Intelligent agricultural greenhouse control system based on adaptive distributed edge calculation and AI algorithm

Abstract

The invention provides an intelligent agricultural greenhouse control system based on adaptive distributed edge calculation and an AI algorithm, and relates to the technical field of software research and development supervision, and the system comprises an edge calculation node module, a multi-modal data fusion module, a deep learning framework module, a federated learning cooperation module, an intelligent control module, and a data privacy and security module. The edge computing node module is used for carrying out data processing at an equipment end and reducing transmission delay; the multi-modal data fusion module is used for integrating information from different data sources and providing a comprehensive research and development data view; the federated learning collaboration module realizes cross-node data collaboration through local training and parameter sharing, so that the research and development efficiency is improved and the data privacy is protected; the intelligent control module automatically adjusts research and development resource allocation and optimizes a research and development process based on data fusion and deep learning analysis; the data privacy and security module ensures the security and privacy of sensitive data in the edge computing and federal learning process.

Description

Technical Field

[0001] The present invention belongs to the field of software development and supervision technology, and more specifically, relates to a smart agricultural greenhouse control system based on adaptive distributed edge computing and AI algorithms. Background Art

[0002] Software development refers to the process of designing, developing, testing, deploying, and maintaining software applications through systematic methods and processes. Modern software development, especially the development of large, complex software systems, typically requires dividing software projects into multiple development phases or modules for independent development, followed by integration and delivery. The output of each development phase or module impacts the functionality and quality of the final software product, making effective oversight of each development process crucial. This oversight helps identify issues promptly, ensuring quality at every stage of software development and ultimately ensuring that the delivered software product meets predetermined functional and quality standards.

[0003] Currently, common software development monitoring methods rely primarily on spot checks of development artifacts across various development processes, with regular audits and inspections of development results. However, the parameters involved in software development are diverse and can change in real time. Due to their random nature and limitations, spot checks often fail to fully and accurately reflect the quality status of each development link. Spot checks also require significant manpower and resources, and often fail to update monitoring data in a timely manner, resulting in delays in the monitoring process and a lack of real-time monitoring and feedback.

[0004] In addition, the management of software development processes requires the coordination of a large amount of resources such as manpower, funds, data, and equipment, and the resource requirements at different development stages are different. For each stage, the types and quantities of resources required by the R&D team are different, which requires experienced managers to reasonably allocate resources based on actual conditions. However, this resource allocation process often lacks a standardized and quantitative basis, and the judgment and experience of managers will affect the rationality and effectiveness of resources. Therefore, traditional resource allocation methods are prone to lead to resource Unbalanced use of resources may cause some development links to be constrained due to insufficient resources, thus affecting the quality of software delivery.

[0005] To achieve the highest software delivery quality within limited resources, the development process urgently requires a digital management system that uses precise digital methods to optimize resource allocation and improve oversight efficiency. Digital systems can capture project progress data and development quality information in real time. Through data analysis and intelligent decision support, they help managers rationally allocate resources and avoid the drawbacks of traditional oversight methods. This systematic digital oversight system not only optimizes resource utilization and improves the overall quality of software development, but also enables efficient collaboration across development processes, ensuring on-time software delivery and quality assurance. Summary of the Invention

[0006] In order to solve the above technical problems, the present invention provides a smart agricultural greenhouse control system based on adaptive distributed edge computing and AI algorithm to solve the above problems.

[0007] A smart agricultural greenhouse control system and method based on adaptive distributed edge computing and AI algorithms, including: edge computing module, intelligent decision-making module, data monitoring module, environmental control module and resource optimization module; The edge computing module is used to set up edge computing nodes in agricultural greenhouses to collect and process environmental data (such as temperature, humidity, light intensity, etc.) from the agricultural greenhouses in real time, and realize local pre-processing and preliminary analysis of the data, thereby reducing latency and bandwidth burden; The intelligent decision-making module uses AI algorithms to conduct deep learning analysis on the environmental data of agricultural greenhouses, generate accurate environmental control strategies, and interact with the environmental control module through the edge computing module to ensure the real-time and accuracy of environmental control; The data monitoring module is used to continuously monitor various environmental parameters inside the agricultural greenhouse and transmit the data to the cloud platform for centralized management and storage, facilitating long-term analysis and trend prediction; The environmental control module is used to control the temperature and humidity adjustment, irrigation system, lighting system and other facilities in the agricultural greenhouse according to the strategy generated by the intelligent decision-making module to ensure the optimal environment for crop growth in the agricultural greenhouse; The resource optimization module is used to dynamically optimize the resource allocation of agricultural greenhouses. By analyzing the operation status and energy consumption of various environmental control facilities, energy consumption prediction and optimization allocation are carried out to improve resource utilization efficiency.

[0008] Furthermore, the edge computing module includes: a data collection unit and a data processing unit; The data collection unit is used to collect environmental data in the agricultural greenhouse in real time and perform preliminary filtering processing; The data processing unit is used to preprocess, format and perform preliminary analysis on the collected environmental data, and provide processed data for subsequent deep learning analysis.

[0009] Furthermore, the intelligent decision-making module includes: an AI algorithm unit and a decision-making generation unit; The AI ​​algorithm unit is used to analyze the environmental data of agricultural greenhouses using a deep learning algorithm, train a prediction model, and identify key factors affecting crop growth; The decision generation unit formulates a specific environmental control strategy based on the prediction results generated by the AI ​​algorithm, and transmits the strategy to the environmental control module for execution through the edge computing module.

[0010] Furthermore, the data monitoring module includes: an environmental sensor unit and a data transmission unit; The environmental sensor unit is used to continuously monitor environmental parameters such as temperature, humidity, and light inside the agricultural greenhouse; The data transmission unit is used to transmit the monitored environmental data to the cloud platform or edge computing node in real time to ensure the timeliness and integrity of the data.

[0011] Furthermore, the environmental control module includes: a temperature and humidity control unit, an irrigation control unit and a lighting control unit; The temperature and humidity regulating unit is used to control the temperature and humidity in the agricultural greenhouse to ensure a suitable environment for crop growth; The irrigation control unit is used to adjust the irrigation system according to the instructions of the intelligent decision-making module to ensure that the soil moisture is maintained within a suitable range; The lighting adjustment unit is used to adjust the light intensity in the greenhouse, especially in the case of insufficient light, to supplement the light through artificial lighting.

[0012] Furthermore, the resource optimization module includes: an energy consumption monitoring unit, an optimization scheduling unit, and a dynamic adjustment unit; The energy consumption monitoring unit is used to monitor the energy consumption of various environmental control facilities in the greenhouse in real time; The optimization scheduling unit formulates the optimal resource allocation plan based on environmental data analysis and energy consumption; The dynamic adjustment unit is used to adjust resource configuration in real time, reduce energy waste, and improve resource utilization efficiency.

[0013] A smart agricultural greenhouse control method based on adaptive distributed edge computing and AI algorithm includes the following steps: Step S1. Deploy edge computing nodes in the agricultural greenhouse to collect environmental data in real time and perform preliminary processing on the data; Step S2. Use AI algorithms to conduct in-depth analysis of the collected environmental data and generate precise environmental control strategies; Step S3. Adjust the temperature, humidity, irrigation, and lighting systems in the agricultural greenhouse according to the control strategy generated by the AI ​​algorithm to ensure the optimal environment for crop growth; Step S4. Transmit the greenhouse environmental data to the cloud platform for long-term storage and analysis to support future trend prediction and data analysis; Step S5. Dynamically optimize the resource allocation of the agricultural greenhouse to ensure efficient use of energy and water resources and reduce waste.

[0014] Furthermore, step S1 includes: Step S11. Deploy environmental sensors in the agricultural greenhouse to collect real-time environmental data such as temperature, humidity, and light intensity; Step S12: The collected data is initially filtered and processed by the edge computing node, ready for subsequent analysis.

[0015] Furthermore, step S2 includes: Step S21: Analyze the processed environmental data using a deep learning algorithm to generate an environmental control strategy; Step S22: The analysis results are transmitted to the edge computing module, and real-time decision generation is performed through the intelligent decision module.

[0016] Furthermore, step S3 includes: Step S31. Control the temperature and humidity control, irrigation, and lighting systems in the agricultural greenhouse according to the environmental control strategy generated by the decision; Step S32: Adjust the operation of the facility to maintain a suitable crop growth environment. Further, step S4 includes: Step S41. Transmitting the environmental data in the agricultural greenhouse to the cloud platform for storage; Step S42: Use the cloud platform to conduct long-term data analysis and trend forecasting to facilitate subsequent regulation and optimization.

[0017] Furthermore, step S5 includes: Step S51. Monitor the resources in the agricultural greenhouse and record the consumption of various resources; Step S52: Adjust the energy and water resource allocation in the greenhouse according to resource consumption, optimize resource allocation, and improve efficiency.

[0018] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention integrates edge computing with intelligent decision-making modules to collect and process environmental data from agricultural greenhouses in real time. It then uses AI algorithms for deep learning analysis to generate precise environmental control strategies. This intelligent control method significantly improves greenhouse management efficiency and environmental control precision, avoids the inefficiency of traditional manual intervention, optimizes environmental management processes, and ensures an optimal environment for crop growth.

[0019] 2. This invention utilizes intelligent algorithms to dynamically optimize resource allocation in agricultural greenhouses, predict and adjust energy consumption, and reduce resource waste. Through precise temperature and humidity control, intelligent irrigation, and light regulation, it further improves resource utilization efficiency in agricultural greenhouses, ensuring optimal conditions for crop growth, saving energy, and enhancing overall resource allocation efficiency.

[0020] 3. This system can calculate optimal resource allocation based on the actual needs and environmental requirements of greenhouses, reducing the additional costs caused by resource waste or shortages. Through real-time data monitoring and analysis, the system maximizes resource utilization efficiency, improves the return on agricultural production, and ensures the sustainability and economic benefits of the agricultural production process. This system also improves the management efficiency and control precision of agricultural greenhouses. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 This is a schematic diagram of the structure of the digital supervision system for software development based on the cloud platform of the present invention; Figure 2 It is a schematic diagram of the steps of the smart agricultural greenhouse control method of the present invention. DETAILED DESCRIPTION

[0022] The following embodiments of the present invention are described in further detail with reference to the accompanying drawings and examples. The following examples are used to illustrate the present invention but are not intended to limit the scope of the present invention.

[0023] See also Figure 1 and Figure 2 , the present invention provides a technical solution: a digital supervision system for software R&D based on a cloud platform, comprising: a cloud communication module, an R&D container module, a test estimation module, a quantitative recording module and a resource management module; The cloud communication module is used to build a cloud-based test and supervision platform, uniformly manage the data required in each R&D process, and recover output data from the R&D process. At the same time, it builds a communication channel between software R&D equipment and the cloud platform, and sets up a data unpacking program at the end of the communication channel to capture the output data packets of the R&D equipment.

[0024] The cloud communication module includes: platform storage unit and data communication unit; The platform storage unit is used to divide space in the cloud to build a data processing platform and database, and to centrally manage test data; The data communication unit is composed of communication management software installed in the R&D equipment and the cloud server, which is used to manage the communication process between the equipment and the server.

[0025] The R&D container module is used to add identifier placeholders according to specific rules in the test data provided to the R&D process by the cloud. It also captures the identifier placeholders in the output data of the R&D process, calculates the one-dimensional test parameters of the software in the development process based on the changes in the position and proportion of the placeholders in the output data, and stores the test results in the cloud database, which is kept updated in real time. The R&D container module includes: process closure unit, data circulation unit and one-dimensional testing unit; The process closure unit is used to close the software functions involved in each R&D process and form a separate test unit; The data flow unit is used to insert placeholders according to preset rules in the input data provided by the cloud to the R&D process, and to capture the output data uploaded to the cloud to capture the identification placeholders in the output data; The one-dimensional test unit is used to calculate the one-dimensional test parameters of the software, including software test delay and R&D data redundancy, based on the position and proportion of the placeholders in the output data.

[0026] The test estimation module is used to generate test cases for concurrent testing of R&D software. It analyzes test feedback data in the cloud platform to obtain two-dimensional test results for the software. The two-dimensional test results include all test items in the one-dimensional test parameters. It calculates the correlation coefficient between the two-dimensional test results of the software and the one-dimensional test parameters of each process, and estimates all test item parameters of the process based on the correlation coefficient. The test estimation module includes: concurrent test unit and related estimation unit; The concurrent testing unit is used to perform concurrent testing on the final software, obtaining all the test items of the software and recording them as two-dimensional test results; The correlation estimation unit is used to fit the correlation between one-dimensional test parameters and two-dimensional test results, and to construct correlation to estimate the test item parameters within the process; The quantitative recording module is used to obtain the comprehensive quality parameters of each R&D process according to the preset weighted calculation method based on all test item parameters of each R&D process. It then plots the relationship function between the comprehensive quality parameters and data resource consumption. When the decline rate of the relationship function exceeds the threshold, the person in charge of the corresponding R&D process is reminded through the cloud platform. The quantitative record module includes: quality assessment unit and supervision alarm unit; The quality assessment unit is used to obtain the comprehensive quality parameters of the process according to a preset weighted calculation method; The supervision alarm unit is used to alert the corresponding process when the functional decline rate between the comprehensive quality parameter and the data resource consumption is higher than the threshold.

[0027] The resource management module is used to adjust the resource allocation of the R&D process, recalculate the comprehensive quality weight of the process based on the adjusted test data, fit the joint influence function of the comprehensive quality weight and various types of resources, and use the actual usage requirements of the software as constraints to obtain the resource allocation conditions when the influence function takes the minimum value. R&D resources are allocated according to the calculation results of the resource allocation conditions.

[0028] The resource management module includes: influence fitting unit, constraint planning unit and resource allocation unit; The impact fitting unit is used to adjust the resource allocation of each R&D process during the R&D process, obtain the comprehensive quality parameters of the process, and calculate the impact function of the comprehensive quality parameters of each R&D process with the resource allocation; The constraint programming unit is used to calculate the maximum value of the influence function under the constraints, taking the actual usage requirements of the software as constraints; The resource allocation unit is used to adjust the resource allocation of each R&D process according to the linear programming results and send the allocation plan to the management department.

[0029] A digital supervision method for software development based on a cloud platform, comprising the following steps: Step S1. Build a test supervision platform in the cloud to uniformly manage the data required in each R&D process, and set up communication management software in the R&D equipment and cloud server. The communication management software can capture input and output data packets; Step S1 includes: Step S11. Divide the space in the cloud to build a test supervision platform. The test supervision platform consists of a user interface, database, data packet capture software and data analysis software to uniformly manage the input and output data of each R&D process; Step S12. Install communication management software in each R&D device to receive test data sent from the cloud and upload the data output by the R&D device to the test supervision platform. At the same time, set up receiving software in the cloud to capture the output data packets sent by the R&D device at the end of the channel.

[0030] Step S2. Each software function involved in the R&D process is enclosed as an independent test unit. The platform inserts identifiers into the data provided to each test unit according to the first rule. At the end of the channel, the identifiers are captured in the output data. The distribution of identifiers in the input and output data is analyzed to obtain one-dimensional test parameters. Step S2 includes: Step S21. Seal the software functions developed by each R&D process to form separate test units, numbered {W1, W2, ..., Wn}, where n is the number of R&D processes and Wn is the test unit corresponding to the nth R&D process; Step S22: The platform provides test data to each test unit and inserts an identifier into the test data according to the first rule. The first rule is as follows: Preset the position function F(x), x is the identifier serial number, let G(x)=TD[F(x)·cx], where G(x) is the insertion interval function, c is the preset amplification coefficient, c>0 and c≠1, TD is the rounding function, parse the test data packet, and insert the identifier after the G(1), G(2), ..., G(x)th data in the data packet until there is not enough data in the data packet; Step S23. Send the test data to the test unit, collect the output data in the test unit, capture the output identifiers in the output data, record the location of each identifier, and generate a function g(x) with the identifier number as the independent variable and the identifier location as the dependent variable; The test delay and data redundancy are calculated according to the distribution calculation method. The distribution calculation method is as follows:

[0031] Where tc represents the test delay, Q represents the data redundancy, x0 represents the number of identifiers inserted in the test data, v represents the cloud communication speed, G(i) and g(i) represent the function values ​​of G(x) and g(x) when the independent variable is i, respectively; Step S24: Record the one-dimensional test parameters of each test unit, where the one-dimensional test parameters include test delay and data redundancy.

[0032] Step S3. Perform concurrency testing on all test items of the delivered software. The test results are recorded as two-dimensional test parameters. The correlation between the one-dimensional test parameters and the two-dimensional test parameters is calculated. Based on the correlation, the parameters of the untested items in each test unit are estimated. Step S3 includes: Step S31. Combine all test units to obtain a fully functional delivered software. Use concurrency testing technology to test all test items of the delivered software to obtain two-dimensional test parameters of the delivered software. The two-dimensional test parameters include: data throughput, data utilization, maximum concurrency, test latency, and data redundancy. Step S32. Based on the common items in the two-dimensional test parameters and the one-dimensional test parameters, establish the correlation between the delivered software and each test unit, and list the equation: A1·r1+A2·r2+…+An·rn=A0; Where A1, A2, ...An represent the one-dimensional test parameters in the 1st, 2nd, ...nth test units, respectively; A0 represents the corresponding two-dimensional test parameters; r1, r2, ...rn represent the correlation coefficients of the 1st, 2nd, ...nth test units, respectively. Substitute the test results of n consecutive tests and solve for the values ​​of the correlation coefficients. Step S33. After obtaining the correlation coefficient between each test unit and the delivered software, for each untested item in the one-dimensional test parameter, list the equation E1·r1+E2·r2+…+En·rn=E0, where E1, E2,…En represent the test items not involved in the 1st, 2nd,…nth test units respectively, and E0 represents the test parameters for the test items in the two-dimensional test parameters. Substitute the two-dimensional test parameters in the n tests and output the calculation results as the parameters of the untested items in each test unit.

[0033] Step S4. Preset the weight of each test item, perform weighted calculation based on the test item parameters corresponding to each R&D process, and obtain the comprehensive quality parameter corresponding to the R&D process. When the comprehensive quality parameter is lower than the threshold, an alarm is issued; Step S4 includes: Step S41. Calculate the comprehensive quality parameter P for each R&D process, P = a1·h1+a2·h2+…+am·hm, where m represents the number of test items, a1, a2, …am represent the preset weights of the 1st, 2nd, …mth test items, respectively, and h1, h2, …hm represent the test parameters of the test unit in the R&D process for the 1st, 2nd, …mth test items, respectively; Step S42: For each R&D process, when the quality parameter is lower than a preset threshold, an alarm is sent to the control center of the corresponding R&D process.

[0034] Step S5. Adjust the resource allocation of each R&D process, use the function fitting tool to output the function of the project's comprehensive quality parameters changing with various types of resources, perform linear programming on the change function with the total resource amount as the constraint condition, and use the planning results to allocate resources for the next cycle.

[0035] Step S5 includes: Step S51. Adjust the various R&D resources allocated to the R&D process, record the comprehensive quality parameters of the R&D process before and after the adjustment, and use a function fitting tool to generate an influence function H(g1, g2, …, gt) with the number of each R&D resource as the independent variable and the comprehensive quality parameter as the dependent variable, where t represents the number of R&D resource types and g1, g2, …, gt represent the number of R&D resources in categories 1, 2, …, t, respectively. Step S52: Perform linear programming on the sum of the influence functions of each R&D process and output the maximum value of the total influence function:

[0036] Where Max represents the maximum value of the objective function, St represents the constraint condition, Y(g1, g2, …, gt) represents the total impact function of the delivered software, H1, H2, …, Hn represent the impact functions corresponding to the 1st, 2nd, …nth R&D processes, respectively, and g1max, g2max, …gtmax represent the maximum amount of R&D resources invested in the 1st, 2nd, …tth categories, respectively. Step S53: Allocate resources to each R&D process according to the allocated quantity of each R&D resource when the total impact function takes the maximum value in step S52.

[0037] Example: An R&D project has three R&D processes and five inspection items. The inspection results of the one-dimensional test parameters in the R&D processes are [0.6, 200], [0.2, 100], and [1.1, 120], respectively. The two-dimensional test parameters for the delivered software are [2.1, 1000, 50, 10, 800]. The first two test items are the same as the one-dimensional test parameters. After three consecutive tests, the remaining three parameters of each R&D process are estimated, and the estimation results are [20, 3, 250], [30, 5, 400], and [45, 8, 500], respectively. Calculation is performed using preset weights, and the comprehensive quality parameters of the three R&D processes are obtained as 17, 24, and 14, respectively. The comprehensive quality parameter of R&D process 3 is lower than the threshold of 15, and an alarm is issued for R&D process 3.

[0038] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0039] The embodiments of the present invention are presented for purposes of illustration and description and are not intended to be exhaustive or to limit the invention to the disclosed forms. Many modifications and variations will be apparent to those skilled in the art. The embodiments are chosen and described in order to better illustrate the principles of the invention and its practical application and to enable those skilled in the art to understand the invention and design various embodiments with various modifications as suited for specific applications.

Claims

1. Smart agricultural greenhouse control system based on adaptive distributed edge computing and AI algorithm, characterized by include: Edge computing node module, used to collect and pre-process environmental data in real time on the device side; Multimodal data fusion module, which integrates sensor data and video data and dynamically adjusts data weights through an attention mechanism; Deep learning framework module, using convolutional neural networks (CNN) and long short-term memory networks (LSTM) for data prediction and analysis; Federated learning collaboration module, which enables cross-node collaborative learning through local training and parameter sharing; Intelligent control module, which automatically adjusts temperature, humidity, irrigation and lighting systems based on analysis results; The data privacy and security module ensures data privacy during edge computing and federated learning.

2. The smart agricultural greenhouse control system based on adaptive distributed edge computing and AI algorithm as claimed in claim 1, characterized in that: The edge computing node module includes a data collection unit and a data processing unit. The data collection unit collects temperature, humidity and light intensity data in real time and performs preliminary filtering. The data processing unit formats and preliminarily analyzes the data.

3. The smart agricultural greenhouse control system based on adaptive distributed edge computing and AI algorithm as claimed in claim 1, characterized in that: The multimodal data fusion module introduces an attention mechanism to dynamically adjust the weights of sensor data and visual data to generate an accurate agricultural environment model.

4. The smart agricultural greenhouse control system based on adaptive distributed edge computing and AI algorithm as claimed in claim 1, characterized in that: The deep learning framework module uses CNN to identify crop pests and diseases and growth status, and uses LSTM to analyze time series data to capture long-term environmental change trends.

5. The smart agricultural greenhouse control system based on adaptive distributed edge computing and AI algorithm as claimed in claim 1, characterized in that: The federated learning collaborative module trains the model locally on each edge node and only shares the model parameters to avoid original data transmission and improve system performance.

6. The smart agricultural greenhouse control system based on adaptive distributed edge computing and AI algorithm as claimed in claim 1, characterized in that: The intelligent control module includes a temperature and humidity adjustment unit, an irrigation control unit, and a lighting adjustment unit, and dynamically optimizes the equipment operating status based on real-time environmental parameters.

7. Intelligent agricultural greenhouse control method based on adaptive distributed edge computing and AI algorithm, characterized by The following steps are involved: Step S1: Deploy edge computing nodes in agricultural greenhouses to collect and preprocess environmental data in real time; Step S2: Analyze data using deep learning algorithms to generate environmental control strategies; Step S3: Automatically adjust temperature, humidity, irrigation, and lighting systems according to the strategy; Step S4: Transmit data to the cloud for long-term storage and trend prediction; Step S5: Dynamically optimize resource allocation to reduce energy and water waste.

8. The intelligent agricultural greenhouse control method based on adaptive distributed edge computing and AI algorithm as claimed in claim 7, characterized in that: The step S1 comprises: Step S11: deploying temperature and humidity, soil moisture and light sensors to collect data; Step S12: Perform data filtering, formatting, and preliminary analysis at the edge node.

9. The intelligent agricultural greenhouse control method based on adaptive distributed edge computing and AI algorithm as claimed in claim 7, characterized in that: The step S2 comprises: Step S21: identifying crop pests and diseases and growth status through CNN; Step S22: Analyze time series data through LSTM to predict environmental change trends.

10. The intelligent agricultural greenhouse control method based on adaptive distributed edge computing and AI algorithm as claimed in claim 7, characterized in that: The step S5 comprises: Step S51: Monitor energy and water consumption data and generate an optimization plan; Step S52: Optimize resource allocation in real time based on the dynamic adjustment unit to improve resource utilization.

Images