Hotel guest room management method and system based on cloud computing

Abstract

The invention provides a hotel guest room management method and system based on cloud computing, and is applied to the field of cloud data processing. By monitoring the access resources cached by the hotel terminal in real time, the current access situation can be rapidly obtained, the load state of the system can be reflected in time, whether the current access resources reach the preset load upper limit or not can be judged in real time, it is ensured that corresponding measures are taken in time under the high-load situation, system overload is avoided, and the service life of the system is prolonged. Meanwhile, access resources are predicted through a pre-trained machine learning model, possible peak periods and load conditions can be identified in advance, scheduling can be performed in advance, resource contents of hotel terminals are dynamically scheduled according to prediction results, reasonable allocation and use of resources in peak periods are ensured, and resource waste or shortage is avoided.

Description

Technical Field

[0001] The present invention relates to the field of cloud data processing, and in particular to a hotel room management method and system based on cloud computing. Background Art

[0002] At present, traditional hotel room management faces many challenges, including low resource utilization efficiency, insufficient customer experience and information security risks. To solve these challenges, cloud computing came into being. Through cloud computing infrastructure, hotel terminals can realize virtualization and elastic expansion of computing resources. However, if the hotel terminal needs to face high load, it is difficult to maintain the corresponding stability and high response speed, and it is impossible to ensure that the hotel terminal can stably manage the guest rooms during peak periods. Summary of the invention

[0003] The present invention aims to solve the problem that when a hotel terminal needs to face a high load, it is difficult to maintain the corresponding stability and high response speed, and it is impossible to ensure that the hotel terminal can stably manage the guest rooms during peak hours, and to provide a hotel room management method and system based on cloud computing.

[0004] The present invention adopts the following technical means to solve the technical problem: The present invention provides a hotel room management method based on cloud computing, comprising: Based on the load upper limit preset by the hotel terminal, access resources cached by the hotel terminal are obtained in real time; Determining whether the access resource has reached the load upper limit; If not, a pre-trained prediction model is applied to predict the access resources, resource contents of the hotel terminal are scheduled according to the prediction results, the demand types corresponding to the access resources are identified, and the resource contents are dynamically allocated to the hybrid cloud architecture preset by the hotel terminal according to the demand types, and a cloud processing platform corresponding to the access resources is generated, wherein the hybrid cloud architecture specifically includes a public cloud and a private cloud; Determining whether the cloud processing platform detects a preset fault; If so, activate the distributed application architecture preset by the hotel terminal, allow the access resources to switch and expand between the hybrid cloud architectures, execute synchronization strategies on the access resources in other cloud platforms based on the multi-cloud deployment preset by the hotel terminal, obtain the disaster recovery strategy data of the cloud processing platform, and switch the access resources from the other cloud platforms back to the cloud processing platform according to the disaster recovery strategy data, wherein the synchronization strategy specifically includes full synchronization and incremental synchronization, and the disaster recovery strategy data specifically includes recovery time objectives and recovery point objectives.

[0005] Furthermore, before the step of applying the pre-trained prediction model to predict the access resources and scheduling the resource content of the hotel terminal according to the prediction result, it also includes: Identify the feature target variable corresponding to the access resource, calculate the score data of the feature target variable, divide the feature target variable into a preferred feature subset and a secondary feature subset based on the score data, perform recursive feature elimination on the access resource, and construct the feature data required by the prediction model, wherein the feature target variable specifically includes numerical features, classification features, time features, and text features; Determining whether the characteristic data have the same dimension range; If not, then collect the data distribution characteristics of the feature data, discretize the feature data according to the data distribution characteristics, and calculate the interactive characteristics of the feature data, wherein the data distribution characteristics specifically include long-tail distribution and multimodal distribution, and the discretization specifically includes equal-width discretization and equal-frequency discretization.

[0006] Furthermore, the step of identifying the type of requirement corresponding to the access resource includes: Based on the concurrent requests pre-received by the hotel terminal, extract key features of the concurrent requests to generate priority queue information of the access resources, wherein the concurrent requests specifically include reservation requests, query requests and payment requests; Determine whether the priority queue information has different types of priorities; If so, respond to the corresponding resources to be processed from the priority queue information, dynamically adjust the execution decision of the hotel terminal on the resources to be processed according to the priority queue information, and divide the allocated resources required for the resources to be processed from the preset server according to the execution decision.

[0007] Furthermore, the step of executing a synchronization strategy on the access resource in other cloud platforms based on the preset multi-cloud deployment of the hotel terminal further includes: Identify the storage support service preset for the access resource, build a corresponding storage bucket in the other cloud platform based on the storage support service, and assign access rights corresponding to the storage bucket to the access resource; Determining whether the cloud processing platform can read the access permission; If not, configure a data synchronization task for the access resource from the other cloud platform, verify the synchronization status of the access resource in the cloud processing platform and the other cloud platform, and detect conflicting data of the access resource based on the synchronization status, wherein the data synchronization task specifically includes source and target storage locations, synchronization frequency and synchronization direction.

[0008] Furthermore, the step of determining whether the access resource has reached the load upper limit further includes: Based on the monitoring node preconfigured by the hotel terminal, obtaining the index data of the access resource to the hotel terminal, wherein the index data specifically includes response time, request rate and error rate; Determining whether the indicator data exceeds a static threshold preset by the hotel terminal; If so, a preset distributed tracing tool is applied to track the request link for accessing the resource, collect abnormal information of the request link during the tracing process, and identify corresponding abnormal nodes and abnormal bottlenecks according to the abnormal information.

[0009] Furthermore, the step of determining whether the cloud processing platform detects a preset fault further includes: Obtaining processing progress information of the access resource by the cloud processing platform; Determine whether the processing progress information can be processed when the preset fault occurs; If not, a preset fault analysis tool is applied in the cloud processing platform to perform root cause analysis, locate the fault data block corresponding to the cloud processing platform based on the analysis result, and divide the fault link and fault node from the fault data block.

[0010] Furthermore, before the step of acquiring the access resources cached by the hotel terminal in real time based on the load upper limit preset by the hotel terminal, the method further includes: Based on the detection script preset by the hotel terminal, the operation process of the hotel user is simulated and at least two access requests are initiated from various preset locations to the hotel terminal according to the preset frequency, and the detection data corresponding to the access requests are collected, wherein the operation process specifically includes user login, room query, room reservation and payment, and the detection data specifically includes response time and success rate; Determining whether the detection data exceeds the alarm rules preconfigured by the hotel terminal; If so, the location information corresponding to the access request is marked from the hotel terminal, and the access request is interactively bridged with the manual channel preset by the hotel terminal according to the location information to collect the access resources corresponding to the access request.

[0011] The present invention also provides a hotel room management system based on cloud computing, comprising: An acquisition module, used to acquire access resources cached by the hotel terminal in real time based on a preset load upper limit of the hotel terminal; A judgment module, used to judge whether the access resource has reached the load upper limit; An execution module is used for, if not, applying a pre-trained prediction model to predict the access resource, scheduling the resource content of the hotel terminal according to the prediction result, identifying the demand type corresponding to the access resource, dynamically allocating the resource content to the hybrid cloud architecture preset by the hotel terminal according to the demand type, and generating a cloud processing platform corresponding to the access resource, wherein the hybrid cloud architecture specifically includes a public cloud and a private cloud; A second judgment module is used to judge whether the cloud processing platform detects a preset fault; The second execution module is used to activate the distributed application architecture preset by the hotel terminal, allow the access resources to switch and expand between the hybrid cloud architecture, execute synchronization strategies on the access resources in other cloud platforms based on the multi-cloud deployment preset by the hotel terminal, obtain the disaster recovery strategy data of the cloud processing platform, and switch the access resources from the other cloud platforms back to the cloud processing platform according to the disaster recovery strategy data, wherein the synchronization strategy specifically includes full synchronization and incremental synchronization, and the disaster recovery strategy data specifically includes recovery time objectives and recovery point objectives.

[0012] Furthermore, it also includes: A construction module is used to identify the characteristic target variable corresponding to the access resource, calculate the score data of the characteristic target variable, divide the characteristic target variable into a preferred feature subset and a secondary feature subset based on the score data, perform recursive feature elimination on the access resource, and construct the characteristic data required by the prediction model, wherein the characteristic target variable specifically includes numerical features, classification features, time features and text features; A third judgment module is used to judge whether the characteristic data have the same dimension range; The third execution module is used to collect data distribution characteristics of the feature data, discretize the feature data according to the data distribution characteristics, and calculate the interactive characteristics of the feature data, wherein the data distribution characteristics specifically include long-tail distribution and multimodal distribution, and the discretization specifically includes equal-width discretization and equal-frequency discretization.

[0013] Furthermore, the execution module includes: A generating unit, configured to extract key features of the concurrent requests pre-received by the hotel terminal and generate priority queue information of the access resource, wherein the concurrent requests specifically include reservation requests, query requests and payment requests; A determination unit, configured to determine whether the priority queue information has different types of priorities; An execution unit is used to respond to the corresponding pending resources from the priority queue information, dynamically adjust the execution decision of the hotel terminal on the pending resources according to the priority queue information, and divide the allocated resources required for the pending resources from the preset server according to the execution decision.

[0014] The present invention provides a hotel room management method and system based on cloud computing, which has the following beneficial effects: By real-time monitoring of the access resources cached by the hotel terminal, the present invention can quickly obtain the current access situation, timely reflect the load status of the system, and judge in real time whether the current access resources have reached the preset load upper limit, so as to ensure that corresponding measures are taken in time under high load conditions to avoid system overload. At the same time, the access resources are predicted by the pre-trained machine learning model, and possible peak periods and load conditions can be identified in advance, and advance scheduling can be performed. The resource content of the hotel terminal is dynamically scheduled according to the prediction results to ensure that resources can be reasonably allocated and used during peak periods to avoid resource waste or shortage. Moreover, according to the demand type of access resources, the system can manage resources more finely to ensure that different types of demands are optimized. The hybrid cloud architecture combining public cloud and private cloud is adopted to realize flexible scheduling and expansion of resources, which can not only utilize the elastic expansion advantage of the public cloud, but also ensure the security and reliability of the private cloud. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 A schematic diagram of a process of an embodiment of a hotel room management method based on cloud computing of the present invention; Figure 2 The structure diagram of an embodiment of a hotel room management system based on cloud computing of the present invention. DETAILED DESCRIPTION

[0016] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. The implementation of the objectives, functional features and advantages of the present invention will be further described in conjunction with the embodiments and with reference to the accompanying drawings.

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] Reference Figure 1 , is a hotel room management method based on cloud computing in an embodiment of the present invention, comprising: S1: Based on the preset load upper limit of the hotel terminal, access resources cached by the hotel terminal are obtained in real time; S2: Determine whether the access resource has reached the load upper limit; S3: If not, then use the pre-trained prediction model to predict the access resources, schedule the resource content of the hotel terminal according to the prediction results, identify the demand type corresponding to the access resources, dynamically allocate the resource content to the hybrid cloud architecture preset by the hotel terminal according to the demand type, and generate a cloud processing platform corresponding to the access resources, wherein the hybrid cloud architecture specifically includes a public cloud and a private cloud; S4: Determine whether the cloud processing platform detects a preset fault; S5: If so, activate the distributed application architecture preset by the hotel terminal, allow the access resources to switch and expand between the hybrid cloud architectures, execute synchronization strategies for the access resources in other cloud platforms based on the multi-cloud deployment preset by the hotel terminal, obtain the disaster recovery strategy data of the cloud processing platform, and switch the access resources from the other cloud platforms back to the cloud processing platform according to the disaster recovery strategy data, wherein the synchronization strategy specifically includes full synchronization and incremental synchronization, and the disaster recovery strategy data specifically includes recovery time objectives and recovery point objectives.

[0019] In this embodiment, the system obtains the access resources cached by the hotel terminal in real time based on the server load upper limit pre-set by the hotel terminal, and then the system determines whether these access resources have reached the pre-set server load upper limit to execute the corresponding steps; for example, when the system determines that the access resources cached by the hotel terminal have reached the pre-set server load upper limit, the system will believe that the current server can no longer handle more requests, and further increasing the load will cause performance degradation or system crash. The system will configure a load balancer to evenly distribute requests to different servers, dynamically adjust the load balancing strategy, adjust the distribution strategy according to the real-time load situation, increase the cache layer, use the distributed cache system to improve the cache hit rate, adjust the cache expiration strategy, extend the cache time, reduce the load on the server, and use the microservice architecture to independently deploy and run different service modules, configure the service grid, and realize dynamic routing and isolation between services; for example, when the system determines that the access resources cached by the hotel terminal have not reached the pre-set server load upper limit, the system will believe that the current server can still handle the access request, and the system will apply the pre-trained prediction model Predict access resources, schedule the resource content of the hotel terminal according to different prediction results, identify the demand types corresponding to the access resources, and dynamically allocate the resource content to the pre-set hybrid cloud architecture of the hotel terminal according to these demand types. The hybrid cloud architecture specifically includes public cloud and private cloud, and generates a cloud processing platform corresponding to the access resources; through the prediction model, the system can identify future access needs in advance, optimize resource allocation, avoid resource waste, dynamically schedule resources, ensure that resources can be allocated on demand, reduce resource idleness, and improve overall resource utilization. At the same time, through advance prediction and scheduling, the system can reserve sufficient resources before demand increases to avoid delays caused by temporary expansion. According to the demand type, the resource content is dynamically allocated to the appropriate cloud platform to balance the load and ensure that each node can operate efficiently. The hybrid cloud architecture allows the system to dynamically expand or reduce resources according to actual needs and flexibly respond to different load conditions. The combination of public cloud and private cloud allows the system to flexibly adjust the resource allocation strategy and select the optimal resource deployment method according to different business needs; then the system determines whether the cloud processing platform detects the pre-set fault content to execute the corresponding steps;For example, when the system determines that the cloud processing platform has not detected the pre-set fault content, the system will consider that the access resources in the hotel terminal processing do not need to be transferred from the cloud processing platform to other cloud platforms. The system will use monitoring tools to monitor various system indicators in real time, such as CPU usage, memory usage and network traffic, configure alarm mechanisms, set warning thresholds, and issue alarms in advance when indicators are close to critical values, take timely measures, and regularly update the prediction model to ensure its accuracy and real-time performance. According to the prediction results, resources are prepared and scheduled in advance to ensure that they can meet future access needs, and identify the types of access resource requirements, such as short-term high-frequency access and long-term stable access, and dynamically adjust resource allocation strategies to ensure that resource allocation meets actual needs; for example, when the system determines that the cloud processing platform detects the pre-set fault content, the system will consider that the access resources in the hotel terminal processing need to be transferred from the cloud processing platform to other cloud platforms. The system will activate the pre-set distributed application architecture of the hotel terminal, allowing access resources to switch and expand between hybrid cloud architectures, and access data in other cloud platforms based on the pre-set multi-cloud deployment of the hotel terminal. Execute synchronization strategies, which include full synchronization and incremental synchronization, obtain disaster recovery strategy data of the cloud processing platform, which includes recovery time objective and recovery point objective, and switch access resources from other cloud platforms back to the cloud processing platform according to these disaster recovery strategy data; the system deploys synchronization strategies (full synchronization and incremental synchronization) for access resources on other cloud platforms to ensure that even if the main cloud platform fails, the system can still quickly switch to the backup cloud platform to ensure business continuity and availability. According to the disaster recovery strategy data (recovery time objective and recovery point objective), a reasonable switching plan is formulated to minimize business interruption time and data loss risks. At the same time, the pre-set distributed application architecture is activated to allow access resources to dynamically switch and expand between hybrid cloud architectures. Based on the multi-cloud deployment strategy, the access resources are deployed on multiple cloud platforms to disperse risks and improve the elasticity and reliability of the system. In addition, according to business needs and disaster recovery strategies, the allocation and use of resources are dynamically adjusted to optimize resource utilization efficiency. Through multi-cloud deployment and resource optimization, the cost risk caused by dependence on a single cloud platform is reduced, and more flexible and cost-effective resource management is achieved. ;

[0020] It should be noted that a specific example of applying a pre-trained prediction model to predict the access resources and scheduling the resource content of the hotel terminal according to the prediction result is as follows: Suppose a hotel chain wants to improve the responsiveness and stability of its online reservation system. Especially during holidays or promotions, reservation requests increase dramatically. By applying a pre-trained prediction model, the system can identify peak periods in advance and dynamically schedule resources. The specific steps and implementation process include: Data collection and preprocessing,By collecting historical data, we collected the reservation request data for the past two years,from the hotel reservation system database, including the number of requests per hour; Assume the data example is as follows Date Time Request Quantity xxxx-06-01 00:00 50 xxxx-07-01 01:00 45 ... xxxx-08-01 23:00 75 Then, data cleaning is performed. Missing values ​​are processed first by deleting or filling in records with missing booking request numbers, outlier processing is performed, abnormally high or low request numbers are identified and processed, and feature extraction is performed, including extracting date, time, day of the week, holidays and promotions, user visit time, length of stay and booked room type; LSTM (Long Short-Term Memory Network) model is selected, and the LSTM model is trained using historical data. Input features include time features and user behavior features. The request quantity data is standardized for model training, and then the data set is divided into training set (70%), validation set (15%) and test set (15%). The validation set is used to evaluate the model performance and ensure the accuracy and stability of the model; the current booking request data and related features (such as time, whether Whether there is a holiday, whether there is a promotion, etc.) is input into the trained LSTM model to generate a prediction result and predict the number of booking requests in the next hour. For example, the model predicts that the number of booking requests in the next hour will reach 300 times, which is 100 times more than usual. According to the prediction result, the system needs to prepare more computing resources in advance to cope with the upcoming peak period. By increasing the number of virtual machine instances and optimizing the configuration of the load balancer, the requests are evenly distributed to increase the cache capacity and reduce the database query pressure. According to the monitoring data and alarm information, the system can adjust the resource configuration and load balancing strategy in real time. For example, when the CPU usage rate is monitored to exceed 80% continuously, the number of virtual machine instances or Pods is further increased to ensure the stable operation of the system. In summary, the hotel system can effectively use the pre-trained prediction model to identify the peak visit period in advance and dynamically schedule resources according to the prediction results, thereby improving the response speed and stability of the system. In actual operation, the system can quickly respond to users' reservation requests during peak periods, avoiding system crashes or response delays due to insufficient resources, and significantly improving user experience and business continuity.

[0021] It should be added that, based on the multi-cloud deployment preset by the hotel terminal, a specific example of executing the synchronization strategy for the access resource in other cloud platforms is as follows: Assume that a hotel chain group wants to improve the disaster recovery capability of its online reservation system through multi-cloud deployment and data synchronization strategy to ensure that when the primary cloud platform fails, the system can quickly switch to the backup cloud platform to maintain business continuity; First, we design the architecture of multi-cloud deployment. The hybrid cloud architecture includes the primary cloud platform: AWS, the backup cloud platform: Azure, and the private cloud: the local data center. The multi-cloud deployment strategy includes: Running the main reservation system on AWS; Deploy a backup reservation system on Azure and configure data synchronization policies; Private cloud is used to store historical data and sensitive information; Then the data synchronization strategy is implemented, including Full synchronization: During the initial synchronization, all booking data is copied from AWS to Azure. Use the AWS DataSync tool to configure a full synchronization task. The specific steps are as follows: create a DataSync task, select the source location (AWS S3) and the target location (Azure Blob Storage), configure the synchronization options, and select the full synchronization mode. Incremental synchronization: Perform incremental synchronization every hour to copy data that has changed in the last hour from AWS to Azure. Use the AWS DataSync tool to configure incremental synchronization tasks. The specific steps are: create a DataSync task, select the source location (AWS RDS) and the target location (Azure SQL Database), configure synchronization options, select incremental synchronization mode, and set the synchronization frequency to once an hour; Use AWS DataSync to synchronize data, configure Azure Data Factory on Azure to receive and process synchronized data, configure synchronization tasks on the AWS DataSync console, specify source and target locations, set the execution frequency and synchronization scope of synchronization tasks, use AWS CloudWatch and Azure Monitor to monitor the status of synchronization tasks, and send notification emails or SMS messages to the hotel terminal when synchronization tasks fail or are delayed. By configuring the automatic retry mechanism, when synchronization tasks fail, the system will automatically retry, and when synchronization tasks continue to fail, the hotel terminal’s human workers will intervene to handle the problem; The specific implementation cases are as follows: 1. Configure the full synchronization task On the AWS DataSync console, create a new task, select AWS S3 as the source location and Azure Blob Storage as the target location, configure the full synchronization option to ensure that all data can be completely copied to Azure, start the full synchronization task, monitor the synchronization progress, and ensure that the task is successfully completed; 2. Configure incremental synchronization tasks On the AWS DataSync console, create a new task, select AWS RDS as the source location and Azure SQL Database as the target location, configure the incremental synchronization option, set the synchronization frequency to once an hour, start the incremental synchronization task, monitor the synchronization status, and ensure that the task is executed as planned; 3. Real-time monitoring and management Use AWS CloudWatch to monitor the status of DataSync tasks and configure alarm rules. Use AzureMonitor to monitor the status of Azure Data Factory and configure alarm rules. Configure the automatic retry mechanism. When the synchronization task fails, the system will automatically retry three times. By configuring the emergency plan, when the synchronization task continues to fail, notify the operation and maintenance team for manual intervention. In summary, the hotel reservation system can achieve efficient data synchronization and multi-cloud deployment. When the main cloud platform fails, it can quickly switch to the backup cloud platform to maintain business continuity. At the same time, through real-time monitoring and management, the system can promptly detect and handle abnormal situations in synchronization tasks to ensure data consistency and integrity, thereby significantly improving the system's reliability and disaster recovery capabilities.

[0022] In this embodiment, the pre-trained prediction model is used to predict the access resources, and before step S3 of scheduling the resource content of the hotel terminal according to the prediction result, the method further includes: S301: Identify the feature target variable corresponding to the access resource, calculate the score data of the feature target variable, divide the feature target variable into a preferred feature subset and a secondary feature subset based on the score data, perform recursive feature elimination on the access resource, and construct the feature data required by the prediction model, wherein the feature target variable specifically includes numerical features, classification features, time features, and text features; S302: Determine whether the characteristic data have the same dimension range; S303: If not, then collect data distribution characteristics of the feature data, discretize the feature data according to the data distribution characteristics, and calculate interactive characteristics of the feature data, wherein the data distribution characteristics specifically include long-tail distribution and multimodal distribution, and the discretization specifically includes equal-width discretization and equal-frequency discretization.

[0023] In this embodiment, the system identifies the characteristic target variables corresponding to the access resources, which specifically include numerical features, categorical features, time features, and text features, calculates the scoring data of these characteristic target variables, divides the preferred feature subsets and the secondary feature subsets from the characteristic target variables based on different scoring data, and recursively eliminates the features of the access resources to construct the characteristic data required by the prediction model. Then the system determines whether these characteristic data have the same dimensional range to execute the corresponding steps; for example, when the system determines that the characteristic data required by the prediction model have the same dimensional range, the system will consider that all characteristic data have been standardized and can be directly used for model training. The system will use these standardized characteristic data for model training and prediction, deploy the trained prediction model to the production environment, ensure that the model can process new access resource data of the hotel terminal in real time or in batches, monitor the real-time prediction performance of the model, ensure the stability and accuracy of the model in the production environment, configure the monitoring and alarm mechanism, and when the model prediction performance decreases or anomalies occur, adjust and process them in time, and update the model regularly, retrain the model with the latest data, maintain the prediction accuracy and adaptability of the model, and In scenarios where it is needed, the model can learn new data in real time using online learning technology and gradually improve prediction performance. For example, when the system determines that the feature data required by the prediction model do not have the same dimensional range, the system will consider that the feature data has not been standardized and cannot be directly used for model training. The system will collect the data distribution characteristics of these feature data, which specifically include long-tail distribution and multimodal distribution, discretize the feature data according to different data distribution characteristics, including equal-width discretization and equal-frequency discretization, and calculate the interactive characteristics of these feature data. Through standardization, the system can convert feature data of different scales into the same dimensional range, so that each feature has equal importance in the model training process, avoiding some features from having too much impact on the model due to their large numerical range. At the same time, when the feature data has a unified dimensional range, the gradient descent algorithm in the model training process is more stable and efficient, which helps to speed up the model convergence speed and improve training efficiency. It can also identify data distribution characteristics (such as long-tail distribution and multimodal distribution) and perform targeted discretization, which can effectively handle extreme values ​​and noise in the data and improve the robustness and generalization ability of the model.

[0024] It should be noted that, a specific example of the feature data required for recursively eliminating the access resources and building the prediction model is as follows: Suppose a hotel booking dataset contains the following features: The age of the customer; Type of customer; Booking time; Room type; Length of booking; Whether you are a member; The goal is to predict hotel bookings; First, an initial model (e.g., a linear regression model) is trained using all features through recursive feature elimination, as follows: Booking volume = β0 + β1*age + β2*customer type + β3*booking time + β4*room type + β5*booking duration + β6*membership status + ε; Then, by evaluating the feature importance, the importance of each feature is evaluated according to the absolute value of the model coefficient. Assume that the feature importance score is as follows: Age: 0.5; Customer type: 0.8; Booking time: 0.3; Room type: 1.2; Booking duration: 0.7; Member status: 0.4; Then we eliminate the least important features. By removing the feature with the lowest feature importance score (booking time), the update formula is as follows: Booking volume = β0 + β1*age + β2*customer type + β4*room type + β5*booking duration + β6*membership status + ε; Repeat the above steps, retrain the model using the remaining features, and repeat the feature elimination process until the predetermined number of features is reached or the feature selection is completed; Assume that the second round of feature importance scoring includes the following features: Age: 0.4; Customer type: 0.7; Room type: 1.1; Booking duration: 0.6; Member status: 0.5; Then remove the feature with the lowest feature importance score (age), and the update formula is as follows: Booking volume = β0 + β2 * customer type + β4 * room type + β5 * booking duration + β6 * member status + ε; Continue to repeat until only the three most important features are left, and the final feature importance score is: Customer type: 0.7; Room type: 1.1; Booking duration: 0.6; In summary, through recursive feature elimination, the system can gradually remove features that contribute less to model prediction and retain the features that are most helpful to model prediction, thereby improving the model's prediction performance and generalization ability. This process can not only simplify the model and increase the training speed, but also help the hotel terminal better understand which features in the data are the most important. In the above example of hotel booking prediction, through the recursive feature elimination method, the three most important features of "customer type", "room type" and "booking duration" are finally retained, and a more streamlined and effective prediction model is constructed. It should be added that a specific example of collecting data distribution characteristics of the feature data, discretizing the feature data according to the data distribution characteristics, and calculating the interactive characteristics of the feature data is as follows: Suppose a hotel booking dataset contains the following features: The age of the customer; Type of customer; Booking time; Room type; Length of booking; Whether you are a member; The goal is to predict hotel bookings; By collecting data distribution characteristics to understand the distribution characteristics of each feature, assuming that the customer age presents a long-tail distribution, then younger or older customers are in the minority, while customers in the middle age group are in the majority; assuming that the booking time presents a multimodal distribution, then there are multiple peaks and troughs in the booking volume in different time periods; assuming that the booking duration presents a long-tail distribution, then most customers book for a shorter time, while a few customers book for a longer time; according to the distribution characteristics of the feature, select the appropriate discretization method; Assuming that the customer age is divided into three groups, each group contains the same number of customers, for example, 18-30 years old, 31-45 years old, 46 years old and above, then the equal-width discretization will divide the time of the day into 24 intervals by hour, each interval represents one hour; while the equal-frequency discretization will divide the booking duration into three groups, each group contains the same number of booking records, for example, 1-3 days, 4-7 days, 8 days and above; by calculating the interactive features, combining the existing features, and generating new interactive features, the relationship between the features can be captured; Suppose that "customer type" and "room type" are combined to generate a new feature, which indicates the preference of different types of customers for different room types, and "booking time" and "booking duration" are combined to generate a new feature, which indicates the average booking duration of customers in different time periods; Assume the original data is as follows:

[0025] The data after discretization and interaction feature calculation are as follows:

[0026] In summary, by discretizing feature data and calculating interactive features, the system can effectively handle complex relationships in the data and improve the prediction performance of the model. In the example of hotel booking prediction, by analyzing the distribution characteristics of feature data, selecting appropriate discretization methods, and generating new interactive features, customer behavior patterns and booking needs can be better captured, thereby improving the accuracy and stability of the prediction model.

[0027] In this embodiment, the step S3 of identifying the demand type corresponding to the access resource includes: S31: Based on the concurrent requests pre-received by the hotel terminal, extract key features of the concurrent requests to generate priority queue information of the access resources, wherein the concurrent requests specifically include reservation requests, query requests and payment requests; S32: Determine whether different types of priorities exist in the priority queue information; S33: If so, respond to the corresponding resources to be processed from the priority queue information, dynamically adjust the execution decision of the hotel terminal on the resources to be processed according to the priority queue information, and divide the allocated resources required for the resources to be processed from the preset server according to the execution decision.

[0028] In this embodiment, the system is based on concurrent requests pre-received by the hotel terminal, and the concurrent requests specifically include reservation requests, query requests and payment requests. The system extracts key features of these concurrent requests and generates priority queue information for accessing resources. The system then determines whether these priority queue information have different types of priorities to execute corresponding steps; for example, when the system determines that the priority queue information for accessing resources does not have different types of priorities, the system will consider that all concurrent requests are equally important in the current situation, and there are no requests that need to be prioritized. The system will add all concurrent requests to the processing queue in the order they are received, and process each request in turn in the order of the queue without prioritizing, while ensuring that all requests are processed fairly and that some requests are not delayed due to different priorities; for example, when the system determines that the priority queue information for accessing resources has different types of priorities, the system will consider that concurrent requests have priority processing, and the system will respond to the corresponding resources to be processed from the priority queue information. According to the priority queue information, the hotel terminal's execution decision on the resources to be processed is dynamically adjusted, and the allocated resources required for the resources to be processed are divided from the pre-set servers according to different execution decisions; the system can respond quickly to key business needs by giving priority to high-priority requests (such as payment requests), improve user experience and satisfaction, ensure that key operations can be given priority, and reduce the waiting time of key operations, thereby improving the overall response speed of the system. At the same time, resource allocation is dynamically adjusted according to priority so that resources can be used reasonably and efficiently. For high-priority requests, the system can allocate more computing and network resources to ensure that these requests can be processed in a timely manner. According to the priority queue information, the system can effectively distribute the load, avoid excessive use or waste of resources, and improve the overall performance and stability of the system. By giving priority to requests that are critical to system stability, the risk of system instability or crash caused by failure to handle key tasks in a timely manner is reduced. By dynamically adjusting and optimizing resource allocation, bottleneck problems under high load conditions can be handled in a timely manner to ensure stable operation of the system during peak periods.

[0029] It should be noted that the specific examples are as follows: Assume that the hotel system receives the following concurrent requests: High priority payment request A; Low priority query requirement B; Medium priority reservation request C; Priority queue processing: High priority queue: payment request A; Medium priority queue: reservation request C; Low priority queue: query requirement B; By dynamically adjusting resource allocation, the system allocates more computing resources and network bandwidth to payment request A in the high-priority queue to ensure that it is processed first; medium-priority reservation request C is processed second and allocated appropriate resources; low-priority query request B is processed last and allocated the remaining resources; The system first processes the high-priority payment request A. After processing payment request A, the system processes the medium-priority reservation request C, and finally processes the low-priority query request B; In summary, through the dynamic adjustment of priority queue information and optimal resource configuration, the system can effectively handle concurrent requests, improve response speed and system stability, optimize resource utilization, and enhance user experience. In hotel management systems, this mechanism is particularly important because different types of requests (such as reservations, payments, and inquiries) have different processing requirements and urgency. Reasonable allocation and priority processing can ensure that the system can run efficiently and stably even during peak periods.

[0030] In this embodiment, based on the multi-cloud deployment preset by the hotel terminal, in step S5 of executing the synchronization strategy for the access resource in other cloud platforms, it also includes: S51: Identify the storage support service preset for the access resource, construct a corresponding storage bucket in the other cloud platform based on the storage support service, and allocate access rights corresponding to the storage bucket to the access resource; S52: Determine whether the cloud processing platform can read the access permission; S53: If not, configure a data synchronization task for the access resource from the other cloud platform, verify the synchronization status of the access resource in the cloud processing platform and the other cloud platform, and detect conflicting data of the access resource according to the synchronization status, wherein the data synchronization task specifically includes source and target storage locations, synchronization frequency and synchronization direction.

[0031] In this embodiment, the system identifies the storage support service pre-set for the access resource, builds the corresponding storage bucket in other cloud platforms based on different storage support services, allocates the access rights corresponding to the storage bucket for the access resource, and then the system determines whether the cloud processing platform can read the access rights to execute the corresponding steps; for example, when the system determines that the cloud processing platform can read the access rights corresponding to the storage bucket, the system will consider that the storage support service has successfully built the corresponding storage bucket in other cloud platforms and has allocated appropriate access rights to the access resource. The system will verify that the cloud processing platform does have access rights to read the storage bucket, including verifying whether the platform has the necessary authentication information (such as access keys or tokens) and whether the access policy has been configured correctly. Once it is confirmed that the cloud processing platform has read permissions, the system can start to execute related data operation tasks, such as reading, writing or modifying data in the storage bucket, and monitor the cloud processing platform's access to the storage bucket to ensure that the operation meets expectations and security requirements, and record operation logs for future review and troubleshooting; for example, when the system determines that the cloud processing platform cannot read the access rights corresponding to the storage bucket, the system will consider that the storage support service has not successfully built the storage bucket in other cloud platforms, that is, the hotel terminal cannot obtain the data from the cloud processing platform through Other cloud platforms identify access resources, and the system will configure data synchronization tasks for access resources in other cloud platforms. The data synchronization tasks specifically include source and target storage locations, synchronization frequency and synchronization direction, verify the synchronization status of access resources in the cloud processing platform and other cloud platforms, and detect conflicting data of access resources according to different synchronization statuses; the system can promptly discover and handle the problem that the cloud processing platform cannot access the storage bucket, which helps to discover and correct configuration errors or access permission setting problems of storage support services on other cloud platforms in advance. At the same time, by configuring data synchronization tasks, it ensures that the access resource data stored on different cloud platforms remains consistent, which is particularly important for hotel terminal management systems because it ensures the accuracy and integrity of customer and business data in a multi-cloud environment, and the system can detect and handle conflicting data that may occur during data synchronization, such as data conflicts caused by modifying the same resource on different platforms at the same time. Through effective conflict resolution strategies, the system can ensure data consistency during data synchronization and avoid data damage or loss. When an abnormal synchronization status is found, the system can respond and repair quickly to ensure that the hotel terminal can continue to operate normally. This capability improves the system's fault tolerance and reduces the risk of business interruption caused by cloud platform failures or configuration errors.

[0032] In this embodiment, the step S2 of determining whether the access resource has reached the load upper limit further includes: S21: Based on the monitoring node preconfigured by the hotel terminal, obtain the index data of the access resource to the hotel terminal, wherein the index data specifically includes response time, request rate and error rate; S22: Determine whether the indicator data exceeds a static threshold preset by the hotel terminal; S23: If yes, a preset distributed tracing tool is applied to track the request link for accessing the resource, collect abnormal information of the request link during the tracing process, and identify corresponding abnormal nodes and abnormal bottlenecks according to the abnormal information.

[0033] In this embodiment, the system obtains the index data of the access resource to the hotel terminal based on the monitoring node pre-configured by the hotel terminal, and the index data specifically includes the response time, request rate and error rate, and then the system determines whether these index data exceed the static threshold value pre-set by the hotel terminal to execute the corresponding steps; for example, when the system determines that the index data of the access resource to the hotel terminal does not exceed the static threshold value pre-set by the hotel terminal, the system will consider that the hotel terminal is currently in a normal operating state and the index data is within an acceptable range. The system will continue to monitor and collect the index data of the access resource to ensure that the normal operating state of the hotel terminal is continuously stable, and regularly check whether the setting of the static threshold value is still suitable for the current operating environment and needs. According to new data trends and business changes, it is recommended that hotel staff adjust the static threshold value to better adapt to the actual operating conditions; for example, when the system determines that the index data of the access resource to the hotel terminal exceeds the static threshold value pre-set by the hotel terminal, the system will consider that the hotel terminal is currently in a normal operating state and the index data is within an acceptable range. The client is currently in an abnormal operating state. The system will apply a pre-set distributed tracing tool to track the request link for accessing resources, collect abnormal information of the request link during the tracing process, and identify the corresponding abnormal nodes and abnormal bottlenecks based on different abnormal information; the system uses distributed tracing tools to capture and record abnormal information in the request link in real time, such as high latency, timeout, or error response. This capability enables the system to respond quickly and identify abnormal situations, reducing the time and cost of troubleshooting. At the same time, based on the detailed link data provided by the distributed tracing tool, the system can accurately locate abnormal nodes and bottlenecks, such as the response speed of a specific service may be slow, or the load of a service node may exceed expectations. This precise positioning helps to quickly fix problems and restore normal service operations. By identifying abnormal nodes and bottlenecks, the system can make targeted resource adjustments and optimizations, such as increasing the number of instances of a specific service or adjusting resource allocation strategies to cope with abnormal situations and improve system stability and performance.

[0034] It should be noted that the specific examples are as follows: Assume that the static thresholds pre-set by the hotel terminal include: The response time does not exceed 500 milliseconds; The request rate is maintained at no more than 100 requests per second; The error rate is kept at no more than 1 error per thousand requests; The indicator data monitored by the current system show that The average response time is 800 milliseconds; The request rate is 120 requests per second; The error rate is 5 errors per thousand requests; The system will then analyze the current indicator data and find that it exceeds the pre-set static threshold, and determine that the hotel terminal is in an abnormal operating state; by starting the pre-set distributed tracing tool, it begins to track the request link and capture abnormal information. Because the distributed tracing tool analyzes abnormal conditions in the request link, it may identify that the response time of a service is too long or the load of a service node is too high; based on the abnormal information, the system takes corresponding measures, including increasing the resources of the affected service, optimizing the request scheduling strategy, or performing emergency repairs and restart operations to restore normal service status.

[0035] In this embodiment, the step S4 of determining whether the cloud processing platform detects a preset fault further includes: S41: Obtaining processing progress information of the access resource by the cloud processing platform; S42: Determine whether the processing progress information can be processed when the preset fault occurs; S43: If not, a preset fault analysis tool is applied in the cloud processing platform to perform root cause analysis, locate the fault data block corresponding to the cloud processing platform based on the analysis result, and divide the fault link and fault node from the fault data block.

[0036] In this embodiment, the system obtains the processing progress information of the cloud processing platform for accessing resources, and then the system determines whether the processing progress information can be processed in time when the preset fault is detected, so as to execute the corresponding steps; for example, when the system determines that the processing progress information of the cloud processing platform for accessing resources can be processed in time before the preset fault is detected, the system will consider that the hotel terminal has sufficient fault tolerance and automatic recovery mechanism, and the system will be able to respond immediately and quickly identify the problem when the preset fault is detected, such as identifying the downtime of the service node or the interruption of the network connection, and automatically switch to the backup node or the service instance on other cloud platforms to ensure the continuous availability of access resources, and record all operations and events in the fault handling process, notify the relevant hotel operation and maintenance team or hotel administrator, and help with subsequent fault analysis and improvement; for example, when the system determines that the processing progress information of the cloud processing platform for accessing resources cannot be processed in time before the preset fault is detected Some faults are handled before they occur. At this time, the system will think that the hotel terminal does not have sufficient fault tolerance and recovery mechanisms. The system will use the pre-set fault analysis tools in the cloud processing platform to perform root cause analysis, locate the corresponding fault data blocks in the cloud processing platform based on the analysis results, and divide the fault links and fault nodes from the fault data blocks; the system can use the fault analysis tools to help the system quickly and accurately locate the specific causes of problems in the cloud processing platform, and by analyzing the fault data blocks, the fault links and fault nodes can be identified, and then the specific location that affects the access resource processing can be clarified. At the same time, after locating the root cause of the fault, the system can take targeted repair measures to shorten the time for fault repair. This rapid response helps to minimize service interruption time, improve service availability and user satisfaction, and through detailed analysis of the fault data blocks, the system can improve the fault handling process and strategy to deal with the recurrence of similar problems. This continuous optimization can enhance the stability and reliability of the system.

[0037] In this embodiment, before the step S1 of acquiring the access resources cached by the hotel terminal in real time based on the load upper limit preset by the hotel terminal, the method further includes: S101: Based on the detection script preset in the hotel terminal, simulate the operation process of the hotel user and initiate at least two access requests to the hotel terminal from various preset locations according to the preset frequency, and collect detection data corresponding to the access requests, wherein the operation process specifically includes user login, room query, room reservation and payment, and the detection data specifically includes response time and success rate; S102: Determine whether the detection data exceeds the alarm rule pre-configured by the hotel terminal; S103: If yes, the location information corresponding to the access request is marked from the hotel terminal, and the access request is interactively bridged with a manual channel preset by the hotel terminal according to the location information to collect the required access resources corresponding to the access request.

[0038] In this embodiment, the system is based on the detection script pre-set in the hotel terminal, simulates the operation process of the hotel user and according to the pre-set frequency, the operation process specifically includes user login, room query, room reservation and payment, and initiates at least two access requests to the hotel terminal from each pre-set location, collects the detection data corresponding to these access requests, and the detection data specifically includes response time and success rate. Then the system determines whether these detection data exceed the alarm rules pre-configured by the hotel terminal to execute the corresponding steps; for example, when the system determines that the detection data corresponding to the access request does not exceed the alarm rules pre-configured by the hotel terminal, the system will consider that the current operation process of the hotel terminal is in a normal operating state, and the system can continue to monitor and record the detection data. The system will continue to collect, monitor and record the detection data, including indicators such as response time and success rate, which is helpful to understand the performance of the hotel terminal and the stability of the operation process in real time. At the same time, for the collected detection data, the system will regularly perform performance analysis and trend analysis. By comparing data from different time periods, performance changes and trends can be identified, and potential performance problems or improvement opportunities can be discovered in a timely manner. Based on the results of the detection data analysis, hotel staff can In order to identify bottlenecks or optimization space in the operation process, it involves improving the code, optimizing resource allocation or adjusting the operation process to improve the overall efficiency and performance of the system; for example, when the system determines that the detection data corresponding to the access request exceeds the alarm rules pre-configured by the hotel terminal, the system will consider the operation process of the hotel terminal to be abnormal, and the system will mark the location information corresponding to the access request from the hotel terminal, and connect the access request with the pre-set manual channel of the hotel terminal according to different location information to collect the required access resources corresponding to the access request; the system quickly marks the abnormal access request and takes corresponding measures, thereby shortening the response time of the hotel terminal to the abnormal processing of the preset location, which helps to reduce the impact of the abnormality on the user experience and service quality. At the same time, according to the marked location information, the system can accurately locate the specific location or link where the abnormality occurs, which provides an important reference and basis for subsequent troubleshooting and problem solving, and connects the abnormal access request with the pre-set manual channel to realize the effective combination of automated process and manual intervention, which can not only handle abnormal situations in a timely manner, but also ensure that manual experts intervene to conduct more in-depth problem analysis and processing when necessary.

[0039] Reference Figure 2 , is a hotel room management system based on cloud computing in one embodiment of the present invention, comprising: An acquisition module 10, for acquiring access resources cached by the hotel terminal in real time based on a load upper limit preset by the hotel terminal; A judging module 20, configured to judge whether the access resource has reached the load upper limit; The execution module 30 is used for, if not, applying the pre-trained prediction model to predict the access resources, scheduling the resource content of the hotel terminal according to the prediction result, identifying the demand type corresponding to the access resources, dynamically allocating the resource content to the hybrid cloud architecture preset by the hotel terminal according to the demand type, and generating a cloud processing platform corresponding to the access resources, wherein the hybrid cloud architecture specifically includes a public cloud and a private cloud; A second judgment module 40 is used to judge whether the cloud processing platform detects a preset fault; The second execution module 50 is used to activate the distributed application architecture preset by the hotel terminal, allow the access resources to switch and expand between the hybrid cloud architecture, execute synchronization strategies on the access resources in other cloud platforms based on the multi-cloud deployment preset by the hotel terminal, obtain the disaster recovery strategy data of the cloud processing platform, and switch the access resources from the other cloud platforms back to the cloud processing platform according to the disaster recovery strategy data, wherein the synchronization strategy specifically includes full synchronization and incremental synchronization, and the disaster recovery strategy data specifically includes recovery time objectives and recovery point objectives.

[0040] In this embodiment, the acquisition module 10 acquires the access resources cached by the hotel terminal in real time based on the server load upper limit pre-set by the hotel terminal, and then the judgment module 20 judges whether these access resources have reached the pre-set server load upper limit to execute the corresponding steps; for example, when the system determines that the access resources cached by the hotel terminal have reached the pre-set server load upper limit, the system will believe that the current server can no longer handle more requests, and further increasing the load will cause performance degradation or system crash. The system will configure a load balancer to evenly distribute requests to different servers, dynamically adjust the load balancing strategy, adjust the distribution strategy according to the real-time load situation, increase the cache layer, use the distributed cache system to improve the cache hit rate, adjust the cache expiration strategy, extend the cache time, reduce the load on the server, and use the microservice architecture to independently deploy and run different service modules, configure the service grid, and realize dynamic routing and isolation between services; for example, when the system determines that the access resources cached by the hotel terminal have not reached the pre-set server load upper limit, the execution module 30 will believe that the current server can still handle the access request, and the system will apply the pre-trained pre- The prediction model predicts the access resources, schedules the resource content of the hotel terminal according to different prediction results, identifies the demand type corresponding to the access resources, and dynamically allocates the resource content to the hybrid cloud architecture pre-set by the hotel terminal according to these demand types. The hybrid cloud architecture specifically includes public cloud and private cloud, and generates a cloud processing platform corresponding to the access resources; through the prediction model, the system can identify future access needs in advance, optimize resource allocation, avoid resource waste, dynamically schedule resources, ensure that resources can be allocated on demand, reduce resource idleness, and improve overall resource utilization. At the same time, through advance prediction and scheduling, the system can reserve sufficient resources before the demand increases, avoid delays caused by temporary expansion, and dynamically allocate resource content to a suitable cloud platform according to the demand type, balance the load, and ensure that each node can operate efficiently. The hybrid cloud architecture allows the system to dynamically expand or reduce resources according to actual needs, and flexibly respond to different load conditions. The combination of public cloud and private cloud allows the system to flexibly adjust the resource allocation strategy and select the optimal resource deployment method according to different business needs; then the second judgment module 40 determines whether the cloud processing platform detects the pre-set fault content to execute the corresponding steps;For example, when the system determines that the cloud processing platform has not detected the pre-set fault content, the system will consider that the access resources in the hotel terminal processing do not need to be transferred from the cloud processing platform to other cloud platforms. The system will use monitoring tools to monitor various indicators of the system in real time, such as CPU usage, memory usage and network traffic, configure alarm mechanisms, set warning thresholds, and issue alarms in advance when indicators are close to critical values, take timely measures, and regularly update the prediction model to ensure its accuracy and real-time performance. According to the prediction results, resources are prepared and scheduled in advance to ensure that they can meet future access needs, and identify the types of access resource requirements, such as short-term high-frequency access and long-term stable access, and dynamically adjust resource allocation strategies to ensure that resource allocation meets actual needs; for example, when the system determines that the cloud processing platform detects the pre-set fault content, the second execution module 50 will consider that the access resources in the hotel terminal processing need to be transferred from the cloud processing platform to other cloud platforms. The system will activate the pre-set distributed application architecture of the hotel terminal, allowing access resources to switch and expand between hybrid cloud architectures, and based on the pre-set multi-cloud deployment of the hotel terminal, the access resources in other cloud platforms are allocated. The system executes synchronization strategies for data, including full synchronization and incremental synchronization, obtains disaster recovery strategy data of the cloud processing platform, including recovery time objectives and recovery point objectives, and switches access resources from other cloud platforms back to the cloud processing platform according to these disaster recovery strategy data; the system deploys synchronization strategies (full synchronization and incremental synchronization) for access resources on other cloud platforms to ensure that even if the main cloud platform fails, the system can still quickly switch to the backup cloud platform to ensure business continuity and availability. According to the disaster recovery strategy data (recovery time objectives and recovery point objectives), a reasonable switching plan is formulated to minimize business interruption time and data loss risks. At the same time, the pre-set distributed application architecture is activated to allow access resources to dynamically switch and expand between hybrid cloud architectures. Based on the multi-cloud deployment strategy, the access resources are deployed on multiple cloud platforms to disperse risks and improve the elasticity and reliability of the system. In addition, according to business needs and disaster recovery strategies, the allocation and use of resources are dynamically adjusted to optimize resource utilization efficiency. Through multi-cloud deployment and resource optimization, the cost risk caused by dependence on a single cloud platform is reduced, and more flexible and cost-effective resource management is achieved. ;

[0041] In this embodiment, it also includes: A construction module is used to identify the characteristic target variable corresponding to the access resource, calculate the score data of the characteristic target variable, divide the characteristic target variable into a preferred feature subset and a secondary feature subset based on the score data, perform recursive feature elimination on the access resource, and construct the characteristic data required by the prediction model, wherein the characteristic target variable specifically includes numerical features, classification features, time features and text features; A third judgment module is used to judge whether the characteristic data have the same dimension range; The third execution module is used to collect data distribution characteristics of the feature data, discretize the feature data according to the data distribution characteristics, and calculate the interactive characteristics of the feature data, wherein the data distribution characteristics specifically include long-tail distribution and multimodal distribution, and the discretization specifically includes equal-width discretization and equal-frequency discretization.

[0042] In this embodiment, the system identifies the characteristic target variables corresponding to the access resources, which specifically include numerical features, categorical features, time features, and text features, calculates the scoring data of these characteristic target variables, divides the preferred feature subsets and the secondary feature subsets from the characteristic target variables based on different scoring data, and recursively eliminates the features of the access resources to construct the characteristic data required by the prediction model. Then the system determines whether these characteristic data have the same dimensional range to execute the corresponding steps; for example, when the system determines that the characteristic data required by the prediction model have the same dimensional range, the system will consider that all characteristic data have been standardized and can be directly used for model training. The system will use these standardized characteristic data for model training and prediction, deploy the trained prediction model to the production environment, ensure that the model can process new access resource data of the hotel terminal in real time or in batches, monitor the real-time prediction performance of the model, ensure the stability and accuracy of the model in the production environment, configure the monitoring and alarm mechanism, and when the model prediction performance decreases or anomalies occur, adjust and process them in time, and update the model regularly, retrain the model with the latest data, maintain the prediction accuracy and adaptability of the model, and In scenarios where it is needed, the model can learn new data in real time using online learning technology and gradually improve prediction performance. For example, when the system determines that the feature data required by the prediction model do not have the same dimensional range, the system will consider that the feature data has not been standardized and cannot be directly used for model training. The system will collect the data distribution characteristics of these feature data, which specifically include long-tail distribution and multimodal distribution, discretize the feature data according to different data distribution characteristics, including equal-width discretization and equal-frequency discretization, and calculate the interactive characteristics of these feature data. Through standardization, the system can convert feature data of different scales into the same dimensional range, so that each feature has equal importance in the model training process, avoiding some features from having too much impact on the model due to their large numerical range. At the same time, when the feature data has a unified dimensional range, the gradient descent algorithm in the model training process is more stable and efficient, which helps to speed up the model convergence speed and improve training efficiency. It can also identify data distribution characteristics (such as long-tail distribution and multimodal distribution) and perform targeted discretization, which can effectively handle extreme values ​​and noise in the data and improve the robustness and generalization ability of the model.

[0043] In this embodiment, the execution module includes: A generating unit, configured to extract key features of the concurrent requests pre-received by the hotel terminal and generate priority queue information of the access resource, wherein the concurrent requests specifically include reservation requests, query requests and payment requests; A determination unit, configured to determine whether the priority queue information has different types of priorities; An execution unit is used to respond to the corresponding pending resources from the priority queue information, dynamically adjust the execution decision of the hotel terminal on the pending resources according to the priority queue information, and divide the allocated resources required for the pending resources from the preset server according to the execution decision.

[0044] In this embodiment, the system is based on concurrent requests pre-received by the hotel terminal, and the concurrent requests specifically include reservation requests, query requests and payment requests. The system extracts key features of these concurrent requests and generates priority queue information for accessing resources. The system then determines whether these priority queue information have different types of priorities to execute corresponding steps; for example, when the system determines that the priority queue information for accessing resources does not have different types of priorities, the system will consider that all concurrent requests are equally important in the current situation, and there are no requests that need to be prioritized. The system will add all concurrent requests to the processing queue in the order they are received, and process each request in turn in the order of the queue without prioritizing, while ensuring that all requests are processed fairly and that some requests are not delayed due to different priorities; for example, when the system determines that the priority queue information for accessing resources has different types of priorities, the system will consider that concurrent requests have priority processing, and the system will respond to the corresponding resources to be processed from the priority queue information. According to the priority queue information, the hotel terminal's execution decision on the resources to be processed is dynamically adjusted, and the allocated resources required for the resources to be processed are divided from the pre-set servers according to different execution decisions; the system can respond quickly to key business needs by giving priority to high-priority requests (such as payment requests), improve user experience and satisfaction, ensure that key operations can be given priority, and reduce the waiting time of key operations, thereby improving the overall response speed of the system. At the same time, resource allocation is dynamically adjusted according to priority so that resources can be used reasonably and efficiently. For high-priority requests, the system can allocate more computing and network resources to ensure that these requests can be processed in a timely manner. According to the priority queue information, the system can effectively distribute the load, avoid excessive use or waste of resources, and improve the overall performance and stability of the system. By giving priority to requests that are critical to system stability, the risk of system instability or crash caused by failure to handle key tasks in a timely manner is reduced. By dynamically adjusting and optimizing resource allocation, bottleneck problems under high load conditions can be handled in a timely manner to ensure stable operation of the system during peak periods.

[0045] In this embodiment, the second execution module further includes: An identification unit, used to identify the storage support service preset for the access resource, construct a corresponding storage bucket in the other cloud platform based on the storage support service, and allocate access rights corresponding to the storage bucket to the access resource; A second determination unit, configured to determine whether the cloud processing platform can read the access permission; The second execution unit is used to, if not, configure a data synchronization task for the access resource from the other cloud platform, verify the synchronization status of the access resource in the cloud processing platform and the other cloud platform, and detect conflicting data of the access resource according to the synchronization status, wherein the data synchronization task specifically includes source and target storage locations, synchronization frequency and synchronization direction.

[0046] In this embodiment, the system identifies the storage support service pre-set for the access resource, builds the corresponding storage bucket in other cloud platforms based on different storage support services, allocates the access rights corresponding to the storage bucket for the access resource, and then the system determines whether the cloud processing platform can read the access rights to execute the corresponding steps; for example, when the system determines that the cloud processing platform can read the access rights corresponding to the storage bucket, the system will consider that the storage support service has successfully built the corresponding storage bucket in other cloud platforms and has allocated appropriate access rights to the access resource. The system will verify that the cloud processing platform does have access rights to read the storage bucket, including verifying whether the platform has the necessary authentication information (such as access keys or tokens) and whether the access policy has been configured correctly. Once it is confirmed that the cloud processing platform has read permissions, the system can start to execute related data operation tasks, such as reading, writing or modifying data in the storage bucket, and monitor the cloud processing platform's access to the storage bucket to ensure that the operation meets expectations and security requirements, and record operation logs for future review and troubleshooting; for example, when the system determines that the cloud processing platform cannot read the access rights corresponding to the storage bucket, the system will consider that the storage support service has not successfully built the storage bucket in other cloud platforms, that is, the hotel terminal cannot obtain the data from the cloud processing platform through Other cloud platforms identify access resources, and the system will configure data synchronization tasks for access resources in other cloud platforms. The data synchronization tasks specifically include source and target storage locations, synchronization frequency and synchronization direction, verify the synchronization status of access resources in the cloud processing platform and other cloud platforms, and detect conflicting data of access resources according to different synchronization statuses; the system can promptly discover and handle the problem that the cloud processing platform cannot access the storage bucket, which helps to discover and correct configuration errors or access permission setting problems of storage support services on other cloud platforms in advance. At the same time, by configuring data synchronization tasks, it ensures that the access resource data stored on different cloud platforms remains consistent, which is particularly important for hotel terminal management systems because it ensures the accuracy and integrity of customer and business data in a multi-cloud environment, and the system can detect and handle conflicting data that may occur during data synchronization, such as data conflicts caused by modifying the same resource on different platforms at the same time. Through effective conflict resolution strategies, the system can ensure data consistency during data synchronization and avoid data damage or loss. When an abnormal synchronization status is found, the system can respond and repair quickly to ensure that the hotel terminal can continue to operate normally. This capability improves the system's fault tolerance and reduces the risk of business interruption caused by cloud platform failures or configuration errors.

[0047] In this embodiment, the judgment module further includes: An acquisition unit, configured to acquire index data of the access resource to the hotel terminal based on a monitoring node preconfigured by the hotel terminal, wherein the index data specifically includes response time, request rate and error rate; A third judgment unit, used to judge whether the indicator data exceeds a static threshold preset by the hotel terminal; The third execution unit is used to apply a preset distributed tracing tool to track the request link for accessing the resource, collect abnormal information of the request link during the tracing process, and identify corresponding abnormal nodes and abnormal bottlenecks according to the abnormal information.

[0048] In this embodiment, the system obtains the index data of the access resource to the hotel terminal based on the monitoring node pre-configured by the hotel terminal, and the index data specifically includes the response time, request rate and error rate, and then the system determines whether these index data exceed the static threshold value pre-set by the hotel terminal to execute the corresponding steps; for example, when the system determines that the index data of the access resource to the hotel terminal does not exceed the static threshold value pre-set by the hotel terminal, the system will consider that the hotel terminal is currently in a normal operating state and the index data is within an acceptable range. The system will continue to monitor and collect the index data of the access resource to ensure that the normal operating state of the hotel terminal is continuously stable, and regularly check whether the setting of the static threshold value is still suitable for the current operating environment and needs. According to new data trends and business changes, it is recommended that hotel staff adjust the static threshold value to better adapt to the actual operating conditions; for example, when the system determines that the index data of the access resource to the hotel terminal exceeds the static threshold value pre-set by the hotel terminal, the system will consider that the hotel terminal is currently in a normal operating state and the index data is within an acceptable range. The client is currently in an abnormal operating state. The system will apply a pre-set distributed tracing tool to track the request link for accessing resources, collect abnormal information of the request link during the tracing process, and identify the corresponding abnormal nodes and abnormal bottlenecks based on different abnormal information; the system uses distributed tracing tools to capture and record abnormal information in the request link in real time, such as high latency, timeout, or error response. This capability enables the system to respond quickly and identify abnormal situations, reducing the time and cost of troubleshooting. At the same time, based on the detailed link data provided by the distributed tracing tool, the system can accurately locate abnormal nodes and bottlenecks, such as the response speed of a specific service may be slow, or the load of a service node may exceed expectations. This precise positioning helps to quickly fix problems and restore normal service operations. By identifying abnormal nodes and bottlenecks, the system can make targeted resource adjustments and optimizations, such as increasing the number of instances of a specific service or adjusting resource allocation strategies to cope with abnormal situations and improve system stability and performance.

[0049] In this embodiment, the second determination module further includes: A second acquisition unit, configured to acquire processing progress information of the access resource by the cloud processing platform; A fourth judgment unit, used to judge whether the processing progress information can be processed completely when the preset fault occurs; The fourth execution unit is used to apply a preset fault analysis tool in the cloud processing platform to perform root cause analysis if no, locate the fault data block corresponding to the cloud processing platform based on the analysis result, and divide the fault link and fault node from the fault data block.

[0050] In this embodiment, the system obtains the processing progress information of the cloud processing platform for accessing resources, and then the system determines whether the processing progress information can be processed in time when the preset fault is detected, so as to execute the corresponding steps; for example, when the system determines that the processing progress information of the cloud processing platform for accessing resources can be processed in time before the preset fault is detected, the system will consider that the hotel terminal has sufficient fault tolerance and automatic recovery mechanism, and the system will be able to respond immediately and quickly identify the problem when the preset fault is detected, such as identifying the downtime of the service node or the interruption of the network connection, and automatically switch to the backup node or the service instance on other cloud platforms to ensure the continuous availability of access resources, and record all operations and events in the fault handling process, notify the relevant hotel operation and maintenance team or hotel administrator, and help with subsequent fault analysis and improvement; for example, when the system determines that the processing progress information of the cloud processing platform for accessing resources cannot be processed in time before the preset fault is detected Some faults are handled before they occur. At this time, the system will think that the hotel terminal does not have sufficient fault tolerance and recovery mechanisms. The system will use the pre-set fault analysis tools in the cloud processing platform to perform root cause analysis, locate the corresponding fault data blocks in the cloud processing platform based on the analysis results, and divide the fault links and fault nodes from the fault data blocks; the system can use the fault analysis tools to help the system quickly and accurately locate the specific causes of problems in the cloud processing platform, and by analyzing the fault data blocks, the fault links and fault nodes can be identified, and then the specific location that affects the access resource processing can be clarified. At the same time, after locating the root cause of the fault, the system can take targeted repair measures to shorten the time for fault repair. This rapid response helps to minimize service interruption time, improve service availability and user satisfaction, and through detailed analysis of the fault data blocks, the system can improve the fault handling process and strategy to deal with the recurrence of similar problems. This continuous optimization can enhance the stability and reliability of the system.

[0051] In this embodiment, it also includes: A simulation module is used to simulate the operation process of the hotel user based on the detection script preset in the hotel terminal and initiate at least two access requests to the hotel terminal from various preset locations according to a preset frequency, and collect detection data corresponding to the access requests, wherein the operation process specifically includes user login, room query, room reservation and payment, and the detection data specifically includes response time and success rate; A fourth judgment module, used to judge whether the detection data exceeds the alarm rule pre-configured by the hotel terminal; The fourth execution module is used to, if yes, mark the location information corresponding to the access request from the hotel terminal, interactively bridge the access request with the manual channel preset by the hotel terminal according to the location information, and collect the required access resources corresponding to the access request.

[0052] In this embodiment, the system is based on the detection script pre-set in the hotel terminal, simulates the operation process of the hotel user and according to the pre-set frequency, the operation process specifically includes user login, room query, room reservation and payment, and initiates at least two access requests to the hotel terminal from each pre-set location, collects the detection data corresponding to these access requests, and the detection data specifically includes response time and success rate. Then the system determines whether these detection data exceed the alarm rules pre-configured by the hotel terminal to execute the corresponding steps; for example, when the system determines that the detection data corresponding to the access request does not exceed the alarm rules pre-configured by the hotel terminal, the system will consider that the current operation process of the hotel terminal is in a normal operating state, and the system can continue to monitor and record the detection data. The system will continue to collect, monitor and record the detection data, including indicators such as response time and success rate, which is helpful to understand the performance of the hotel terminal and the stability of the operation process in real time. At the same time, for the collected detection data, the system will regularly perform performance analysis and trend analysis. By comparing data from different time periods, performance changes and trends can be identified, and potential performance problems or improvement opportunities can be discovered in a timely manner. Based on the results of the detection data analysis, hotel staff can In order to identify bottlenecks or optimization space in the operation process, it involves improving the code, optimizing resource allocation or adjusting the operation process to improve the overall efficiency and performance of the system; for example, when the system determines that the detection data corresponding to the access request exceeds the alarm rules pre-configured by the hotel terminal, the system will consider the operation process of the hotel terminal to be abnormal, and the system will mark the location information corresponding to the access request from the hotel terminal, and connect the access request with the pre-set manual channel of the hotel terminal according to different location information to collect the required access resources corresponding to the access request; the system quickly marks the abnormal access request and takes corresponding measures, thereby shortening the response time of the hotel terminal to the abnormal processing of the preset location, which helps to reduce the impact of the abnormality on the user experience and service quality. At the same time, according to the marked location information, the system can accurately locate the specific location or link where the abnormality occurs, which provides an important reference and basis for subsequent troubleshooting and problem solving, and connects the abnormal access request with the pre-set manual channel to realize the effective combination of automated process and manual intervention, which can not only handle abnormal situations in a timely manner, but also ensure that manual experts intervene to conduct more in-depth problem analysis and processing when necessary.

[0053] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A hotel room management method based on cloud computing, characterized in that: The following steps are involved: Based on the load upper limit preset by the hotel terminal, access resources cached by the hotel terminal are obtained in real time; Determining whether the access resource has reached the load upper limit; If not, a pre-trained prediction model is applied to predict the access resources, resource contents of the hotel terminal are scheduled according to the prediction results, the demand types corresponding to the access resources are identified, and the resource contents are dynamically allocated to the hybrid cloud architecture preset by the hotel terminal according to the demand types, and a cloud processing platform corresponding to the access resources is generated, wherein the hybrid cloud architecture specifically includes a public cloud and a private cloud; Determining whether the cloud processing platform detects a preset fault; If so, activate the distributed application architecture preset by the hotel terminal, allow the access resources to switch and expand between the hybrid cloud architectures, execute synchronization strategies on the access resources in other cloud platforms based on the multi-cloud deployment preset by the hotel terminal, obtain the disaster recovery strategy data of the cloud processing platform, and switch the access resources from the other cloud platforms back to the cloud processing platform according to the disaster recovery strategy data, wherein the synchronization strategy specifically includes full synchronization and incremental synchronization, and the disaster recovery strategy data specifically includes recovery time objectives and recovery point objectives.

2. The hotel room management method based on cloud computing according to claim 1, characterized in that: Before the step of applying the pre-trained prediction model to predict the access resources and scheduling the resource content of the hotel terminal according to the prediction result, the method further includes: Identify the feature target variable corresponding to the access resource, calculate the score data of the feature target variable, divide the feature target variable into a preferred feature subset and a secondary feature subset based on the score data, perform recursive feature elimination on the access resource, and construct the feature data required by the prediction model, wherein the feature target variable specifically includes numerical features, classification features, time features, and text features; Determining whether the characteristic data have the same dimension range; If not, then collect the data distribution characteristics of the feature data, discretize the feature data according to the data distribution characteristics, and calculate the interactive characteristics of the feature data, wherein the data distribution characteristics specifically include long-tail distribution and multimodal distribution, and the discretization specifically includes equal-width discretization and equal-frequency discretization.

3. The hotel room management method based on cloud computing according to claim 1, characterized in that: The step of identifying the type of requirement corresponding to the access resource includes: Based on the concurrent requests pre-received by the hotel terminal, extract key features of the concurrent requests to generate priority queue information of the access resources, wherein the concurrent requests specifically include reservation requests, query requests and payment requests; Determine whether the priority queue information has different types of priorities; If so, respond to the corresponding resources to be processed from the priority queue information, dynamically adjust the execution decision of the hotel terminal on the resources to be processed according to the priority queue information, and divide the allocated resources required for the resources to be processed from the preset server according to the execution decision.

4. The hotel room management method based on cloud computing according to claim 1, characterized in that: The step of executing a synchronization strategy on the access resource in other cloud platforms based on the preset multi-cloud deployment of the hotel terminal also includes: Identify the storage support service preset for the access resource, build a corresponding storage bucket in the other cloud platform based on the storage support service, and assign access rights corresponding to the storage bucket to the access resource; Determining whether the cloud processing platform can read the access permission; If not, configure a data synchronization task for the access resource from the other cloud platform, verify the synchronization status of the access resource in the cloud processing platform and the other cloud platform, and detect conflicting data of the access resource based on the synchronization status, wherein the data synchronization task specifically includes source and target storage locations, synchronization frequency and synchronization direction.

5. The hotel room management method based on cloud computing according to claim 1, characterized in that: The step of determining whether the access resource has reached the load upper limit further includes: Based on the monitoring node preconfigured by the hotel terminal, obtaining the index data of the access resource to the hotel terminal, wherein the index data specifically includes response time, request rate and error rate; Determining whether the indicator data exceeds a static threshold preset by the hotel terminal; If so, a preset distributed tracing tool is applied to track the request link for accessing the resource, collect abnormal information of the request link during the tracing process, and identify corresponding abnormal nodes and abnormal bottlenecks according to the abnormal information.

6. The hotel room management method based on cloud computing according to claim 1, characterized in that: The step of determining whether the cloud processing platform has detected a preset fault also includes: Obtaining processing progress information of the access resource by the cloud processing platform; Determine whether the processing progress information can be processed when the preset fault occurs; If not, a preset fault analysis tool is applied in the cloud processing platform to perform root cause analysis, locate the fault data block corresponding to the cloud processing platform based on the analysis result, and divide the fault link and fault node from the fault data block.

7. The hotel room management method based on cloud computing according to claim 1, characterized in that: Before the step of acquiring the access resources cached by the hotel terminal in real time based on the load upper limit preset by the hotel terminal, the method further includes: Based on the detection script preset by the hotel terminal, the operation process of the hotel user is simulated and at least two access requests are initiated from various preset locations to the hotel terminal according to the preset frequency, and the detection data corresponding to the access requests are collected, wherein the operation process specifically includes user login, room query, room reservation and payment, and the detection data specifically includes response time and success rate; Determining whether the detection data exceeds the alarm rules preconfigured by the hotel terminal; If so, the location information corresponding to the access request is marked from the hotel terminal, and the access request is interactively bridged with the manual channel preset by the hotel terminal according to the location information to collect the access resources corresponding to the access request.

8. A hotel room management system based on cloud computing, characterized in that: include: An acquisition module, used to acquire access resources cached by the hotel terminal in real time based on a preset load upper limit of the hotel terminal; A judgment module, used to judge whether the access resource has reached the load upper limit; An execution module is used for, if not, applying a pre-trained prediction model to predict the access resource, scheduling the resource content of the hotel terminal according to the prediction result, identifying the demand type corresponding to the access resource, dynamically allocating the resource content to the hybrid cloud architecture preset by the hotel terminal according to the demand type, and generating a cloud processing platform corresponding to the access resource, wherein the hybrid cloud architecture specifically includes a public cloud and a private cloud; A second judgment module is used to judge whether the cloud processing platform detects a preset fault; The second execution module is used to activate the distributed application architecture preset by the hotel terminal, allow the access resources to switch and expand between the hybrid cloud architecture, execute synchronization strategies on the access resources in other cloud platforms based on the multi-cloud deployment preset by the hotel terminal, obtain the disaster recovery strategy data of the cloud processing platform, and switch the access resources from the other cloud platforms back to the cloud processing platform according to the disaster recovery strategy data, wherein the synchronization strategy specifically includes full synchronization and incremental synchronization, and the disaster recovery strategy data specifically includes recovery time objectives and recovery point objectives.

9. The hotel room management system based on cloud computing according to claim 8, characterized in that: Also includes: A construction module is used to identify the characteristic target variable corresponding to the access resource, calculate the score data of the characteristic target variable, divide the characteristic target variable into a preferred feature subset and a secondary feature subset based on the score data, perform recursive feature elimination on the access resource, and construct the characteristic data required by the prediction model, wherein the characteristic target variable specifically includes numerical features, classification features, time features and text features; A third judgment module is used to judge whether the characteristic data have the same dimension range; The third execution module is used to collect data distribution characteristics of the feature data, discretize the feature data according to the data distribution characteristics, and calculate the interactive characteristics of the feature data, wherein the data distribution characteristics specifically include long-tail distribution and multimodal distribution, and the discretization specifically includes equal-width discretization and equal-frequency discretization.

10. The hotel room management system based on cloud computing according to claim 8, characterized in that: The execution module includes: A generating unit, configured to extract key features of the concurrent requests pre-received by the hotel terminal and generate priority queue information of the access resource, wherein the concurrent requests specifically include reservation requests, query requests and payment requests; A determination unit, configured to determine whether the priority queue information has different types of priorities; An execution unit is used to respond to the corresponding pending resources from the priority queue information, dynamically adjust the execution decision of the hotel terminal on the pending resources according to the priority queue information, and divide the allocated resources required for the pending resources from the preset server according to the execution decision.

Images