Intelligent property management and deployment method and system based on big data

Abstract

The invention provides an intelligent property management and allocation method and system based on big data, and the method comprises the steps: carrying out the data compression and anomaly detection of the operation state information of property facilities and user demand feedback through an edge computing technology, obtaining screened data, constructing a facility health assessment model, and carrying out the data compression and anomaly detection. The method comprises the steps of generating a facility health assessment report, constructing a service response mode library based on the facility health assessment report in combination with user demand feedback to generate an optimal service response scheme and generate a task work order, performing monitoring and assessment by using a big data analysis technology and a machine learning algorithm, and collecting actual effect feedback. According to the technical scheme provided by the invention, the data transmission quantity is effectively reduced, the response speed of the system is improved, the system can be ensured to continuously evolve and adapt to new situations, and the management efficiency and the service quality are further improved.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of big data technology, and in particular to a property intelligent management and allocation method based on big data. Background Art

[0002] With the acceleration of the urbanization process and the increasing complexity of property management requirements, the intelligent management of property facilities has become the key to improving service quality and reducing operating costs. The traditional property management method relies on manual inspections and empirical judgments, making it difficult to achieve real-time monitoring and precise maintenance of a large number of property facilities. In recent years, the development of big data technology and intelligent algorithms has provided new solutions for property intelligent management. By integrating advanced technologies such as edge computing, Bayesian networks, time series analysis, genetic algorithms, and reinforcement learning, it is possible to dynamically evaluate the operating status of property facilities and predict potential faults, thereby optimizing the service response mode and improving the reliability of facilities and user satisfaction.

[0003] However, there are still some deficiencies in the existing technologies in practical applications. On the one hand, the traditional property management system lacks effective integration and in-depth mining of multi-source heterogeneous data, and cannot fully utilize historical maintenance records and real-time operating status information for accurate facility health assessment. On the other hand, the existing service response schemes often rely on fixed patterns and do not fully consider the specific conditions of facilities and the personalized needs of users, resulting in unreasonable resource allocation and low service efficiency. In addition, most systems lag in updating model parameters and optimizing decision rules and cannot adapt to the changes in the latest situation in a timely manner, affecting the overall management effect. Summary of the Invention

[0004] The embodiments of the present invention provide a property intelligent management and allocation method and system based on big data to solve the problems in the existing technologies, such as the lack of effective integration and in-depth mining of multi-source heterogeneous data, the inability to fully utilize historical maintenance records and real-time operating status information for accurate facility health assessment, and the failure to fully consider the specific conditions of facilities and the personalized needs of users, resulting in unreasonable resource allocation and low service efficiency.

[0005] In a first aspect, the embodiments of the present invention provide a property intelligent management and allocation method based on big data, including:

[0006] Using edge computing technology to perform preliminary data compression and anomaly detection on the real-time operating status information of multiple property facilities and user demand feedback to obtain the filtered data;

[0007] Based on the filtered data and historical maintenance records, a facility health assessment model is constructed by combining Bayesian network and time series analysis. The facility health assessment model is used to dynamically evaluate the facility health status and predict potential failures, generating a facility health assessment report. At the same time, the incremental learning algorithm is used to update the parameters of the facility health assessment model in real time, optimizing the facility health assessment model to adapt to the latest situation, and obtaining an adaptive facility health assessment model;

[0008] Based on the performance score, expected failure probability, and recommended maintenance cycle in the facility health assessment report, combined with user requirement feedback, a service response pattern library is constructed. The genetic algorithm and reinforcement learning technology are used to optimize and match the service patterns in the service response pattern library, generating an optimal service response plan. The service response pattern library includes: historical data, best practices, and industry standards;

[0009] Using natural language generation technology and geographic information system, a task work order is automatically generated according to the optimal service response plan, providing operation guidelines for the property management team. At the same time, the expected completion time and service content notification are sent to relevant users through the message queue mechanism to obtain optimized task allocation and user notification;

[0010] Based on the optimized task allocation and user notification, the big data analysis technology and machine learning algorithm are used to monitor and evaluate the effect of the optimal service response plan, collecting actual effect feedback. According to the search for actual effect feedback, the weight allocation and decision rules in the adaptive facility health assessment model and the optimal service response plan are adjusted to form a closed-loop optimization mechanism.

[0011] Optionally, based on the filtered data and historical maintenance records, a facility health assessment model is constructed by combining Bayesian network and time series analysis. The facility health assessment model is used to dynamically evaluate the facility health status and predict potential failures, generating a facility health assessment report. At the same time, the incremental learning algorithm is used to update the parameters of the facility health assessment model in real time, optimizing the facility health assessment model to adapt to the latest situation, and obtaining an adaptive facility health assessment model, including:

[0012] Using the method of combining Bayesian network and time series analysis, multi-dimensional feature fusion processing is performed on the filtered data and historical maintenance records received by the cloud data center, obtaining a concise and information-rich high-dimensional feature representation. At the same time, the autoencoder in the deep learning framework is used to reduce the dimension and extract features of the filtered data, generating a concise and information-rich feature set;

[0013] Based on the concise and information-rich feature set, the facility health assessment model is constructed and trained using a deep neural network and a generative adversarial network, obtaining an initial facility health assessment model;

[0014] Using the facility health assessment model to dynamically assess the health status of the facility and predict potential failures, a facility health assessment report is generated that includes the performance score of each facility, the expected failure probability, and the recommended maintenance cycle;

[0015] According to the facility health assessment report, an adaptive incremental learning algorithm is designed and applied to update the parameters of the facility health assessment model in real time to obtain an adaptive facility health assessment model.

[0016] Optionally, according to the facility health assessment report, an adaptive incremental learning algorithm is designed and applied to update the parameters of the facility health assessment model in real time to obtain an adaptive facility health assessment model, including:

[0017] Based on the performance score, expected failure probability and recommended maintenance cycle in the facility health assessment report, a feedback analysis model is constructed using a decision tree and a random forest algorithm, and the feedback analysis model is used to identify the factors that have the greatest impact on the health status of the facility, thereby obtaining a list of key influencing factors;

[0018] Based on the list of key influencing factors, the parameter adjustment strategy of the facility health assessment model is optimized using the Q-learning algorithm in the reinforcement learning framework to obtain an optimized parameter adjustment strategy;

[0019] Combined with the optimized parameter adjustment strategy, a transfer learning mechanism is introduced to enhance the adaptive incremental learning algorithm. By comparing the similarity between the current facility and the historical case, the most relevant experience is automatically selected for transfer to generate an enhanced adaptive incremental learning algorithm.

[0020] Using online learning technology and sliding window method, combined with enhanced adaptive incremental learning algorithm, the parameters of the facility health assessment model are dynamically updated one by one to obtain real-time updated parameters of the facility health assessment model.

[0021] Based on the real-time updated facility health assessment model parameters, an enhanced adaptive incremental learning algorithm is regularly executed to update the facility health assessment model parameters. After each update, the performance difference between the new and old facility health assessment models is verified through A / B testing, and the facility health assessment model parameters are adjusted according to actual effect feedback to obtain an adaptive facility health assessment model.

[0022] Optionally, according to the performance score, expected failure probability and recommended maintenance cycle in the facility health assessment report, a feedback analysis model is constructed using a decision tree and a random forest algorithm, and the feedback analysis model is used to identify the factors that have the greatest impact on the health status of the facility, and a list of key influencing factors is obtained, including:

[0023] Using online learning technology, each new data stream received is processed one by one, and the micro-batch gradient descent algorithm is used to perform preliminary update processing on the parameters of the facility health assessment model to obtain the updated parameters of the facility health assessment model;

[0024] Based on the updated parameters of the facility health assessment model, the sliding window method is applied to retain the data stream within a preset time as training samples to generate a training sample set;

[0025] Combined with the training sample set, the transfer learning mechanism in the enhanced adaptive incremental learning algorithm is used to automatically select the most relevant experience for transfer, adjust the parameters of the facility health assessment model, and obtain the transfer-adjusted parameters of the facility health assessment model;

[0026] Based on the transfer-adjusted parameters of the facility health assessment model, an anomaly detection module is introduced to perform a stability check on the updated parameters of the facility health assessment model, and the isolation forest algorithm is used to identify abnormal parameter values to obtain the real-time updated parameters of the facility health assessment model after anomaly detection.

[0027] Optionally, based on the transfer-adjusted parameters of the facility health assessment model, an anomaly detection module is introduced to perform a stability check on the updated parameters of the facility health assessment model, and the isolation forest algorithm is used to identify abnormal parameter values to obtain the real-time updated parameters of the facility health assessment model after anomaly detection, including:

[0028] Based on the transfer-adjusted parameters of the facility health assessment model, a multi-level anomaly detection framework is constructed, and the multi-level anomaly detection framework is used to perform a preliminary screening on the updated parameters of the facility health assessment model to obtain a preliminary list of abnormal parameters;

[0029] According to the preliminary list of abnormal parameters, the context awareness mechanism is applied to calculate the historical data of facility operation and the current environmental conditions, and the abnormal parameter values are verified twice to generate a list of abnormal parameters;

[0030] According to the list of abnormal parameters, a rollback mechanism is triggered to restore to the parameters of the facility health assessment model in the previous stable version, and the specific conditions of the abnormal parameters are recorded to form an anomaly log;

[0031] Combined with the anomaly log, the enhanced adaptive incremental learning algorithm is executed regularly to re-evaluate and adjust the parameters of the facility health assessment model. After each update, the performance difference between the new and old facility health assessment models is verified through A / B testing, and the model parameters are adjusted according to the actual effect feedback. Finally, the real-time updated parameters of the facility health assessment model after anomaly detection are obtained.

[0032] Optionally, based on the performance score, expected failure probability, and recommended maintenance cycle in the facility health assessment report, combined with user requirement feedback, construct a service response pattern library, and use genetic algorithms and reinforcement learning techniques to optimize and match the service patterns in the service response pattern library to generate an optimal service response plan. The service response pattern library includes: historical data, best practices, and industry standards, including:

[0033] Based on the performance score, expected failure probability, and recommended maintenance cycle in the facility health assessment report, combined with user requirement feedback, sort out and summarize the existing service patterns to form a service response pattern library containing historical data, best practices, and industry standards;

[0034] Use genetic algorithms to preliminarily screen the service patterns in the service response pattern library to generate a set of candidate service response patterns;

[0035] According to the candidate service response patterns, introduce reinforcement learning techniques, simulate the execution of each service pattern and evaluate the effects of each service pattern, and select the optimal solution among the service patterns as the preliminary optimal service response plan;

[0036] Combined with the preliminary optimal service response plan, apply scenario analysis and Monte Carlo simulation techniques to predict the service response effects under different scenarios, and generate a service response plan evaluation report;

[0037] Based on the service response plan evaluation report, determine the optimal service response plan and provide specific operation guidelines for the property management team.

[0038] Optionally, based on the optimized task assignment and user notification, use big data analysis techniques and machine learning algorithms to monitor and evaluate the effects of the optimal service response plan, collect actual effect feedback, and adjust the weight allocation and decision rules in the adaptive facility health assessment model and the optimal service response plan according to the search for actual effect feedback to form a closed-loop optimization mechanism, including:

[0039] Based on the optimized task assignment and user notification, use big data analysis techniques to monitor the actual execution of service responses in real time and collect actual effect feedback data from the property management team and users;

[0040] According to the actual effect feedback data, apply machine learning algorithms to construct an effect evaluation model to comprehensively evaluate the effects of the optimal service response plan and generate an effect evaluation report;

[0041] Combined with the effect evaluation report, use the A / B test method to compare different versions of the service response plan, verify the performance differences between the old and new service response plans, and obtain a performance comparison result;

[0042] Based on the performance comparison results, an incremental learning algorithm is introduced to dynamically adjust the weight allocation and decision rules in the adaptive facility health assessment model and the best service response plan, and the adjusted weight allocation and decision rules are generated.

[0043] Using the adjusted weight allocation and decision rules, the adaptive facility health assessment model and the service response pattern library are updated regularly. After each update, the effectiveness of the update is verified through a backtesting mechanism, and adjustments are made according to the verification results to form a closed-loop optimization mechanism.

[0044] In a second aspect, an embodiment of the present invention provides a system, including:

[0045] A compression module for initially compressing data and detecting anomalies in the real-time operating status information and user demand feedback from multiple property facilities using edge computing technology to obtain filtered data.

[0046] An evaluation module for constructing a facility health assessment model by combining Bayesian networks and time series analysis based on the filtered data and historical maintenance records, and using the facility health assessment model to dynamically evaluate the facility health status and predict potential failures, generating a facility health assessment report. At the same time, the parameters of the facility health assessment model are updated in real time using an incremental learning algorithm to optimize the facility health assessment model to adapt to the latest situation, obtaining an adaptive facility health assessment model.

[0047] A construction module for constructing a service response pattern library based on the performance score, expected failure probability, and recommended maintenance cycle in the facility health assessment report, combined with user demand feedback, and optimizing and matching the service patterns in the service response pattern library using genetic algorithms and reinforcement learning techniques to generate the best service response plan. The service response pattern library includes historical data, best practices, and industry standards.

[0048] A sending module for automatically generating a task work order using natural language generation technology and geographic information systems, providing operation guidelines for the property management team, and sending notifications of the estimated completion time and service content to relevant users through a message queue mechanism to obtain optimized task allocation and user notifications.

[0049] An adjustment module, configured to monitor and evaluate the effectiveness of the optimal service response plan by using big data analysis technology and machine learning algorithms based on the optimized task allocation and user notifications, collect actual effect feedback, and adjust the weight allocation and decision rules in the adaptive facility health assessment model and the optimal service response plan according to the searched actual effect feedback, so as to form a closed-loop optimization mechanism. In a third aspect, an embodiment of the present invention provides a computing device, including a processor and a memory, where a computer program is stored in the memory, and the processor is configured to run the computer program to execute the method for intelligent management and allocation of property based on big data according to any one of the first aspects.

[0050] In a fourth aspect, an embodiment of the present invention provides a computer storage medium, on which computer program instructions are stored, and when the computer program instructions are executed by a processor, the method for intelligent management and allocation of property based on big data according to any one of the first aspects is implemented.

[0051] In the embodiments of the present invention, edge computing technology is used to perform preliminary data compression and anomaly detection on real-time operation status information and user demand feedback from multiple property facilities to obtain screened data. According to the screened data and historical maintenance records, a Bayesian network and time series analysis are combined to construct a facility health assessment model, and the facility health assessment model is used to dynamically evaluate the facility health status and predict potential failures, generating a facility health assessment report. At the same time, an incremental learning algorithm is used to update the parameters of the facility health assessment model in real time, optimizing the facility health assessment model to adapt to the latest situation, obtaining an adaptive facility health assessment model. Based on the performance score, expected failure probability, and recommended maintenance cycle in the facility health assessment report, combined with user demand feedback, a service response pattern library is constructed, and genetic algorithm and reinforcement learning technology are used to optimize and match the service patterns in the service response pattern library to generate an optimal service response plan. The service response pattern library includes: historical data, best practices, and industry standards. Using natural language generation technology and geographic information system, a task work order is automatically generated according to the optimal service response plan, and operation guidelines are provided for the property management team. At the same time, the expected completion time and service content notice are sent to relevant users through the message queue mechanism to obtain optimized task allocation and user notification. Based on the optimized task allocation and user notification, big data analysis technology and machine learning algorithms are used to monitor and evaluate the effect of the optimal service response plan, collect actual effect feedback, and adjust the weight allocation and decision rules in the adaptive facility health assessment model and the optimal service response plan according to the search for actual effect feedback, forming a closed-loop optimization mechanism. The technical solution provided by the present invention uses edge computing technology to perform preliminary data compression and anomaly detection on real-time operation status information and user demand feedback from multiple property facilities, effectively reducing the data transmission volume, improving the system response speed, and ensuring that abnormal situations can be detected and processed in a timely manner. Combining Bayesian network and time series analysis to construct a facility health assessment model can not only dynamically evaluate the facility health status but also predict potential failures. This helps to take preventive maintenance measures in advance, reduce losses caused by sudden failures, and improve the reliability and service life of facilities. Based on the performance score, expected failure probability, and recommended maintenance cycle in the facility health assessment report, combined with user demand feedback, a service response pattern library is constructed, and genetic algorithm and reinforcement learning technology are used to optimize and match the optimal service response plan.This solution fully considers the specific conditions of the facilities and the personalized needs of users, achieving the optimal allocation of resources and maximizing service efficiency. It uses natural language generation technology and geographic information systems to automatically generate task work orders according to the optimal service response plan and provides detailed operation guidelines for the property management team. At the same time, through the message queue mechanism, it sends notifications of the estimated completion time and service content to relevant users to ensure the immediacy and accuracy of information transmission, improving the user experience and service satisfaction. Based on the optimized task allocation and user notifications, it uses big data analysis technology and machine learning algorithms to monitor and evaluate the effectiveness of the optimal service response plan, collects actual effect feedback, and adjusts the weight allocation and decision rules in the adaptive facility health assessment model and the optimal service response plan according to the feedback, forming a closed-loop optimization mechanism. This method ensures that the system can continuously self-evolve, adapt to new situations, and further improve management efficiency and service quality;

[0052] Furthermore, a method combining Bayesian networks and time series analysis is used to perform multi-dimensional feature fusion processing on the filtered data received by the cloud data center and historical maintenance records, obtaining a concise and information-rich high-dimensional feature representation. At the same time, an autoencoder in the deep learning framework is used to reduce the dimension and extract features of the filtered data, generating a concise and information-rich feature set. This method not only improves the quality of the features but also simplifies the data input for subsequent model training, enhancing the accuracy and generalization ability of the model; Based on the concise and information-rich feature set, a deep neural network (DNN) and a generative adversarial network (GAN) are used to construct and train the facility health assessment model, obtaining an initial facility health assessment model. By introducing GAN, simulating the competition mechanism between real and virtual data, the robustness and generalization ability of the model are enhanced, enabling it to better handle complex and changing actual environments; The initial facility health assessment model is used to dynamically evaluate the facility health status and predict potential failures, generating a facility health assessment report containing the performance scores of each facility, the expected failure probability, and the recommended maintenance cycle. This process uses Monte Carlo simulation technology to estimate uncertainties and provides multiple possible development paths through scenario analysis, providing more reference information for decision-makers to help them make more scientific and reasonable decisions; According to the facility health assessment report, an adaptive incremental learning algorithm is designed and applied to update the parameters of the facility health assessment model in real time, finally obtaining an adaptive facility health assessment model. This algorithm can dynamically adjust the model parameters without retraining the entire model, ensuring that the model always maintains the latest state and can quickly adapt to new operating environments and technological advancements, thereby improving the real-time performance and prediction accuracy of the model.

[0053] These aspects or other aspects of the present invention will be more clearly understood in the following description of the embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can also be obtained based on these drawings.

[0055] Figure 1 It is a flowchart of a property intelligent management and allocation method based on big data provided by an embodiment of the present invention;

[0056] Figure 2 It is a schematic structural diagram of a property intelligent management and allocation system based on big data provided by an embodiment of the present invention;

[0057] Figure 3 It is a schematic structural diagram of a computing device provided by an embodiment of the present invention. Detailed implementation manners

[0058] In order to enable those skilled in the art of the present technology to better understand the solutions of the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention.

[0059] In some processes described in the specification and claims of the present invention and the above accompanying drawings, there are multiple operations that appear in a specific order. However, it should be clearly understood that these operations can be executed not in the order in which they appear in this article or in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations can be executed in sequence or in parallel. It should be noted that the descriptions such as "first" and "second" in this article are used to distinguish different messages, devices, modules, etc., and do not represent a sequence, nor do they limit that "first" and "second" are of different types.

[0060] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative efforts belong to the scope of protection of the present invention.

[0061] Figure 1 An embodiment of the present invention provides a flowchart of a property intelligent management and allocation method based on big data, as Figure 1 shown, the method includes:

[0062] Step 101: Use edge computing technology to perform preliminary data compression and anomaly detection on the real-time operation status information and user demand feedback from multiple property facilities, and obtain the filtered data;

[0063] In this step, edge computing technology refers to the technology of processing data at the network edge close to the data source, aiming to reduce data transmission latency and bandwidth usage; real-time operation status information refers to various status data generated during the operation of property facilities (such as elevators, air conditioning systems, etc.), such as temperature, humidity, vibration frequency, etc.; user demand feedback refers to the opinions and suggestions on the facility usage experience provided by users through applications or directly to the property management team; preliminary data compression and anomaly detection refer to the preprocessing of a large amount of collected raw data, including removing redundant information, identifying and marking abnormal data points; filtered data refers to the valid and reliable data retained after compression and anomaly detection.

[0064] Collect the real-time operation status information and user demand feedback from multiple property facilities, use edge computing devices to perform preliminary processing on this data on-site, remove unnecessary redundant information, detect possible abnormal situations, and synchronize the valid data after compression and anomaly detection to the cloud data center to ensure high-quality and timely updated data for subsequent analysis.

[0065] Step 102: According to the filtered data and historical maintenance records, combine Bayesian network and time series analysis to construct a facility health assessment model, and use the facility health assessment model to dynamically evaluate the facility health status and predict potential failures, generate a facility health assessment report, and at the same time use the incremental learning algorithm to update the parameters of the facility health assessment model in real time, optimize the facility health assessment model to adapt to the latest situation, and obtain an adaptive facility health assessment model;

[0066] In this step, Bayesian network refers to a probabilistic graphical model used to represent the dependence relationship between variables and support uncertainty reasoning; time series analysis refers to the method of studying data sequences that change over time, often used to predict future trends; facility health assessment model refers to a mathematical model constructed based on historical maintenance records and real-time operation status information, used to evaluate the health status of facilities; dynamic assessment and potential failure prediction refer to real-time monitoring of facility status and predicting possible future problems; incremental learning algorithm refers to a machine learning method that allows the model to gradually update its parameters without retraining the entire dataset.

[0067] Adaptive Facility Health Assessment Model: A facility health assessment model that can self-adjust and optimize according to the latest situation. By combining the screened data and historical maintenance records, a facility health assessment model is constructed using a method that combines Bayesian networks and time series analysis. This model is used to dynamically assess the health status of the facility, predict the probability of potential failures, generate a facility health assessment report containing performance scores, expected failure probabilities, and recommended maintenance cycles, and use incremental learning algorithms to update the model parameters in real time to adapt to the latest situation, thereby continuously optimizing the facility health assessment model.

[0068] Step 103: Based on the performance scores, expected failure probabilities, and recommended maintenance cycles in the facility health assessment report, combined with user requirement feedback, construct a service response pattern library, and use genetic algorithms and reinforcement learning techniques to optimize and match the service patterns in the service response pattern library to generate the best service response plan. The service response pattern library includes: historical data, best practices, and industry standards.

[0069] In this step, the service response pattern library refers to a database containing various possible service response plans, covering historical data, best practices, and industry standards; performance scores, expected failure probabilities, and recommended maintenance cycles refer to the key indicators in the facility health assessment report, which are used to guide the selection of service response plans; genetic algorithms refer to a search algorithm that simulates the process of natural selection and is used to find the optimal solution; reinforcement learning techniques refer to a machine learning method that enables an agent to learn how to take actions to maximize cumulative rewards through a trial-and-error mechanism.

[0070] Based on the performance scores, expected failure probabilities, and recommended maintenance cycles in the facility health assessment report, combined with user requirement feedback, construct a service response pattern library, apply genetic algorithms and reinforcement learning techniques to optimize and match the service patterns in the service response pattern library to generate the best service response plan, ensuring that the generated service response plan not only considers the specific conditions of the facility but also fully meets the personalized needs of users, achieving the best allocation of resources and maximizing service efficiency.

[0071] Step 104: Use natural language generation technology and geographic information systems to automatically generate task work orders according to the best service response plan, provide operation guidelines for the property management team, and at the same time send notifications of the estimated completion time and service content to relevant users through a message queue mechanism to obtain optimized task allocation and user notifications.

[0072] In this step, natural language generation technology refers to the technology of converting structured data into natural language text, enabling the computer to generate human-readable information; Geographic Information System (GIS) refers to a system used to capture, store, manipulate, analyze, manage, and display all types of geographic data; A task work order refers to a document that details the tasks to be performed and their related details, usually automatically generated by the system and assigned to the corresponding staff; The message queue mechanism refers to an asynchronous communication method that allows messages to be passed between different components without waiting for an immediate response from the other party;

[0073] According to the optimal service response plan, use natural language generation technology and Geographic Information System to automatically generate task work orders, providing detailed written operation guidelines for the property management team to ensure that they clearly understand the specific requirements of each task. Send notifications of the estimated completion time and service content to relevant users through the message queue mechanism to ensure the immediacy and accuracy of information transmission, and enhance the user experience and service satisfaction.

[0074] Step 105: Based on the optimized task assignment and user notification, use big data analysis technology and machine learning algorithms to monitor and evaluate the effectiveness of the optimal service response plan, collect actual effect feedback, and adjust the weight assignment and decision rules in the adaptive facility health assessment model and the optimal service response plan according to the search for actual effect feedback, forming a closed-loop optimization mechanism;

[0075] In this step, big data analysis technology refers to the technology of processing and analyzing massive data sets to extract valuable information and insights; Machine learning algorithms refer to a class of algorithms that can automatically improve through data and can be used for various tasks such as classification, regression, and clustering; Actual effect feedback refers to the data obtained from the actual execution results, reflecting the effectiveness of the service response plan; Weight assignment and decision rules refer to the key factors that determine how the model makes predictions and service responses; The closed-loop optimization mechanism refers to a continuous improvement process that optimizes performance by continuously collecting feedback and adjusting system parameters;

[0076] Based on the optimized task assignment and user notification, use big data analysis technology and machine learning algorithms to monitor and evaluate the actual effectiveness of the optimal service response plan, collect actual effect feedback, analyze the performance of the service response plan, and adjust the weight assignment and decision rules in the adaptive facility health assessment model and the optimal service response plan according to these feedbacks, forming a closed-loop optimization mechanism to ensure that the system can continuously self-evolve, adapt to new situations, and further improve management efficiency and service quality;

[0077] Suppose that in a modern property management system, in order to achieve intelligent management of multiple property facilities, a property intelligent management and allocation method based on big data is adopted. This method first collects real-time operation status information and user demand feedback from each property facility through edge computing technology. These data are preliminarily processed near the data source, including data compression and anomaly detection, to remove redundant information and mark possible abnormal situations. After processing, the filtered valid data is synchronized to the cloud data center through a secure encrypted channel to ensure the security and integrity of data transmission;

[0078] Next, based on the filtered data transmitted from the edge devices and historical maintenance records, we use a method combining Bayesian network and time series analysis to construct a facility health assessment model. This model can not only dynamically evaluate the current health status of the facilities, but also predict the occurrence probability of potential failures. The generated facility health assessment report includes key indicators such as the performance scores of each facility, the expected failure probability, and the recommended maintenance cycle. At the same time, in order to make the model adapt to the latest situation changes, an incremental learning algorithm is introduced to realize real-time update of the facility health assessment model parameters, so as to optimize the model to cope with the latest situation, and finally an adaptive facility health assessment model is obtained;

[0079] Based on the performance scores, expected failure probabilities, and recommended maintenance cycles in the above-mentioned facility health assessment report, and combined with the user's demand feedback, a service response pattern library is constructed. This library integrates various resources such as historical data, best practices, and industry standards. To find the most practical service response plan, genetic algorithm and reinforcement learning technology are applied to optimize and match the service patterns in the service response pattern library. The generated best service response plan not only considers the specific conditions of the facilities, but also fully meets the personalized needs of users, achieving the best allocation of resources and the maximization of service efficiency;

[0080] In order to transform the best service response plan into actual operations, natural language generation technology and Geographic Information System (GIS) are adopted. According to the generated best service response plan, task work orders are automatically generated and detailed operation guidelines are provided for the property management team. At the same time, notifications of the estimated completion time and specific service content are sent to relevant users through the message queue mechanism to ensure the immediacy and accuracy of information transmission, improving the user experience and service satisfaction;

[0081] Finally, to continuously improve the entire management process, a closed-loop optimization mechanism is established. Based on the optimized task allocation and user notifications, big data analysis techniques and machine learning algorithms are used to monitor and evaluate the effectiveness of the best service response plan. By collecting actual effect feedback, the weight allocation and decision rules in the adaptive facility health assessment model and the best service response plan are continuously adjusted, enabling the model to self-evolve, adapt to new environmental changes and technological advancements, and further improve management efficiency and service quality.

[0082] The property intelligent management and deployment method based on big data not only improves the efficiency of facility management and maintenance, but also significantly enhances the user experience, bringing a revolutionary change to property management.

[0083] Based on this, the present invention provides a specific embodiment. In step 102, based on the screened data and historical maintenance records, a Bayesian network and time series analysis are combined to construct a facility health assessment model, and the facility health assessment model is used to dynamically evaluate the facility health status and predict potential failures, generating a facility health assessment report. At the same time, an incremental learning algorithm is used to update the parameters of the facility health assessment model in real time, optimize the facility health assessment model to adapt to the latest situation, and obtain an adaptive facility health assessment model, which specifically includes the following steps:

[0084] Step 201: Use a method that combines Bayesian network and time series analysis to perform multi-dimensional feature fusion processing on the screened data and historical maintenance records received by the cloud data center, obtaining a concise and information-rich high-dimensional feature representation. At the same time, an autoencoder in the deep learning framework is used to reduce the dimension and extract features of the screened data, generating a concise and information-rich feature set.

[0085] In this step, a Bayesian network refers to a probabilistic graphical model used to represent the dependence relationship between variables and support uncertainty reasoning; time series analysis refers to a method for studying data sequences that change over time, often used to predict future trends; a cloud data center refers to a server cluster that centrally stores and processes a large amount of data and usually provides services through the Internet; the screened data and historical maintenance records refer to the effective data after preliminary processing and past maintenance information; multi-dimensional feature fusion processing refers to combining features from different sources or types into a comprehensive feature set to capture more comprehensive information.

[0086] High-dimensional feature representation refers to a data representation with multiple feature dimensions that can more precisely describe object attributes; an autoencoder refers to a deep learning model used for feature extraction and data compression in unsupervised learning; dimension reduction and feature extraction: reducing the data dimension while retaining important information, generating a concise and information-rich feature set.

[0087] By using a method that combines Bayesian network and time series analysis, multi-dimensional feature fusion processing is performed on the screened data received by the cloud data center and historical maintenance records to obtain a concise and information-rich high-dimensional feature representation. An autoencoder in the deep learning framework is used to reduce the dimension and extract features from the screened data, generating a concise and information-rich feature set, which not only reduces the data complexity but also enhances the quality of the features, providing a solid foundation for subsequent modeling.

[0088] Step 202: Based on the concise and information-rich feature set, use a deep neural network and a generative adversarial network to construct and train a facility health assessment model to obtain an initial facility health assessment model.

[0089] In this step, the concise and information-rich feature set refers to the high-quality feature set generated in Step 201; the deep neural network (DNN) refers to a multi-layer neural network structure that can handle complex non-linear mapping problems; the generative adversarial network (GAN) refers to a generative adversarial model that includes a generator and a discriminator, and improves the quality of the generated data through the game between the two; the initial facility health assessment model refers to the first version of the facility health assessment model trained based on existing data.

[0090] Based on the concise and information-rich feature set, use a deep neural network and a generative adversarial network to construct and train a facility health assessment model. The deep neural network is used to capture the complex relationships between features, while the generative adversarial network simulates the competition mechanism between real and virtual data, enhancing the robustness and generalization ability of the model. Finally, an initial facility health assessment model is obtained, which can accurately predict the health status of the facility when given an input.

[0091] Step 203: Use the facility health assessment model to perform dynamic assessment and potential fault prediction processing on the facility health status, and generate a facility health assessment report including the performance scores of each facility, the expected fault probability, and the recommended maintenance cycle.

[0092] In this step, potential fault prediction refers to predicting possible future problems based on the current state; the facility health assessment report refers to a document that details the performance scores of the facility, the expected fault probability, and the recommended maintenance cycle.

[0093] Use the facility health assessment model to perform dynamic assessment and potential fault prediction processing on the facility health status, and generate a facility health assessment report including the performance scores of each facility, the expected fault probability, and the recommended maintenance cycle. The Monte Carlo simulation technique is used to estimate the uncertainty, and multiple possible development paths are provided through scenario analysis to help decision-makers better understand the current state and future trends of the facility.

[0094] Step 204: According to the facility health assessment report, design and apply an adaptive incremental learning algorithm to perform real-time update processing on the parameters of the facility health assessment model, and obtain an adaptive facility health assessment model;

[0095] In this step, the adaptive incremental learning algorithm refers to a machine learning method that allows the model to gradually update its parameters without retraining the entire dataset; the parameters of the facility health assessment model refer to the key factors that determine how the model makes predictions and service responses; the adaptive facility health assessment model: a facility health assessment model that can self-adjust and optimize according to the latest situation; according to the facility health assessment report, design and apply an adaptive incremental learning algorithm to perform real-time update processing on the parameters of the facility health assessment model;

[0096] The adaptive incremental learning algorithm can dynamically adjust the model parameters when new data arrives, ensuring that the model always maintains the latest state, adapts to the new operating environment and technological progress, thereby improving the real-time performance and prediction accuracy of the model, and finally obtaining an adaptive facility health assessment model, ensuring that the model can be continuously optimized to cope with changing situations.

[0097] Based on this, the present invention provides a specific embodiment. In step 204, according to the facility health assessment report, design and apply an adaptive incremental learning algorithm to perform real-time update processing on the parameters of the facility health assessment model, and obtain an adaptive facility health assessment model, which specifically includes the following steps:

[0098] Step 301: According to the performance score, expected failure probability, and recommended maintenance cycle in the facility health assessment report, construct a feedback analysis model using decision tree and random forest algorithms, and use the feedback analysis model to identify the influencing factors that have the greatest impact on the facility health status, and obtain a list of key influencing factors;

[0099] In this step, the decision tree and random forest algorithms refer to two common machine learning methods used for classification and regression problems. Random forest is an ensemble of multiple decision trees; the feedback analysis model refers to a model used to analyze the information in the facility health assessment report and identify the factors that have the greatest impact on the facility health status; the list of key influencing factors refers to the list of factors that have the greatest impact on the facility health status determined by the feedback analysis model;

[0100] According to the performance score, expected failure probability, and recommended maintenance cycle in the facility health assessment report, construct a feedback analysis model using decision tree and random forest algorithms, and use the feedback analysis model to identify the influencing factors that have the greatest impact on the facility health status, and obtain a list of key influencing factors, which helps to focus on the most important variables and provide a clear direction for subsequent optimization.

[0101] Step 302: Based on the list of key influencing factors, the parameter adjustment strategy of the facility health assessment model is optimized using the Q-learning algorithm in the reinforcement learning framework to obtain an optimized parameter adjustment strategy;

[0102] In this step, the Q-learning algorithm in the reinforcement learning framework refers to a learning method based on reward and punishment mechanism, which is used to find the optimal action strategy; the parameter adjustment strategy refers to the method of determining how to adjust the parameters of the facility health assessment model to optimize the prediction effect; the optimized parameter adjustment strategy refers to the strategy that can more effectively guide the adjustment of model parameters after being optimized by the Q-learning algorithm;

[0103] Based on the list of key influencing factors, the parameter adjustment strategy of the facility health assessment model is optimized using the Q-learning algorithm in the reinforcement learning framework. The effects of different parameter adjustment schemes are explored through the Q-learning algorithm, and the optimal solution is selected as the optimized parameter adjustment strategy. Multi-dimensional indicators such as the cost-effectiveness ratio of different maintenance operations, resource utilization and service quality are considered to ensure that parameter adjustment not only improves the prediction accuracy, but also achieves the optimal configuration of resources.

[0104] Step 303: In combination with the optimized parameter adjustment strategy, a transfer learning mechanism is introduced to enhance the adaptive incremental learning algorithm. By comparing the similarity between the current facility and the historical cases, the most relevant experience is automatically selected for transfer to generate an enhanced adaptive incremental learning algorithm.

[0105] In this step, the transfer learning mechanism refers to a machine learning technique that allows the knowledge learned by the model from one field to be applied to another related field; the adaptive incremental learning algorithm refers to an algorithm that can gradually update the model parameters without retraining the entire data set; the enhanced adaptive incremental learning algorithm refers to the adaptive incremental learning algorithm that is improved by combining the transfer learning mechanism, which enhances its learning and generalization capabilities;

[0106] Combined with the optimized parameter adjustment strategy, a transfer learning mechanism is introduced to enhance the adaptive incremental learning algorithm. By comparing the similarities between current facilities and historical cases, the most relevant experience is automatically selected for migration to generate an enhanced adaptive incremental learning algorithm, which accelerates the convergence speed of the model and improves the prediction accuracy, allowing the model to adapt to new environments and technological advances more quickly.

[0107] Step 304: using online learning technology and a sliding window method, combined with an enhanced adaptive incremental learning algorithm, dynamically updating the facility health assessment model parameters one by one, and obtaining real-time updated facility health assessment model parameters;

[0108] In this step, online learning technology refers to a technology that can process continuously arriving new data and update the model in real time; the sliding window method refers to a method of keeping the latest data within a certain period of time as training samples to ensure that the model reflects the latest operating status; the real-time updated facility health assessment model parameters refer to the latest version of the model parameters obtained through dynamic updates to ensure that the model is always in the optimal state;

[0109] Using online learning technology and sliding window method, combined with enhanced adaptive incremental learning algorithm, the parameters of the facility health assessment model are dynamically updated one by one. Whenever new data arrives, the Mini-batch Gradient Descent algorithm is used to minimize the prediction error to ensure that the model parameters are always kept up to date and optimal. This process retains the data in the recent period as training samples through the sliding window, maintains the model's sensitivity to the latest situation, and finally obtains the real-time updated parameters of the facility health assessment model, ensuring that the model can reflect the latest status of the facility in real time and give early warnings for possible problems in the future.

[0110] Step 305: Based on the real-time updated facility health assessment model parameters, the enhanced adaptive incremental learning algorithm is periodically executed to update the facility health assessment model parameters. After each update, the performance difference between the new and old facility health assessment models is verified through A / B testing, and the facility health assessment model parameters are adjusted according to actual effect feedback to obtain an adaptive facility health assessment model.

[0111] In this step, A / B testing refers to an experimental method used to compare the performance differences between two or more versions (version A and version B) to determine which version is better; actual effect feedback refers to the data obtained from the actual execution results, reflecting the changes in performance before and after the model is updated; adaptive facility health assessment model refers to a facility health assessment model that can self-adjust and optimize according to the latest situation;

[0112] Based on the parameters of the facility health assessment model updated in real time, the enhanced adaptive incremental learning algorithm is periodically executed to update the parameters of the facility health assessment model. This process ensures that the model can continuously adapt to the new operating environment and technological progress. After each update, the performance differences between the old and new facility health assessment models are verified through A / B testing. Specifically, the existing model (version A) and the updated model (version B) are simultaneously applied to a part of the same data set, and the performance indicators such as prediction accuracy, response time, and resource utilization are compared between the two. The user feedback and other actual effect data are analyzed to evaluate the actual impact of the model update, and the parameters of the facility health assessment model are adjusted according to the actual effect feedback. If the updated model shows better performance, the new parameters are formally adopted; otherwise, further fine-tuning or rolling back to the previous version may be required. Finally, an adaptive facility health assessment model is obtained. Through this closed-loop optimization mechanism, it is ensured that the model can not only reflect the latest status of the facility in real time, but also give early warnings of possible problems in the future and continuously self-optimize to adapt to the changing environment.

[0113] Based on this, the present invention provides a specific embodiment. In step 301, according to the performance score, expected failure probability, and recommended maintenance cycle in the facility health assessment report, a feedback analysis model is constructed using decision tree and random forest algorithms, and the influencing factors that have the greatest impact on the facility health status are identified using the feedback analysis model to obtain a list of key influencing factors, which specifically includes the following steps:

[0114] Step 401: Using online learning technology, each new data stream received is processed one by one, and the parameters of the facility health assessment model are initially updated using the mini-batch gradient descent algorithm to obtain the updated parameters of the facility health assessment model;

[0115] In this step, online learning technology refers to a technology that can process continuously arriving new data and update the model in real time; the new data stream refers to continuously generated new data, such as the real-time operating status information of property facilities; the mini-batch gradient descent algorithm refers to an optimization algorithm that updates the model parameters through small batches of data to balance computational efficiency and model accuracy; the updated parameters of the facility health assessment model refer to the new version of the model parameters after the initial update process.

[0116] Using online learning technology, each new data stream received is processed one by one, and the parameters of the facility health assessment model are initially updated using the mini-batch gradient descent algorithm. Whenever new data arrives, the mini-batch gradient descent algorithm is used to minimize the prediction error, ensuring that the model parameters always remain up-to-date and optimal, not only improving the response speed of the model, but also enhancing its adaptability. Finally, the updated parameters of the facility health assessment model are obtained.

[0117] Step 402: Based on the updated parameters of the facility health assessment model, apply the sliding window method to retain the data stream within a preset time as training samples, and generate a training sample set;

[0118] In this step, the sliding window method refers to a method of keeping the latest data within a certain time period as training samples to ensure that the model reflects the latest operating status; the data stream within the preset time refers to the data collected within the time period set according to requirements; the training sample set refers to the data set used to train or update the model;

[0119] Based on the updated parameters of the facility health assessment model, apply the sliding window method to retain the data stream within a preset time as training samples, and generate a training sample set. By setting a reasonable window size and sliding step, maintain the sensitivity of the model to the latest situation, ensure that the model can timely reflect the latest operating status, and provide high-quality data support for subsequent parameter adjustment.

[0120] Step 403: Combine the training sample set, and use the transfer learning mechanism in the enhanced adaptive incremental learning algorithm to automatically select the most relevant experience for transfer, adjust the parameters of the facility health assessment model, and obtain the parameters of the facility health assessment model after transfer adjustment;

[0121] In this step, the training sample set refers to the data set generated in Step 402 for training or updating the model; the enhanced adaptive incremental learning algorithm refers to the adaptive incremental learning algorithm improved by combining the transfer learning mechanism, which enhances its learning ability and generalization ability; the transfer learning mechanism refers to a machine learning technology that allows the knowledge learned by the model in one domain to be applied to another related domain;

[0122] Combine the training sample set, and use the transfer learning mechanism in the enhanced adaptive incremental learning algorithm to automatically select the most relevant experience for transfer, adjust the parameters of the facility health assessment model. By comparing the similarity between the current facility and historical cases, automatically select the most relevant experience for transfer, generate the parameters of the facility health assessment model after transfer adjustment, accelerate the convergence speed of the model, and improve the prediction accuracy, enabling the model to adapt to the new environment and technological progress faster;

[0123]

[0124] Evaluate the model parameters; η is the learning rate, which controls the step size of parameter update; N is the number of samples in the training sample set; w i is the weight of the i-th sample, reflecting its importance for transfer learning; is the gradient of the loss function L with respect to the parameter θ, which is used to indicate the direction and degree of parameter update; x i and y iThey are the feature vector and label of the i-th sample respectively;

[0125] By utilizing the transfer learning mechanism in the enhanced adaptive incremental learning algorithm, the most relevant experiences are automatically selected for transfer, and the parameters of the facility health assessment model are adjusted, solving the problem in the prior art that the model is difficult to quickly adapt to new situations. Especially when the environment changes, traditional static models may not be able to respond to new demands in a timely manner. And by introducing the transfer learning mechanism, the model can automatically select the most relevant experiences for transfer, thus adapting to the new environment faster and improving the prediction accuracy.

[0126] Step 404: Based on the parameters of the facility health assessment model after transfer adjustment, an anomaly detection module is introduced to check the stability of the updated parameters of the facility health assessment model. The Isolation Forest algorithm is used to identify the abnormal parameter values, and the parameters of the real-time updated facility health assessment model after anomaly detection are obtained;

[0127] In this step, the anomaly detection module refers to a tool for identifying outliers or abnormal patterns in data; the Isolation Forest algorithm refers to an unsupervised learning method based on decision trees for identifying outliers; the parameters of the real-time updated facility health assessment model after anomaly detection refer to the new version of model parameters that are confirmed to be stable after anomaly detection and stability check;

[0128] Based on the parameters of the facility health assessment model after transfer adjustment, an anomaly detection module is introduced to check the stability of the updated parameters of the facility health assessment model. The Isolation Forest algorithm is used to identify the abnormal parameter values to ensure the stability and reliability of the model parameters during the update process. If abnormal parameter values are found, a rollback mechanism is triggered to restore to the model parameters of the previous stable version.

[0129] Finally, the parameters of the real-time updated facility health assessment model after anomaly detection are obtained, ensuring that the model is always in the optimal state while maintaining its stability and reliability.

[0130] Based on this, the present invention provides a specific embodiment. In step 404, based on the parameters of the facility health assessment model after transfer adjustment, an anomaly detection module is introduced to check the stability of the updated parameters of the facility health assessment model. The Isolation Forest algorithm is used to identify the abnormal parameter values, and the parameters of the real-time updated facility health assessment model after anomaly detection are obtained, which specifically includes the following steps:

[0131] Step 501: Based on the parameters of the facility health assessment model after transfer adjustment, a multi-level anomaly detection framework is constructed, and the updated parameters of the facility health assessment model are preliminarily screened using the multi-level anomaly detection framework to obtain a preliminary list of abnormal parameters;

[0132] In this step, the multi-level anomaly detection framework refers to a comprehensive framework that combines multiple anomaly detection techniques (such as Isolation Forest, LOF, One-Class SVM) and is used to identify abnormal parameter values; the preliminary abnormal parameter list refers to the potential abnormal parameter values preliminarily screened by the multi-level anomaly detection framework.

[0133] Based on the parameters of the facility health assessment model after migration adjustment, a multi-level anomaly detection framework is constructed, and this framework is used to preliminarily screen the updated parameters of the facility health assessment model. The multi-level anomaly detection framework combines the Isolation Forest algorithm and multiple other anomaly detection techniques (such as LOF, One-Class SVM) to improve the accuracy and reliability of anomaly detection, identify possible abnormal parameter values, generate a preliminary abnormal parameter list, and provide a basis for subsequent verification.

[0134] Step 502: According to the preliminary abnormal parameter list, apply the context-aware mechanism to calculate the historical data of facility operation and the current environmental conditions, and perform a secondary verification on the abnormal parameter values to generate an abnormal parameter list.

[0135] In this step, the preliminary abnormal parameter list refers to the list of potential abnormal parameter values generated in Step 501; the context-aware mechanism is used to consider the historical data of facility operation and the current environmental conditions to more comprehensively understand the true situation of the abnormal parameter values; the abnormal parameter list refers to the list of abnormal parameter values confirmed after secondary verification.

[0136] According to the preliminary abnormal parameter list, apply the context-aware mechanism to calculate the historical data of facility operation and the current environmental conditions, and perform a secondary verification on the abnormal parameter values. The context-aware mechanism takes into account the historical operation mode of the facility and the current environmental conditions to ensure that the abnormal parameter values are not false alarms caused by normal fluctuations. After secondary verification, a refined and reliable abnormal parameter list is generated to provide accurate information for subsequent processing.

[0137] Step 503: According to the abnormal parameter list, trigger the rollback mechanism to restore to the parameters of the previous stable version of the facility health assessment model, and at the same time record the specific situation of the abnormal parameters to form an anomaly log.

[0138] In this step, the rollback mechanism refers to restoring to the model parameters of the previous stable version when an anomaly is detected to ensure the stability and reliability of the system; the anomaly log refers to the log file that records the specific situation of the abnormal parameters and their processing process.

[0139] According to the abnormal parameter list, trigger the rollback mechanism to restore the facility health assessment model parameters to the previous stable version. At the same time, record the specific conditions of the abnormal parameters to form an abnormal log, which not only helps to trace the root cause of the problem but also provides a reference basis for future optimization, ensuring that the system can quickly resume normal operation when encountering abnormal situations.

[0140] Step 504: Combine the abnormal log and regularly execute the enhanced adaptive incremental learning algorithm to re-evaluate and adjust the facility health assessment model parameters. After each update, verify the performance difference between the new and old facility health assessment models through A / B testing, and adjust the model parameters according to the actual effect feedback, finally obtaining the real-time updated facility health assessment model parameters after anomaly detection.

[0141] In this step, combine the abnormal log and regularly execute the enhanced adaptive incremental learning algorithm to re-evaluate and adjust the facility health assessment model parameters. After each update, verify the performance difference between the new and old facility health assessment models through A / B testing, and adjust the model parameters according to the actual effect feedback. This method ensures that the model can be continuously optimized in a changing environment, and at the same time, through a strict anomaly detection mechanism, it guarantees the stability and reliability of the model. Finally, the real-time updated facility health assessment model parameters after anomaly detection are obtained, ensuring that the model is always in the optimal state while maintaining its stability and reliability.

[0142] Based on this, the present invention provides a specific embodiment. In step 103, based on the performance score, expected failure probability, and recommended maintenance cycle in the facility health assessment report, combined with user demand feedback, construct a service response pattern library, and use genetic algorithms and reinforcement learning techniques to optimize and match the service patterns in the service response pattern library to generate the best service response plan. The service response pattern library includes: historical data, best practices, and industry standards, and specifically includes the following steps:

[0143] Step 601: Based on the performance score, expected failure probability, and recommended maintenance cycle in the facility health assessment report, combined with user demand feedback, sort out and summarize the existing service patterns to form a service response pattern library containing historical data, best practices, and industry standards.

[0144] In this step, based on the performance score, expected failure probability, and recommended maintenance cycle in the facility health assessment report, combined with user demand feedback, sort out and summarize the existing service patterns to form a service response pattern library containing historical data, best practices, and industry standards, ensuring that the selection of service patterns is not only based on the specific conditions of the facility but also fully considers the personalized needs of users and the standard practices of the service industry, providing a solid foundation for subsequent optimization.

[0145] Step 602: Use a genetic algorithm to preliminarily screen the service patterns in the service response pattern library to generate a set of candidate service response patterns;

[0146] In this step, the candidate service response patterns refer to a set of potential service response solutions generated after preliminary screening;

[0147] Use a genetic algorithm to preliminarily screen the service patterns in the service response pattern library to generate a set of candidate service response patterns. The genetic algorithm explores various possible combinations of service patterns through selection, crossover, and mutation operations to adapt to different facility conditions and user requirements, quickly finding a set of potential candidate solutions from a large number of possible service patterns, providing a basis for further optimization.

[0148] Step 603: According to the candidate service response patterns, introduce reinforcement learning technology, simulate the execution of each service pattern and evaluate the effect of each service pattern, and select the optimal solution among the service patterns as the preliminary best service response plan;

[0149] In this step, the reinforcement learning technology refers to a machine learning method that enables an agent to learn how to take actions to maximize cumulative rewards through a trial-and-error mechanism; the preliminary best service response plan refers to the optimal service pattern selected after simulated execution and effect evaluation;

[0150] According to the candidate service response patterns, introduce reinforcement learning technology, simulate the execution of each service pattern and evaluate its effect. The reinforcement learning technology simulates the actual execution of different service patterns and selects the optimal solution as the preliminary best service response plan based on multi-dimensional indicators such as cost-benefit ratio, resource utilization rate, and service quality, ensuring that the selected plan can achieve the best allocation of resources and maximize service efficiency in practical applications.

[0151] Step 604: Combine the preliminary best service response plan, apply scenario analysis and Monte Carlo simulation techniques to predict the service response effects under different scenarios, and generate a service response plan evaluation report;

[0152] In this step, scenario analysis refers to analyzing different possible future situations by setting different hypothetical conditions; the Monte Carlo simulation technique refers to a statistical method that estimates the probability distribution of results through multiple random samplings; the service response plan evaluation report refers to a document that details the service response effects under different scenarios;

[0153] Combined with the preliminary optimal service response plan, apply scenario analysis and Monte Carlo simulation techniques to predict the service response effects under different scenarios. Scenario analysis considers various possible development paths and their corresponding expected results, while Monte Carlo simulation provides more accurate probability estimates through a large number of random samplings. Finally, generate a detailed service response plan evaluation report to provide more reference information for decision-makers and help them make more scientific and reasonable decisions.

[0154] Step 605: Based on the service response plan evaluation report, determine the optimal service response plan and provide specific operation guidelines for the property management team;

[0155] In this step, based on the service response plan evaluation report, determine the optimal service response plan and provide specific operation guidelines for the property management team. The optimal service response plan not only meets personalized needs but also achieves the optimal allocation of resources. At the same time, it provides detailed written guidance for the property management team to ensure that they clearly understand the specific requirements of each task, ensuring the effective implementation of the service response plan and improving the overall management efficiency and service quality.

[0156] Based on this, the present invention provides a specific embodiment. In step 105, based on the optimized task allocation and user notification, use big data analysis technology and machine learning algorithms to monitor and evaluate the effects of the optimal service response plan, collect actual effect feedback, and adjust the weight allocation and decision rules in the adaptive facility health assessment model and the optimal service response plan according to the search for actual effect feedback to form a closed-loop optimization mechanism, which specifically includes the following steps:

[0157] Step 701: Based on the optimized task allocation and user notification, use big data analysis technology to monitor the actual execution of the service response in real time and collect actual effect feedback data from the property management team and users;

[0158] In this step, big data analysis technology refers to the technology of processing and analyzing massive data sets to extract valuable information and insights; actual effect feedback data refers to the data obtained from the actual execution results, reflecting the effects of the service response plan.

[0159] Based on the optimized task allocation and user notification, use big data analysis technology to monitor the actual execution of the service response in real time and collect actual effect feedback data from the property management team and users, ensuring that the system can timely obtain the latest operation status and service effect information, providing a solid foundation for subsequent evaluation.

[0160] Step 702: According to the actual effect feedback data, apply machine learning algorithms to construct an effect evaluation model, comprehensively evaluate the effects of the optimal service response plan, and generate an effect evaluation report.

[0161] In this step, the machine learning algorithm refers to a class of algorithms that can automatically improve through data and can be used for various tasks such as classification, regression, and clustering; the effect evaluation model refers to a mathematical model used to comprehensively evaluate the effect of the best service response plan; the effect evaluation report refers to a document that details the effect of the best service response plan.

[0162] According to the actual effect feedback data, apply the machine learning algorithm to construct an effect evaluation model to comprehensively evaluate the effect of the best service response plan. This model considers multiple dimensions such as task completion time, user satisfaction, resource utilization rate, etc., generates a detailed effect evaluation report, provides a scientific basis, and helps decision-makers better understand the actual performance of the service response plan.

[0163] Step 703: Combine the effect evaluation report and use the A / B testing method to compare different versions of the service response plan, verify the performance differences between the old and new service response plans, and obtain the performance comparison result.

[0164] In this step, the performance comparison result refers to the performance difference between the old and new service response plans obtained through A / B testing.

[0165] Combine the effect evaluation report and use the A / B testing method to compare different versions of the service response plan, verify the performance differences between the old and new service response plans. A / B testing applies the existing plan (version A) and the updated plan (version B) to a part of the same data set at the same time, compares the performance indicators of the two in terms of prediction accuracy, response time, resource utilization rate, etc., and finally obtains the performance comparison result, ensuring the effectiveness and superiority of the new plan in actual application.

[0166] Step 704: Based on the performance comparison result, introduce the incremental learning algorithm to dynamically adjust the weight allocation and decision rules in the adaptive facility health assessment model and the best service response plan, and generate the adjusted weight allocation and decision rules.

[0167] In this step, based on the performance comparison result, introduce the incremental learning algorithm to dynamically adjust the weight allocation and decision rules in the adaptive facility health assessment model and the best service response plan, ensure that the model and the plan can be continuously optimized to adapt to the new operating environment and technological progress, and finally generate the adjusted weight allocation and decision rules, ensuring that the model and the plan are always in the optimal state, improving the management efficiency and service quality.

[0168] Step 705: Use the adjusted weight allocation and decision rules to regularly update the adaptive facility health assessment model and the service response pattern library. After each update, verify the effectiveness of the update through the backtest mechanism and make adjustments according to the verification results to form a closed-loop optimization mechanism.

[0169] In this step, the adjusted weight assignment and decision rules refer to the new version of weight assignment and decision rules generated by step 704; the backtesting mechanism refers to a way to verify the effectiveness of a model or strategy through historical data; the closed-loop optimization mechanism refers to a continuous improvement process that optimizes performance by continuously collecting feedback and adjusting system parameters;

[0170] Using the adjusted weight assignment and decision rules, the adaptive facility health assessment model and the service response pattern library are updated regularly. After each update, the effectiveness of the update is verified through the backtesting mechanism, and fine-tuning is performed according to the verification results, forming a closed-loop optimization mechanism, which ensures that the system can continuously self-evolve, adapt to new situations, and further improve management efficiency and service quality. In this way, the system can not only reflect the latest status of the facilities in real time, but also give early warnings about possible future problems and continuously optimize itself.

[0171] Figure 2 The following is a schematic structural diagram of a property intelligent management and deployment system based on big data provided by an embodiment of the present invention, as Figure 2 shown, the system includes:

[0172] A compression module 21, configured to perform preliminary data compression and anomaly detection on real-time operation status information from multiple property facilities and user demand feedback using edge computing technology to obtain filtered data;

[0173] An evaluation module 22, configured to construct a facility health assessment model by combining Bayesian network and time series analysis according to the filtered data and historical maintenance records, and use the facility health assessment model to dynamically evaluate the facility health status and predict potential failures, generate a facility health assessment report, and at the same time use the incremental learning algorithm to update the parameters of the facility health assessment model in real time, optimize the facility health assessment model to adapt to the latest situation, and obtain an adaptive facility health assessment model;

[0174] A construction module 23, configured to construct a service response pattern library based on the performance score, expected failure probability, and recommended maintenance cycle in the facility health assessment report, combined with user demand feedback, and optimize and match the service patterns in the service response pattern library using genetic algorithm and reinforcement learning technology to generate an optimal service response plan, where the service response pattern library includes: historical data, best practices, and industry standards;

[0175] A sending module 24, configured to automatically generate a task order using natural language generation technology and geographic information system, provide operation guidelines for the property management team, and at the same time send the estimated completion time and service content notification to relevant users through the message queue mechanism to obtain optimized task allocation and user notification;

[0176] Adjustment module 25, which is used to monitor and evaluate the effect of the optimal service response plan based on the optimized task assignment and user notification by using big data analysis technology and machine learning algorithms, collect actual effect feedback, and adjust the weight assignment and decision rules in the adaptive facility health assessment model and the optimal service response plan according to the searched actual effect feedback, so as to form a closed-loop optimization mechanism.

[0177] Figure 2 The described property intelligent management and allocation system based on big data can execute Figure 1 The described property intelligent management and allocation method based on big data in the illustrated embodiment, the implementation principle and technical effect will not be elaborated. For the property intelligent management and allocation system in the above embodiment, the specific ways for each module and unit to execute operations have been described in detail in the embodiment related to the method, and will not be elaborated here.

[0178] In a possible design, Figure 2 The property intelligent management and allocation system in the illustrated embodiment can be implemented as a computing device, such as Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0179] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32.

[0180] The processing component 32 is used for: performing preliminary data compression and anomaly detection on the real-time operation status information and user demand feedback from multiple property facilities by using edge computing technology to obtain filtered data;

[0181] According to the filtered data and historical maintenance records, a Bayesian network and time series analysis are combined to construct a facility health assessment model, and the facility health assessment model is used to dynamically evaluate the facility health status and predict potential failures, generate a facility health assessment report, and at the same time, the incremental learning algorithm is used to update the facility health assessment model parameters in real time to optimize the facility health assessment model to adapt to the latest situation, so as to obtain an adaptive facility health assessment model;

[0182] Based on the performance score, expected failure probability and recommended maintenance cycle in the facility health assessment report, combined with user demand feedback, a service response pattern library is constructed, and genetic algorithm and reinforcement learning technology are used to optimize and match the service patterns in the service response pattern library to generate an optimal service response plan. The service response pattern library includes: historical data, best practices and industry standards;

[0183] Utilize natural language generation technology and geographic information system to automatically generate task work orders according to the best service response plan, provide operation guidelines for the property management team, and at the same time send notifications of the estimated completion time and service content to relevant users through the message queue mechanism to obtain optimized task allocation and user notifications;

[0184] Based on the optimized task allocation and user notifications, utilize big data analysis technology and machine learning algorithms to monitor and evaluate the effectiveness of the best service response plan, collect actual effect feedback, and adjust the weight allocation and decision rules in the adaptive facility health assessment model and the best service response plan according to the search for actual effect feedback to form a closed-loop optimization mechanism.

[0185] Among them, the processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component may also be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components for executing the above method.

[0186] The storage component 31 is configured to store various types of data to support the operation of the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0187] Of course, the computing device may also necessarily include other components, such as input / output interfaces, display components, communication components, etc.

[0188] The input / output interface provides an interface between the processing component and the peripheral interface module, and the above peripheral interface module can be an output device, an input device, etc.

[0189] The communication component is configured to facilitate communication between the computing device and other devices in a wired or wireless manner, etc.

[0190] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. At this time, the computing device can refer to a cloud server, and the above processing component, storage component, etc. can be basic server resources leased or purchased from the cloud computing platform.

[0191] An embodiment of the present invention also provides a computer storage medium storing a computer program, which when executed by a computer can implement the above-mentioned Figure 1 A property intelligent management and allocation method based on big data in the embodiment shown.

[0192] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the above-described systems, devices, and units can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0193] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative labor.

[0194] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product, which can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., including several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0195] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments or perform equivalent replacements for some of the technical features. These modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for intelligent property management and deployment based on big data, characterized in that: include: Use edge computing technology to perform preliminary data compression and anomaly detection on real-time operating status information and user demand feedback from multiple property facilities to obtain filtered data; Based on the screened data and historical maintenance records, a Bayesian network and time series analysis are combined to construct a facility health assessment model, and the facility health assessment model is used to dynamically assess the health status of the facility and predict potential failures, and a facility health assessment report is generated. At the same time, an incremental learning algorithm is used to update the parameters of the facility health assessment model in real time, and the facility health assessment model is optimized to adapt to the latest situation, thereby obtaining an adaptive facility health assessment model; Based on the performance score, expected failure probability and recommended maintenance cycle in the facility health assessment report, combined with user demand feedback, a service response pattern library is constructed, and the service patterns in the service response pattern library are optimized and matched using genetic algorithms and reinforcement learning techniques to generate the best service response plan, wherein the service response pattern library includes: historical data, best practices and industry standards; Using natural language generation technology and geographic information systems, automatically generate task work orders based on the best service response plan, provide operation guidance to the property management team, and send estimated completion time and service content notifications to relevant users through a message queue mechanism to obtain optimized task allocation and user notifications; Based on the optimized task allocation and user notification, big data analysis technology and machine learning algorithms are used to monitor and evaluate the effectiveness of the best service response plan, collect actual effect feedback, and adjust the weight allocation and decision rules in the adaptive facility health assessment model and the best service response plan according to the search actual effect feedback to form a closed-loop optimization mechanism.

2. The method according to claim 1, characterized in that Based on the screened data and historical maintenance records, a Bayesian network and time series analysis are combined to construct a facility health assessment model, and the facility health assessment model is used to dynamically assess the facility health status and predict potential failures, and a facility health assessment report is generated. At the same time, an incremental learning algorithm is used to update the facility health assessment model parameters in real time, and the facility health assessment model is optimized to adapt to the latest situation, thereby obtaining an adaptive facility health assessment model, including: Using a method combining Bayesian network and time series analysis, multi-dimensional feature fusion processing is performed on the screened data and historical maintenance records received by the cloud data center to obtain a concise and information-rich high-dimensional feature representation. At the same time, an automatic encoder in a deep learning framework is used to reduce the dimension and extract features of the screened data to generate a concise and information-rich feature set; Based on the streamlined and information-rich feature set, a facility health assessment model is constructed and trained using a deep neural network and a generative adversarial network to obtain an initial facility health assessment model; Using the facility health assessment model to dynamically assess the health status of the facility and predict potential failures, a facility health assessment report is generated that includes the performance score of each facility, the expected failure probability, and the recommended maintenance cycle; According to the facility health assessment report, an adaptive incremental learning algorithm is designed and applied to update the parameters of the facility health assessment model in real time to obtain an adaptive facility health assessment model.

3. The method according to claim 2, characterized in that According to the facility health assessment report, an adaptive incremental learning algorithm is designed and applied to update the parameters of the facility health assessment model in real time to obtain an adaptive facility health assessment model, including: Based on the performance score, expected failure probability and recommended maintenance cycle in the facility health assessment report, a feedback analysis model is constructed using a decision tree and a random forest algorithm, and the feedback analysis model is used to identify the factors that have the greatest impact on the health status of the facility, thereby obtaining a list of key influencing factors; Based on the list of key influencing factors, the parameter adjustment strategy of the facility health assessment model is optimized using the Q-learning algorithm in the reinforcement learning framework to obtain an optimized parameter adjustment strategy; Combined with the optimized parameter adjustment strategy, a transfer learning mechanism is introduced to enhance the adaptive incremental learning algorithm. By comparing the similarity between the current facility and the historical case, the most relevant experience is automatically selected for transfer to generate an enhanced adaptive incremental learning algorithm. Using online learning technology and sliding window method, combined with enhanced adaptive incremental learning algorithm, the parameters of the facility health assessment model are dynamically updated one by one to obtain real-time updated parameters of the facility health assessment model. Based on the real-time updated facility health assessment model parameters, an enhanced adaptive incremental learning algorithm is regularly executed to update the facility health assessment model parameters. After each update, the performance difference between the new and old facility health assessment models is verified through A / B testing, and the facility health assessment model parameters are adjusted according to actual effect feedback to obtain an adaptive facility health assessment model.

4. The method according to claim 3, characterized in that According to the performance score, expected failure probability and recommended maintenance cycle in the facility health assessment report, a feedback analysis model is constructed using decision tree and random forest algorithms. The feedback analysis model is used to identify the factors that have the greatest impact on the health status of the facility, and a list of key influencing factors is obtained, including: Using online learning technology, the received new data streams are processed one by one, and the parameters of the facility health assessment model are preliminarily updated using the micro-batch gradient descent algorithm to obtain the updated parameters of the facility health assessment model; Based on the updated facility health assessment model parameters, a sliding window method is applied to retain data streams within a preset time as training samples to generate a training sample set; In combination with the training sample set, the transfer learning mechanism in the enhanced adaptive incremental learning algorithm is used to automatically select the most relevant experience for transfer, adjust the facility health assessment model parameters, and obtain the facility health assessment model parameters after transfer adjustment; Based on the facility health assessment model parameters after the migration and adjustment, an anomaly detection module is introduced to perform stability check on the updated facility health assessment model parameters, and an isolation forest algorithm is used to identify abnormal parameter values ​​to obtain real-time updated facility health assessment model parameters after anomaly detection.

5. The method according to claim 4, characterized in that Based on the facility health assessment model parameters adjusted after migration, an anomaly detection module is introduced to perform stability check on the updated facility health assessment model parameters, and an isolation forest algorithm is used to identify abnormal parameter values ​​to obtain facility health assessment model parameters updated in real time after anomaly detection, including: Based on the migrated and adjusted facility health assessment model parameters, a multi-level anomaly detection framework is constructed, and the updated facility health assessment model parameters are preliminarily screened using the multi-level anomaly detection framework to obtain a preliminary anomaly parameter list; Based on the preliminary abnormal parameter list, a context-aware mechanism is applied to calculate the historical data of the operation of the facility and the current environmental conditions, and the abnormal parameter values ​​are secondary verified to generate an abnormal parameter list; According to the abnormal parameter list, a rollback mechanism is triggered to restore the facility health assessment model parameters to the last stable version, and at the same time, specific conditions of the abnormal parameters are recorded to form an abnormal log; In combination with the anomaly log, an enhanced adaptive incremental learning algorithm is regularly executed to re-evaluate and adjust the parameters of the facility health assessment model. After each update, the performance difference between the new and old facility health assessment models is verified through A / B testing, and the model parameters are adjusted based on actual effect feedback, ultimately obtaining the facility health assessment model parameters that are updated in real time after anomaly detection.

6. The method according to claim 1, characterized in that Based on the performance scores, expected failure probabilities, and recommended maintenance cycles in the facility health assessment report, combined with user demand feedback, a service response pattern library is constructed, and the service patterns in the service response pattern library are optimized and matched using genetic algorithms and reinforcement learning techniques to generate the best service response plan. The service response pattern library contains historical data, best practices, and industry standards, including: Based on the performance scores, expected failure probabilities, and recommended maintenance cycles in the facility health assessment report, combined with user demand feedback, the existing service models are sorted and summarized to form a service response model library that includes historical data, best practices, and industry standards; Using a genetic algorithm to preliminarily screen the service patterns in the service response pattern library to generate a set of candidate service response patterns; According to the candidate service response modes, reinforcement learning technology is introduced to simulate the execution of each service mode and evaluate the effect of each service mode, and the optimal solution in the service mode is selected as the preliminary optimal service response solution; Combined with the preliminary optimal service response plan, scenario analysis and Monte Carlo simulation technology are used to predict the service response effects under different scenarios and generate a service response plan evaluation report; Based on the service response plan evaluation report, determine the best service response plan and provide specific operational guidance to the property management team.

7. The method according to claim 1, characterized in that Based on the optimized task allocation and user notification, the effect of the best service response plan is monitored and evaluated using big data analysis technology and machine learning algorithms, and actual effect feedback is collected. The weight allocation and decision rules in the adaptive facility health assessment model and the best service response plan are adjusted according to the search actual effect feedback to form a closed-loop optimization mechanism, including: Based on the optimized task allocation and user notification, the actual execution of service response is monitored in real time using big data analysis technology, and actual effect feedback data from the property management team and users is collected; Based on the actual effect feedback data, a machine learning algorithm is applied to build an effect evaluation model, comprehensively evaluate the effect of the best service response plan, and generate an effect evaluation report; In combination with the effect evaluation report, an A / B test method is used to compare different versions of service response solutions, verify the performance difference between the new and old service response solutions, and obtain performance comparison results; Based on the performance comparison results, an incremental learning algorithm is introduced to dynamically adjust the weight distribution and decision rules in the adaptive facility health assessment model and the optimal service response solution to generate adjusted weight distribution and decision rules; The adaptive facility health assessment model and service response pattern library are regularly updated using the adjusted weight distribution and decision rules. After each update, the effectiveness of the update is verified through a backtesting mechanism, and adjustments are made based on the verification results to form a closed-loop optimization mechanism.

8. A property intelligent management and deployment system based on big data, characterized in that: include: The compression module is used to perform preliminary data compression and anomaly detection on the real-time operation status information and user demand feedback from multiple property facilities using edge computing technology to obtain filtered data; An evaluation module is used to construct a facility health evaluation model based on the screened data and historical maintenance records by combining Bayesian network and time series analysis, and to dynamically evaluate the health status of the facility and predict potential failures by using the facility health evaluation model, to generate a facility health evaluation report, and to update the parameters of the facility health evaluation model in real time by using an incremental learning algorithm, to optimize the facility health evaluation model to adapt to the latest situation, and to obtain an adaptive facility health evaluation model; A construction module is used to construct a service response pattern library based on the performance score, expected failure probability and recommended maintenance cycle in the facility health assessment report, combined with user demand feedback, and optimize the matching of service patterns in the service response pattern library by using genetic algorithm and reinforcement learning technology to generate the best service response plan, wherein the service response pattern library includes: historical data, best practices and industry standards; A sending module, which is used to automatically generate task work orders according to the optimal service response plan by using natural language generation technology and geographic information systems, and provide operation guidance to the property management team, and send estimated completion time and service content notifications to relevant users through a message queue mechanism to obtain optimized task allocation and user notifications; The adjustment module is used to monitor and evaluate the effect of the best service response plan based on the optimized task allocation and user notification, using big data analysis technology and machine learning algorithms, collect actual effect feedback, and adjust the weight allocation and decision rules in the adaptive facility health assessment model and the best service response plan according to the search actual effect feedback to form a closed-loop optimization mechanism.

9. A computing device, characterized in that It comprises a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a property intelligent management and deployment method based on big data as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that: A computer program is stored, and when the computer program is executed by a computer, a property intelligent management and deployment method based on big data as described in any one of claims 1 to 7 is implemented.

Images