.

[0119] The above are only the preferred embodiments of the present invention, and do not limit the patent scope of the present invention accordingly. Any equivalent structural or equivalent process transformation made by using the specification and drawings of the present invention, or directly or indirectly applied in other related technical fields, shall be included in the patent protection scope of the present invention by the same token.

Claims

1. A method for managing after-school service business based on the SaaS model, characterized in that, Including: Based on the SaaS multi-tenant isolation strategy, allocate independent database shards for each educational end; Dynamically generate an electronic protocol template according to the course type and real-time venue capacity, and use placeholder replacement rules to fill in parent information to generate a document to be signed; Asynchronously push the document to be signed according to the tenant priority through the message middleware and the elastic cloud server cluster; Receive the electronic signature completed by the parent end, fuse the electronic signature with the document to be signed to generate an electronic protocol; When the protocol signing completion rate exceeds the preset threshold, combine the teacher qualification matching degree, classroom capacity limit and course time continuity constraint, and use the resource scheduling model to generate a scheduling plan.

2. The after-school service business management method based on the SaaS model according to claim 1, wherein After generating the scheduling plan using the resource scheduling model, the method further includes: Extract student interaction features based on classroom video stream data, and combine with an evaluation model to determine the course effect trend; wherein, the parameters of the evaluation model are determined according to different course types; Based on the microservice architecture gateway, feedback the course effect trend to the resource scheduling model to dynamically adjust the scheduling plan.

3. The SaaS mode-based after-school service business management method according to claim 1, wherein The allocation method of the SaaS multi-tenant isolation strategy includes: allocating a tenant ID for each educational end, and storing the course data shards of each educational end in different cloud database nodes based on the consistent hashing algorithm; wherein, the course data includes course information, teacher configuration and parent data; The dynamically generating an electronic protocol template according to the course type and real-time venue capacity includes: setting a protocol push priority queue according to the tenant service level agreement, and implementing traffic grading control through the message middleware; based on the tenant's historical course data, using a causal inference model to analyze the causal relationship between the course type, real-time venue capacity and protocol terms, and dynamically generate an electronic protocol template corresponding to the tenant.

4. The after-school service business management method based on the SaaS model according to claim 3, characterized in that, The resource scheduling model includes a mixed integer programming model; when the protocol signing completion rate exceeds the preset threshold, combining the teacher qualification matching degree, classroom capacity limit and course time continuity constraint, and using the resource scheduling model to generate a scheduling plan, including: Taking teacher qualification, classroom capacity and course time continuity as hard constraints, and verifying the relevance between teachers and courses through a knowledge graph; Introduce a cross-tenant resource sharing conflict cost rule to quantify the competition relationship between different educational ends for shared teachers or venues; Real-time adjust the weights of teacher idle rate, venue conflict rate and cross-tenant conflict cost to adapt to the priority requirements of different tenants; Model teachers, courses and venues as hypergraph nodes, and decompose them into multiple subgraph problems for parallel processing; For the subgraph combination, adopt the quantum annealing heuristic algorithm to quickly approximate the global optimal solution; When it is detected that the resource utilization rate of a certain tenant is lower than the utilization rate threshold, recycle the idle resources through Kubernetes and re-allocate them to high-demand tenants; wherein, a high-demand tenant is a tenant whose course scheduling request volume exceeds the resource supply capacity within a preset time period.

5. The after-school service business management method based on the SaaS model according to claim 4, wherein The extracting student interaction features based on classroom video stream data and combining with an evaluation model to determine the course effect trend includes: Identifying key interaction behavior chains in videos through spatio-temporal causal convolutional networks; Eliminating differences in classroom scenarios of different tenants according to the adversarial domain adaptation strategy; Performing speech emotion analysis and facial expression recognition on the scenarios in the key interaction behavior chains to determine the comprehensive student engagement index; Using a meta-reinforcement learning framework to extract evaluation strategy features from high-scoring tenant courses and transfer them to the initial model of new tenants; wherein, the comprehensive student engagement index of high-scoring tenants is greater than the index threshold; When outputting the trend of course effects, annotating the policy reusability label; wherein, the policy reusability label supports cross-tenant policy sharing.

6. The method for managing after-school service business based on the SaaS model according to claim 5, wherein Based on the microservice architecture gateway, feedback the course effect trend to the resource scheduling model to dynamically adjust the scheduling plan, including: Aggregating multi-tenant evaluation results through a federated reinforcement learning framework to update the global resource scheduling policy; wherein, the multi-tenant evaluation results include course completion rate, course effect trend, policy reusability label, and comprehensive student engagement index; Embedding an interpretable rule engine in the mixed integer programming model to dynamically adjust the constraint weights; When the course effect of a certain tenant continuously falls below the threshold, trigger a resource recovery mechanism to dynamically allocate idle teaching staff or venues to high-demand tenants through Kubernetes; Based on the meta-reinforcement learning framework, transfer the scheduling policy features of high-scoring tenants to the new tenant model.

7. The after-school service business management method based on the SaaS model according to claim 1, characterized in that The SaaS-based after-school service business management method is implemented based on a microservice architecture, and the microservice architecture includes: Pay-as-you-go service, used to dynamically charge according to the computing resources used by tenants, and realize usage statistics and bill generation through the API gateway; Security authentication microservice, used to transmit data through the OAuth2.0 protocol and SSL / TLS encryption; Deployment service, used to distribute protocol files through the CDN acceleration protocol, and combine edge computing nodes to reduce access latency.

8. The SaaS-based after-school service business management method according to claim 1, characterized in that Based on the Kubernetes container orchestration technology, real-time monitor the resource utilization rate of each tenant and dynamically scale the cloud server instances; Among them, elastic expansion is achieved in the following way: when the tenant concurrent request volume increases to a preset number, start a new cloud server instance and allocate traffic through a load balancer; dynamically adjust the cloud server instance specifications according to the complexity of the resource scheduling model.

9. The after-school service business management method based on the SaaS model according to claim 1, characterized in that After generating the electronic protocol, the method further includes: Using an Ethereum smart contract to encrypt and store the signing records of the document to be signed, and implementing decentralized archiving of the electronic protocol through IPFS to ensure cross-tenant data isolation and immutability.

10. A business management system for after-school service based on the SaaS model, characterized in that, Including: Data storage module, used to allocate independent database shards for each educational end based on the SaaS multi-tenant isolation strategy; Protocol filling module, used to dynamically generate an electronic protocol template according to the course type and real-time venue capacity, and fill in the parent information using placeholder replacement rules to generate a document to be signed; Protocol generation module, asynchronously push the document to be signed according to tenant priorities through a message middleware and an elastic cloud server cluster; Receive the electronic signature completed by the parent end, integrate the electronic signature with the document to be signed, and generate an electronic agreement; A scheduling plan generation module, which is used to generate a scheduling plan using a resource scheduling model by combining the teacher qualification matching degree, classroom capacity limit, and course time continuity constraint when the agreement signing completion rate exceeds a preset threshold.

Images