Smart park policy intelligent matching calculation method and system based on large model

Abstract

The invention relates to a smart park policy intelligent matching calculation method based on a large model, belongs to the field of policy optimization, and particularly belongs to the field of data processing systems or methods specially suitable for administrative, commercial, financial, management, supervision or prediction purposes. The method comprises the following steps: taking each number proportion obtained when servers with different number proportions are allocated to each settled enterprise of the smart park as a number proportion set; and determining an intelligent matching quantity proportion set based on the delay time proportion of each settled enterprise corresponding to each quantity proportion set predicted intelligently by the large model. The invention further relates to an intelligent matching calculation system for the intelligent park policy based on the large model. According to the method and the device, aiming at the technical problem that time-sharing dynamic intelligent matching of shared server resources in a smart park is difficult in the prior art, a traversal mode is adopted to carry out directional intelligent prediction of various operation states corresponding to various server configuration modes respectively, so that the technical problem is solved.

Description

Technical Field

[0001] The strategy optimization of this invention relates more specifically to the field of data processing systems or methods specifically applicable to administrative, commercial, financial, management, supervision or forecasting purposes, and in particular to a smart park policy intelligent matching calculation method and system based on a large model. Background Technology

[0002] Strategy optimization, in the field of data processing systems or methods specifically applied to administrative, commercial, financial, managerial, supervisory, or predictive purposes, manifests as the application of data processing to perform administrative, commercial, financial, managerial, supervisory, or predictive actions to achieve strategic optimization objectives in the corresponding fields. A park refers to a centrally planned and designated area by the government, where specific types of enterprises or companies of a particular industry or form are established and managed uniformly; typical examples include industrial parks, free trade zones, and industrial parks. A smart park refers to the integration of resources inside and outside the park, and, based on the park's energy management planning requirements, the construction of a highly efficient smart park service platform through the Internet of Things and network technologies. This platform helps the park achieve functions and services such as community environmental monitoring, artificial intelligence analysis, security area monitoring, high-definition monitoring of key areas, and emergency command, creating an intelligent smart park. For each smart park, it is necessary to intelligently optimize the management or allocation strategies of its various shared resources to achieve efficient and high-quality development of the overall smart park economy.

[0003] For example, Chinese invention patent publication CN119090336A proposes a comprehensive index analysis system for the operation and construction of smart parks. The system includes: a data acquisition and processing module for collecting various categories of data from smart parks; an energy consumption monitoring module for acquiring comprehensive energy system planning information and operational optimization data for the park, and updating energy consumption data in real time; a policy analysis and talent evaluation module for developing standardized interfaces for policy recommendations, evaluating enterprise talent, and updating enterprise operational vitality data in real time; a park environmental maintenance module for updating the park environmental maintenance index in real time; and an index analysis module for analyzing population density data within the monitoring period of each target park, and combining this with real-time updated energy consumption data, enterprise operational vitality data, and park environmental maintenance data to evaluate the comprehensive operation and construction index of each target park within the monitoring period. This invention can solve the problem of comprehensive quantitative reference evaluation of the operation and construction of smart parks, providing effective data support for the evaluation of smart park operation.

[0004] For example, Chinese invention patent publication CN119295048A proposes a smart park integrated operation and maintenance intelligent management platform, relating to the field of smart park technology. The platform includes: a data acquisition module, a data storage module, an operation and maintenance management analysis module, an operation and maintenance service analysis module, and a comprehensive evaluation module. Its key technical points are: analyzing smart park operation and maintenance management; calculating the relative deviation between the current park equipment performance change value and the standard value of equipment performance change by analyzing equipment performance parameters to obtain equipment operation and maintenance evaluation values; evaluating the effectiveness of equipment operation and maintenance management to promptly identify and resolve problems, improving equipment operating efficiency and productivity; and obtaining park environmental operation and maintenance evaluation values ​​by correlating current park environmental data with environmental data of the city, which can assess the effectiveness of park environmental governance measures, improve park environmental quality, and enhance the accuracy and rationality of evaluation results.

[0005] However, the aforementioned existing technologies only involve a rough analysis and assessment of the operational status of smart parks. They cannot perform unified and effective overall dynamic time-sharing allocation of key resources shared by all resident companies in a smart park, such as shared server resources. They cannot improve the utilization rate of shared resources while ensuring the reliable operation of each resident company. The bottleneck lies in the fact that different resident companies have different operational needs for shared resources at different times, making it difficult to carry out unified and effective overall dynamic time-sharing allocation for formulating time-sharing allocation policies for different time intervals. Summary of the Invention

[0006] To address the technical problems in existing technologies, this invention provides a smart park policy intelligent matching calculation method and system based on a large model. Building upon a mechanism that iterates through various server configurations, a mechanism for customizing different large models for different smart parks, and a mechanism for targeted filtering of basic data for intelligent prediction of operational status, this invention addresses the application scenario where multiple resident companies within a smart park simultaneously share servers of a set scale. It employs a traversal approach to intelligently predict various operational statuses corresponding to different server configurations for future time intervals, thereby selecting the server configuration corresponding to the optimal operational status as the best server allocation policy for the smart park in the future time interval. This achieves time-sharing dynamic intelligent matching of the optimal configuration policy for shared server resources in the smart park.

[0007] According to a first aspect of the present invention, a method for intelligent matching calculation of smart park policies based on a large model is provided, the method comprising: The various quantities of servers obtained when allocating different quantities of servers to each enterprise in the smart park are collected as a set of quantities. The various allocations are then iterated to obtain the sets of quantities. Obtain the operational status information of each enterprise in the smart park for each past time interval before the current moment, with each time interval having an equal duration; To build an intelligent status prediction model for smart parks, based on a customized structural design of the total number of servers provided in the computer rooms of smart parks; The intelligent state prediction model uses the proportion of each quantity in each quantity proportion set, the operational status information of each enterprise's operation program in each past time interval before the current time, the duration of each time interval, and the enterprise-related data of each enterprise to intelligently predict the proportion of each delay time in the current time interval under the configuration of the quantity proportion set. The current time interval starts from the current time. The intelligent matching quantity percentage set is determined based on the percentage of delay time corresponding to each obtained quantity percentage set.

[0008] According to a second aspect of the present invention, a smart park policy intelligent matching calculation system based on a large model is provided, the system comprising: The first parsing device is used to obtain a set of quantity proportions when allocating different number proportions of servers to each enterprise in the smart park, and to traverse various allocations to obtain a set of quantity proportions. The second analysis device is used to obtain the operational status information of each enterprise in the smart park in each past time interval before the current moment, with each time interval having an equal duration. The third analytical device is used to build an intelligent state prediction model for smart parks based on the total number of servers provided in the computer room of the smart park and the customized structural design. The state prediction device is connected to the first analysis device, the second analysis device, and the third analysis device respectively. It is used to use an intelligent state prediction model to intelligently predict the proportion of delay time that the enterprise operation program of each enterprise in the current time interval will appear under the configuration of the quantity proportion set, based on the proportion of each quantity in each quantity proportion set, the operation status information corresponding to each enterprise operation program of each enterprise in the past time interval before the current time, the duration of each time interval, and the enterprise association data corresponding to each enterprise in the current time interval. The current time interval starts from the current time. The intelligent matching device, connected to the state prediction device, is used to determine the intelligent matching quantity proportion set based on the proportion of each delay time corresponding to each acquired quantity proportion set.

[0009] Compared with the prior art, the present invention has at least the following five key inventive points: Invention Point A: For the application scenario where all resident enterprises in a smart park share the same servers in the same data center, various server allocation modes are traversed based on the number of servers in the smart park data center and the number of resident enterprises to obtain various quantity percentage sets. Each quantity percentage set sequentially shows the quantity percentage of each server occupied by each resident enterprise under the current allocation mode. Then, an intelligent state prediction model designed for the smart park's customized structure is used to intelligently predict the latency percentage of each resident enterprise in the current time interval, which is a future time interval. The quantity percentage set with the optimal latency state is used as the intelligent matching quantity percentage set for the current time interval, and thus serves as the optimal server allocation policy for the smart park in the current time interval. In this way, the optimal server allocation policy for the smart park in the current time interval is determined in advance by using a traversal mode, which improves the utilization rate of smart park servers while ensuring the operational effect and efficiency of the enterprise operation programs of each resident enterprise in the smart park. Invention Point B: The intelligent state prediction model designed for the customized structure of smart parks is a deep neural network that has been trained multiple times. The number of training times of the deep neural network changes with the total number of servers provided in the smart park's computer room. The deep neural network has multiple hidden layers, a single output layer, and a single input layer. The multiple hidden layers are located between the single input layer and the single output layer. The number of hidden layers of the deep neural network is positively correlated with the number of enterprises in the smart park. This allows for the design of intelligent state prediction models with different customized structures for different smart parks, improving the stability and reliability of the intelligent prediction results of the delay time proportions corresponding to each enterprise obtained using each set of quantity proportions in the current time interval. Invention Point C: For the intelligent prediction of the delay time percentages of each resident enterprise obtained by using each set of quantity percentages for the current time interval, various basic data are selectively filtered. These basic data include the quantity percentages in each set of quantity percentages, the operational status information of each resident enterprise's business operation program in each past time interval before the current moment, the duration of each time interval, and the enterprise-related data of each resident enterprise. The selective filtering of the above basic data further improves the stability and reliability of the intelligent prediction results of the delay time percentages of each resident enterprise obtained by using each set of quantity percentages for the current time interval. Invention Point D: Specifically, the operational status information of each enterprise's operating program in the smart park for each past time interval before the current moment includes the peak computing power, peak memory consumption, peak communication bandwidth, server allocation ratio, and latency ratio of the enterprise's operating program in the past time interval. The enterprise-related data for each enterprise includes the number of employees of that enterprise, the maximum computing power of the office computer per unit time, the memory capacity of the office computer, and the communication bandwidth of the office computer. Each enterprise uses the same model of office computer, and the number of time intervals for each past time interval before the current moment is proportional to the number of enterprises in the smart park, thereby designing a customized data structure for the targeted selection of various basic data. Invention Point E: In each training iteration of the deep neural network, the percentage of delay time of a known resident enterprise in a smart park within a certain past time interval, configured with a certain set of quantity percentages, is used as a single output of the deep neural network. The quantity percentage of the resident enterprise within the certain set of quantity percentages, the operational status information of the resident enterprise's business process corresponding to each past time interval before the certain past time interval, the duration of each time interval, and the enterprise association data corresponding to the resident enterprise are used as various inputs of the deep neural network to complete the training, thereby ensuring the training effect of each iteration of the deep neural network. Attached Figure Description

[0010] The embodiments of the present invention will now be described with reference to the accompanying drawings, wherein: Figure 1 This is a schematic diagram of the working scenario of the intelligent matching calculation method and system for smart park policies based on a large model according to the present invention.

[0011] Figure 2 The following is a flowchart illustrating the steps of a smart park policy intelligent matching calculation method based on a large model, according to Embodiment 1 of the present invention.

[0012] Figure 3 The flowchart illustrates the steps of the intelligent matching calculation method for smart park policies based on a large model, as shown in Embodiment 2 of the present invention.

[0013] Figure 4 The following is a flowchart illustrating the steps of the intelligent matching calculation method for smart park policies based on a large model, as shown in Embodiment 3 of the present invention.

[0014] Figure 5 The following is a flowchart illustrating the steps of the intelligent matching calculation method for smart park policies based on a large model, as shown in Embodiment 4 of the present invention.

[0015] Figure 6 The following is a flowchart illustrating the steps of the intelligent matching calculation method for smart park policies based on a large model, as shown in Embodiment 5 of the present invention.

[0016] Figure 7 This is a schematic diagram of the structure of a smart park policy intelligent matching calculation system based on a large model, according to Embodiment 6 of the present invention. Detailed Implementation

[0017] Powered by technologies such as artificial intelligence, deep learning, and digital twins, smart industrial parks become organic entities based on data-driven automatic control, autonomous learning, self-evolution, and autonomous decision-making. By establishing industry alliances, conducting technological innovation, and promoting resource sharing, they integrate core systems within the park, including production, transportation, living, municipal services, energy, commerce, and business, fostering cooperation and collaborative development among enterprises and achieving sustainable development and shared benefits for the park.

[0018] However, how to dynamically and intelligently match shared resources in a smart park, such as shared server resources, among the resident companies, and thus determine the optimal server allocation policy for future time periods in advance, in order to improve server utilization while ensuring the operational effectiveness and efficiency of the business processes of each resident company, remains a technical challenge in the existing technology. This invention aims to solve this technical challenge.

[0019] like Figure 1 The diagram illustrates a working scenario of the intelligent matching calculation method and system for smart park policies based on a large model, as presented in this invention. The strategy optimization of this invention more specifically relates to the field of data processing systems or methods applicable to administrative, commercial, financial, management, supervisory, or predictive purposes.

[0020] The specific technical process of this invention is as follows: Technical Process 1: For application scenarios where various resident companies in a smart park share servers within the same data center, based on the number of servers in the smart park's data center and the number of resident companies, various server allocation modes are iterated to obtain a set of proportions for each mode, such as... Figure 1 As shown, N sets of quantity percentages were obtained, where N is a large number of values ​​with relatively large values; Specifically, in each set of quantity proportions, the percentage of each server occupied by each resident enterprise under this allocation mode is given in sequence. The cumulative value of each quantity proportion in each set of quantity proportions is equal to 1. Each resident enterprise in the smart park only uses the servers provided in the smart park's computer room, and each server has the same maximum computing power, memory capacity and communication bandwidth per unit time. Specifically, the sequence number of each quantity percentage corresponding to each resident enterprise in each quantity percentage set is equal to the enterprise sequence number of each resident enterprise in the smart park, that is, the sequence number of each quantity percentage corresponding to each resident enterprise in each quantity percentage set is fixed. In this way, for the application scenario where all the companies in the smart park share the same server in the same computer room, a traversal approach is used to provide a massive number of server configuration methods. Specifically, the set of each quantity percentage is given, thus providing a configuration basis for intelligent prediction of the overall operation status of the smart park in a massive number of subsequent times. Technical Process Two: Designing intelligent state prediction models with different customized structures for smart parks of different sizes, such as... Figure 1 As shown; Specifically, the structural customization of the intelligent state prediction model designed for smart parks is mainly reflected in the following aspects: Aspect A: The intelligent state prediction model designed for smart parks is a deep neural network that has been trained multiple times. The deep neural network has multiple hidden layers, a single output layer, and a single input layer. Aspect B: The trend of the number of training iterations of deep neural networks as a function of the total number of servers provided in the computer room of the smart park; For example, the total number of servers provided in the computer room of the smart park is 100, and the number of training times for the deep neural network is 500; the total number of servers provided in the computer room of the smart park is 150, and the number of training times for the deep neural network is 750; the total number of servers provided in the computer room of the smart park is 200, and the number of training times for the deep neural network is 1000; the total number of servers provided in the computer room of the smart park is 300, and the number of training times for the deep neural network is 1500, and so on. Aspect C: In the deep neural network used, multiple hidden layers are located between a single input layer and a single output layer, and the number of hidden layers in the deep neural network is positively correlated with the number of enterprises in the smart park; For example, if the number of companies in a smart park is 10, the number of hidden layers in a deep neural network is 3; if the number of companies in a smart park is 15, the number of hidden layers in a deep neural network is 5; if the number of companies in a smart park is 20, the number of hidden layers in a deep neural network is 7, and so on. It can be seen that aspects B and C reflect the customized structural design of different intelligent state prediction models for different smart parks; Aspect D: In each training iteration of the deep neural network, the percentage of delay time of a known resident enterprise in the smart park within a certain past time interval under a certain quantity percentage set is used as the single output of the deep neural network. The quantity percentage of the resident enterprise in the certain quantity percentage set, the operational status information of the resident enterprise's business process in each past time interval before the certain past time interval, the duration of each time interval, and the enterprise association data corresponding to the resident enterprise are used as the inputs of the deep neural network to complete the training, thereby ensuring the training effect of each training iteration of the deep neural network. In this way, through the above-mentioned customized structural design, the stability and reliability of the intelligent prediction results of the delay time ratio of each resident enterprise obtained by using each quantity ratio set in the current time interval are improved. Technical Process 3: Intelligent prediction of the delay time percentage of each resident enterprise obtained by using each quantity percentage set for the current time interval, that is, intelligent prediction of the overall operation status of the smart park corresponding to each quantity percentage set for the current time interval, and targeted screening of various basic data. Specifically, the basic data includes the percentage of each quantity in each quantity percentage set, the operational status information of each enterprise's operation program in each past time interval before the current moment, the duration of each time interval, and the enterprise association data corresponding to each enterprise. More specifically, the operational status information of each enterprise's operating program in the smart park for each past time interval before the current moment includes the peak computing volume, peak memory consumption, peak communication bandwidth, server allocation ratio, and latency ratio of the enterprise's operating program in the past time interval. The enterprise-related data for each enterprise includes the number of employees of that enterprise, the maximum computing volume of the office computer per unit time, the memory capacity of the office computer, and the communication bandwidth of the office computer. Each enterprise uses the same model of office computer, and the number of time intervals for each past time interval before the current moment is proportional to the number of enterprises in the smart park. This allows for the design of a customized data structure for the targeted selection of various basic data. For example, the number of time intervals in each past time interval before the current moment is proportional to the number of enterprises in the smart park, including: the number of enterprises in the smart park is 10, the number of time intervals in each past time interval before the current moment is 5, the number of enterprises in the smart park is 16, the number of time intervals in each past time interval before the current moment is 8, the number of enterprises in the smart park is 20, the number of time intervals in each past time interval before the current moment is 10, and so on; In this way, by selectively filtering the above-mentioned basic data, the stability and reliability of the intelligent prediction results of the delay time percentages for each resident company obtained using each set of quantity percentages in the current time interval are further improved. Technical Process Four: Utilizing the intelligent state prediction model designed for the smart park using Technical Process Two, and based on the targeted basic data selected in Technical Process Three, before the current time interval arrives, it iterates through the various server allocation patterns obtained in Technical Process One, i.e., iterates through the various quantity percentage sets obtained in Technical Process One, to obtain the overall operational status of the smart park corresponding to each quantity percentage set in the current time interval, such as... Figure 1 As shown, N items of the overall operation status of the smart park were obtained, where N is a large number of values ​​with a large range. Specifically, the overall operational status of the smart park corresponding to each set of quantity proportions in the current time interval is obtained by intelligently predicting the respective delay time proportions of each resident enterprise in the smart park under the server allocation mode of each set of quantity proportions in the current time interval. Technical Process 5: Based on the sets of quantity proportions obtained in Technical Process 4 for each time interval corresponding to the overall operational status of the smart park, a customized data analysis model is adopted to analyze the set of quantity proportions with the optimal overall latency status of the smart park as the intelligent matching set of quantity proportions for the current time interval, and then use it as the optimal server allocation policy for the smart park in the current time interval. That is, from Figure 1 Select the set of quantities corresponding to one of the N smart park overall operation statuses as the optimal server allocation policy for the smart park in the current time interval. Specifically, the customized data analysis model is as follows: obtain the latency time percentage of each resident enterprise under each quantity percentage set; take the quantity percentage set whose mean square deviation of each latency time percentage is greater than the set mean square deviation threshold as the candidate quantity percentage set to obtain multiple candidate quantity percentage sets; take the candidate quantity percentage set with the smallest mean of each latency time percentage as the intelligent matching quantity percentage set, that is, the quantity percentage set with the optimal latency status of the smart park, and use it as the best server allocation policy of the smart park for the current time interval to provide to each resident enterprise. The specific allocation is triggered and implemented by the smart park's data center at the current moment. It is evident that the coordinated operation of the above five technical processes enables the intelligent matching and pre-configuration of the optimal server allocation policy for the smart park, which is a future time interval, using a traversal mode. This improves the utilization rate of the smart park's servers while ensuring the operational effectiveness and efficiency of the business operation programs of each enterprise in the smart park.

[0021] The key points of this invention are: a customized data analysis mode for intelligent matching of quantity proportion set analysis, a mechanism for traversing various server configurations item by item, a mechanism for customizing different large models for different smart parks, and a targeted screening mechanism for various basic data for intelligent prediction of operational status.

[0022] The following will describe in detail the intelligent matching calculation method and system for smart park policies based on large models of the present invention through specific embodiments.

[0023] Example 1 Figure 2 The following is a flowchart illustrating the steps of a smart park policy intelligent matching calculation method based on a large model, according to Embodiment 1 of the present invention.

[0024] like Figure 2 As shown, the intelligent matching calculation method for smart park policies based on a large model includes the following specific steps: Step S201: When allocating servers with different numbers of proportions to each enterprise in the smart park, the various proportions are collected as a set of proportions. The various allocations are traversed to obtain the sets of proportions. For example, in a smart park with 200 shared servers, there are 16 companies. These 16 companies are numbered from 1 to 16, and each set of quantity proportions contains 16 quantity proportions from 1 to 16, which are assigned to these 16 companies respectively. Obviously, with 200 shared servers allocated to 16 resident companies, there is a massive set of distribution ratios, meaning there is a massive number of server allocation policies. The technical problem this invention aims to solve is how to select the optimal allocation policy for the current time segment suitable for the smart park from this massive number of server allocation policies. Here, the current time segment is based on the current moment as the starting moment, and therefore belongs to a type of future time segmentation. Step S202: Obtain the operational status information of each enterprise in the smart park for each past time interval before the current moment, with each time interval having an equal duration; For example, obtain the operational status information of each enterprise in the smart park for each past time interval before the current moment, with each time interval having an equal duration, including: each time interval lasting 2 hours. Step S203: Construct an intelligent state prediction model for the smart park based on the total number of servers provided in the computer room of the smart park and a customized structural design; Specifically, since different total numbers of servers represent smart parks with different scales of shared servers, different customized intelligent state prediction models are designed for different smart parks. Step S204: Using an intelligent state prediction model, based on the proportion of each quantity in each quantity proportion set, the operational status information corresponding to each past time interval before the current time of each enterprise's operation program, the duration of each time interval, and the enterprise-related data corresponding to each enterprise, the model intelligently predicts the proportion of each delay time that the enterprise's operation program will experience in the current time interval under the configuration of the quantity proportion set, with the current time interval starting from the current time. Specifically, the intelligent state prediction model can be used to intelligently predict the proportion of delay time of each enterprise's operating procedure in the current time interval under the configuration of the set of proportions of the number of enterprises. Alternatively, the intelligent state prediction model can be used to intelligently predict the proportion of delay time of one enterprise's operating procedure in the current time interval under the configuration of the set of proportions of the number of enterprises, and multiple intelligent predictions can be performed. Step S205: Determine the intelligent matching quantity proportion set based on the delay time proportions corresponding to each acquired quantity proportion set; In this way, the set of intelligent matching quantity proportions obtained represents the optimal server allocation policy for the smart park in the current time segment. The process of determining the intelligent matching quantity percentage set based on the delay time percentages corresponding to each acquired quantity percentage set includes: acquiring the delay time percentages corresponding to each quantity percentage set; taking the quantity percentage set whose mean square error of each delay time percentage is greater than a set mean square error threshold as a candidate quantity percentage set to obtain multiple candidate quantity percentage sets; taking the candidate quantity percentage set with the smallest mean of each delay time percentage as the intelligent matching quantity percentage set; and providing each quantity percentage in the intelligent matching quantity percentage set to each resident enterprise. Therefore, it can be seen that the customized data analysis mode of the above intelligent matching quantity proportion set reflects the search for a quantity proportion set where the difference in the delay time proportion of each enterprise is not large and the overall delay time proportion value is small. Among them, the number of servers obtained when allocating different numbers of servers to each enterprise in the smart park is a set of number percentages. The various allocations are traversed to obtain each set of number percentages, including: the cumulative value of each number percentage in each set of number percentages is equal to 1, and each enterprise in the smart park only uses the servers provided in the smart park's computer room, and each server has the same maximum computing power, memory capacity and communication bandwidth per unit time. Among them, obtaining the operational status information of each enterprise's operating program in the smart park for each past time interval before the current moment includes: the operational status information of each enterprise's operating program in the smart park for each past time interval before the current moment is the peak computing volume, peak memory consumption, peak communication bandwidth, server allocation ratio, and latency ratio of the enterprise's operating program in the past time interval. Among them, the enterprise-related data for each enterprise includes the number of employees of that enterprise, the maximum computing power of the office computer per unit time, the memory capacity of the office computer, and the communication bandwidth of the office computer. All office computers used by each enterprise are of the same model. The intelligent state prediction model for building a smart park based on the total number of servers provided in the smart park's computer room includes: the intelligent state prediction model is a deep neural network that has been trained multiple times, and the training times of the deep neural network change with the numerical trend of the total number of servers provided in the smart park's computer room. For example, the total number of servers provided in the computer room of the smart park is 100, and the number of training times for the deep neural network is 500; the total number of servers provided in the computer room of the smart park is 150, and the number of training times for the deep neural network is 750; the total number of servers provided in the computer room of the smart park is 200, and the number of training times for the deep neural network is 1000; the total number of servers provided in the computer room of the smart park is 300, and the number of training times for the deep neural network is 1500, and so on. The intelligent state prediction model is a deep neural network that has been trained multiple times, and the trend of the number of training times of the deep neural network with the total number of servers provided in the smart park's computer room includes: the curve of the number of training times of the deep neural network has the same curve curvature as the curve of the total number of servers provided in the smart park's computer room. The intelligent state prediction model is a deep neural network that has been trained multiple times. The trend of the number of training times of the deep neural network with the total number of servers provided in the smart park's computer room also includes: the deep neural network has multiple hidden layers, a single output layer and a single input layer. The multiple hidden layers are located between the single input layer and the single output layer. The number of hidden layers of the deep neural network is positively correlated with the number of enterprises in the smart park. For example, if the number of companies in a smart park is 10, the number of hidden layers in a deep neural network is 3; if the number of companies in a smart park is 15, the number of hidden layers in a deep neural network is 5; if the number of companies in a smart park is 20, the number of hidden layers in a deep neural network is 7, and so on. In each training iteration of the deep neural network, the percentage of delay time of a known resident company in the smart park within a certain past time interval, configured with a certain set of quantity percentages, is used as the single output of the deep neural network. The percentage of the resident company obtained in the certain set of quantity percentages, the operational status information of the resident company's business operation program corresponding to each past time interval before the certain past time interval, the duration of each time interval, and the enterprise association data corresponding to the resident company are used as the inputs of the deep neural network to complete this training.

[0025] Example 2 Figure 3 The flowchart illustrates the steps of the intelligent matching calculation method for smart park policies based on a large model, as shown in Embodiment 2 of the present invention.

[0026] like Figure 3 As shown, with Figure 2Unlike the previous implementation, in the smart park policy intelligent matching calculation method based on a large model, before obtaining the operational status information of each enterprise's operational procedures in the smart park for each past time interval before the current moment, and before the duration of each time interval is equal, i.e. before step S201, the method further includes: Step S206: Determine the number of time intervals for each past time interval before the current moment based on the number of enterprises in the smart park; Among them, determining the number of time intervals for each past time interval before the current moment based on the number of enterprises in the smart park includes: the number of time intervals for each past time interval before the current moment is proportional to the number of enterprises in the smart park; For example, if the number of companies in a smart park is 10 and the number of time intervals in each past time interval before the current time is 5, if the number of companies in a smart park is 16 and the number of time intervals in each past time interval before the current time is 8, if the number of companies in a smart park is 20 and the number of time intervals in each past time interval before the current time is 10, and so on.

[0027] Example 3 Figure 4 The following is a flowchart illustrating the steps of the intelligent matching calculation method for smart park policies based on a large model, as shown in Embodiment 3 of the present invention.

[0028] like Figure 4 As shown, with Figure 2 Unlike the previous implementation, in the smart park policy intelligent matching calculation method based on a large model, before traversing various allocations to obtain each set of quantity proportions by taking the different quantity proportions of servers allocated to each enterprise in the smart park as a set of quantity proportions, i.e. before step S201, the method further includes: Step S207: Collect the total number of servers provided in the computer room of the smart park and the number of enterprises in the smart park. For example, in a smart park with 200 shared servers, there are 16 resident companies. These 16 resident companies are numbered from 1 to 16. Each quantity percentage set contains 16 quantity percentages from 1 to 16. These are assigned to the application scenarios of these 16 resident companies. The total number of servers provided in the smart park's computer room is 200, and the number of resident companies in the smart park is 16.

[0029] Example 4 Figure 5 The following is a flowchart illustrating the steps of the intelligent matching calculation method for smart park policies based on a large model, as shown in Embodiment 4 of the present invention.

[0030] like Figure 5 As shown, with Figure 2 Unlike the previous implementation, in the intelligent matching calculation method for smart park policies based on a large model, after determining the intelligent matching quantity proportion set based on the delay time proportions corresponding to each acquired quantity proportion set, i.e., after step S205, the method further includes: Step S208: The determined set of intelligent matching quantity proportions is used as the optimal server allocation policy for the smart park in the current time interval and is wirelessly transmitted to the remote allocation policy service network element through the wireless communication network. For example, the remote allocation policy service network element is a big data service network element, a cloud computing service network element, or a blockchain service network element; The allocation policy service network element that wirelessly transmits the determined intelligent matching quantity ratio set as the optimal server allocation policy for the smart park in the current time interval to the remote end via the wireless communication network includes: the wireless communication network is based on time division duplex communication mode or frequency division duplex communication mode.

[0031] Example 5 Figure 6 The following is a flowchart illustrating the steps of the intelligent matching calculation method for smart park policies based on a large model, as shown in Embodiment 5 of the present invention.

[0032] like Figure 6 As shown, with Figure 2 Unlike the previous implementation, in the intelligent matching calculation method for smart park policies based on a large model, after determining the intelligent matching quantity proportion set based on the delay time proportions corresponding to each acquired quantity proportion set, i.e. after step S205, the method further includes: Step S209: Use the determined set of intelligent matching quantity proportions as the optimal server allocation policy for the smart park in the current time interval, and display the optimal server allocation policy for the smart park in the current time interval in the central control room of the smart park. Specifically, the determined set of intelligent matching quantity proportions is used as the optimal server allocation policy for the smart park in the current time interval. The optimal server allocation policy for the smart park in the current time interval is displayed on-site in the central control room of the smart park. This includes selecting a giant screen display in the central control room of the smart park to display the optimal server allocation policy for the current time interval on-site.

[0033] Next, the various method embodiments of the present invention will be described in detail.

[0034] In the intelligent matching calculation method for smart park policies based on a large model according to various method embodiments of the present invention: The various percentages obtained when allocating different numbers of servers to each enterprise in the smart park are collected as a set of percentages. The set of percentages obtained by traversing various allocations also includes: the ratio of the number of servers allocated to each enterprise in the smart park to the total number of servers provided in the smart park's data center is the percentage of the number of servers allocated to that enterprise. For example, in a smart park with 200 shared servers, there are 16 resident companies. These 16 resident companies are numbered from 1 to 16. Each quantity percentage set contains 16 quantity percentages from 1 to 16. In the application scenario where these 16 resident companies are allocated their respective quantities, if 20 shared servers are allocated to resident company number 10 in a certain quantity percentage set, then the specific value of the quantity percentage of set sequence 10 in that certain quantity percentage set is 20 / 200, which is 1 / 10. Among them, the operational status information of each enterprise's operating program in the smart park for each past time interval before the current moment includes the peak computing volume, peak memory consumption, peak communication bandwidth, server allocation ratio, and latency ratio of the enterprise's operating program in the past time interval. The latency ratio of the enterprise's operating program in the past time interval is the proportion of the total time of operational latency of the enterprise's operating program in the past time interval to the duration of the past time interval. And wherein, the proportion of the delay time of the enterprise operation program in the past time interval is the proportion of the total time of the enterprise operation program's operation delay in the past time interval to the duration of the past time interval, including: taking the cumulative value of each delay time corresponding to each operation delay of the enterprise operation program in the past time interval as the total time of the enterprise operation program's operation delay in the past time interval.

[0035] And in the intelligent matching calculation method for smart park policies based on a large model according to various method embodiments of the present invention: The intelligent state prediction model is based on the proportion of each quantity in each quantity proportion set, the operational status information of each enterprise's business operation program in each past time interval before the current time, the duration of each time interval, and the enterprise association data corresponding to each enterprise. It intelligently predicts the proportion of delay time that each enterprise's business operation program will have in the current time interval under the configuration of the quantity proportion set. The method is as follows: For each enterprise as the target enterprise, the proportion of the target enterprise in each quantity proportion set, the operational status information of the target enterprise's business operation program in each past time interval before the current time, the duration of each time interval, and the enterprise association data corresponding to the target enterprise are synchronously input into the intelligent state prediction model. For example, numerical simulation mode can be used to simulate and test the quantity proportion of the target resident enterprises in each quantity proportion set, the operation status information of the target resident enterprises in each past time interval before the current time, the duration of each time interval, and the synchronous input of the enterprise-related data of the target resident enterprises to the intelligent status prediction model. The method of using an intelligent state prediction model to intelligently predict the percentage of delay time of each enterprise's business operation program in the current time interval under the configuration of the number percentage set, based on the percentage of each quantity in each quantity percentage set, the operational status information corresponding to each enterprise's business operation program in each past time interval before the current time, the duration of each time interval, and the enterprise association data corresponding to each enterprise, also includes: running the intelligent state prediction model to obtain the percentage of delay time of the target enterprise's business operation program in the current time interval under the configuration of the number percentage set. Specifically, for each resident enterprise as a target resident enterprise, the following steps are taken: the target resident enterprise's quantity percentage in each quantity percentage set, the target resident enterprise's operational procedures for each past time interval before the current moment, the duration of each time interval, and the enterprise-related data corresponding to the target resident enterprise are synchronously input into the intelligent status prediction model. Specifically, the programmable logic device is used to synchronously input the quantity percentage of the target resident enterprise in each quantity percentage set, the operation status information of the target resident enterprise's enterprise operation program in each past time interval before the current time, the duration of each time interval, and the enterprise-related data corresponding to the target resident enterprise into the intelligent status prediction model. The programmable logic device can be an FPGA chip designed with VHDL language. In addition, the number percentage of the target enterprise in each number percentage set, the operational status information of the target enterprise's enterprise operation program in each past time interval before the current moment, the duration of each time interval, the enterprise association data corresponding to the target enterprise, and the delay time percentage of the target enterprise's enterprise operation program in the current time interval under the configuration of the number percentage set are all numerically normalized representations.

[0036] Example 6 Figure 7 This is a schematic diagram of the structure of a smart park policy intelligent matching calculation system based on a large model, according to Embodiment 6 of the present invention.

[0037] like Figure 7 As shown, the intelligent matching and calculation system for smart park policies based on a large model includes the following components: The first parsing device is used to obtain a set of quantity proportions when allocating different number proportions of servers to each enterprise in the smart park, and to traverse various allocations to obtain a set of quantity proportions. For example, in a smart park with 200 shared servers, there are 16 companies. These 16 companies are numbered from 1 to 16, and each set of quantity proportions contains 16 quantity proportions from 1 to 16, which are assigned to these 16 companies respectively. Obviously, with 200 shared servers allocated to 16 resident companies, there is a massive set of distribution ratios, meaning there is a massive number of server allocation policies. The technical problem this invention aims to solve is how to select the optimal allocation policy for the current time segment suitable for the smart park from this massive number of server allocation policies. Here, the current time segment is based on the current moment as the starting moment, and therefore belongs to a type of future time segmentation. The second analysis device is used to obtain the operational status information of each enterprise in the smart park in each past time interval before the current moment, with each time interval having an equal duration. For example, obtain the operational status information of each enterprise in the smart park for each past time interval before the current moment, with each time interval having an equal duration, including: each time interval lasting 2 hours. The third analytical device is used to build an intelligent state prediction model for smart parks based on the total number of servers provided in the computer room of the smart park and the customized structural design. Specifically, since different total numbers of servers represent smart parks with different scales of shared servers, different customized intelligent state prediction models are designed for different smart parks. The state prediction device is connected to the first analysis device, the second analysis device, and the third analysis device respectively. It is used to use an intelligent state prediction model to intelligently predict the proportion of delay time that the enterprise operation program of each enterprise in the current time interval will appear under the configuration of the quantity proportion set, based on the proportion of each quantity in each quantity proportion set, the operation status information corresponding to each enterprise operation program of each enterprise in the past time interval before the current time, the duration of each time interval, and the enterprise association data corresponding to each enterprise in the current time interval. The current time interval starts from the current time. Specifically, the intelligent state prediction model can be used to intelligently predict the proportion of delay time of each enterprise's operating procedure in the current time interval under the configuration of the set of proportions of the number of enterprises. Alternatively, the intelligent state prediction model can be used to intelligently predict the proportion of delay time of one enterprise's operating procedure in the current time interval under the configuration of the set of proportions of the number of enterprises, and multiple intelligent predictions can be performed. The intelligent matching device, connected to the state prediction device, is used to determine the intelligent matching quantity proportion set based on the proportion of each delay time corresponding to each acquired quantity proportion set. In this way, the set of intelligent matching quantity proportions obtained represents the optimal server allocation policy for the smart park in the current time segment. The process of determining the intelligent matching quantity percentage set based on the delay time percentages corresponding to each acquired quantity percentage set includes: acquiring the delay time percentages corresponding to each quantity percentage set; taking the quantity percentage set whose mean square error of each delay time percentage is greater than a set mean square error threshold as a candidate quantity percentage set to obtain multiple candidate quantity percentage sets; taking the candidate quantity percentage set with the smallest mean of each delay time percentage as the intelligent matching quantity percentage set; and providing each quantity percentage in the intelligent matching quantity percentage set to each resident enterprise. Therefore, it can be seen that the customized data analysis mode of the above intelligent matching quantity proportion set reflects the search for a quantity proportion set where the difference in the delay time proportion of each enterprise is not large and the overall delay time proportion value is small. Among them, the number of servers obtained when allocating different numbers of servers to each enterprise in the smart park is a set of number percentages. The various allocations are traversed to obtain each set of number percentages, including: the cumulative value of each number percentage in each set of number percentages is equal to 1, and each enterprise in the smart park only uses the servers provided in the smart park's computer room, and each server has the same maximum computing power, memory capacity and communication bandwidth per unit time. Among them, obtaining the operational status information of each enterprise's operating program in the smart park for each past time interval before the current moment includes: the operational status information of each enterprise's operating program in the smart park for each past time interval before the current moment is the peak computing volume, peak memory consumption, peak communication bandwidth, server allocation ratio, and latency ratio of the enterprise's operating program in the past time interval. Among them, the enterprise-related data for each enterprise includes the number of employees of that enterprise, the maximum computing power of the office computer per unit time, the memory capacity of the office computer, and the communication bandwidth of the office computer. All office computers used by each enterprise are of the same model. The intelligent state prediction model for building a smart park based on the total number of servers provided in the smart park's computer room includes: the intelligent state prediction model is a deep neural network that has been trained multiple times, and the training times of the deep neural network change with the numerical trend of the total number of servers provided in the smart park's computer room. For example, the total number of servers provided in the computer room of the smart park is 100, and the number of training times for the deep neural network is 500; the total number of servers provided in the computer room of the smart park is 150, and the number of training times for the deep neural network is 750; the total number of servers provided in the computer room of the smart park is 200, and the number of training times for the deep neural network is 1000; the total number of servers provided in the computer room of the smart park is 300, and the number of training times for the deep neural network is 1500, and so on. The intelligent state prediction model is a deep neural network that has been trained multiple times, and the trend of the number of training times of the deep neural network with the total number of servers provided in the smart park's computer room includes: the curve of the number of training times of the deep neural network has the same curve curvature as the curve of the total number of servers provided in the smart park's computer room. The intelligent state prediction model is a deep neural network that has been trained multiple times. The trend of the number of training times of the deep neural network with the total number of servers provided in the smart park's computer room also includes: the deep neural network has multiple hidden layers, a single output layer and a single input layer. The multiple hidden layers are located between the single input layer and the single output layer. The number of hidden layers of the deep neural network is positively correlated with the number of enterprises in the smart park. For example, if the number of companies in a smart park is 10, the number of hidden layers in a deep neural network is 3; if the number of companies in a smart park is 15, the number of hidden layers in a deep neural network is 5; if the number of companies in a smart park is 20, the number of hidden layers in a deep neural network is 7, and so on. In each training iteration of the deep neural network, the percentage of delay time of a known resident company in the smart park within a certain past time interval, configured with a certain set of quantity percentages, is used as the single output of the deep neural network. The percentage of the resident company obtained in the certain set of quantity percentages, the operational status information of the resident company's business operation program corresponding to each past time interval before the certain past time interval, the duration of each time interval, and the enterprise association data corresponding to the resident company are used as the inputs of the deep neural network to complete this training.

[0038] In addition, the present invention may also cite the following technical contents to highlight the significant technical advancements of the present invention: The various quantity ratios obtained when allocating different numbers of servers to each resident enterprise in the smart park are collected as a set of quantity ratios. The process of traversing various allocations to obtain each set of quantity ratios also includes: the permutation number of each quantity ratio corresponding to each resident enterprise in each set of quantity ratios is equal to the enterprise number of each resident enterprise in the smart park, that is, the permutation number of each quantity ratio corresponding to each resident enterprise in each set of quantity ratios is fixed. For example, the sequence number of each quantity percentage corresponding to each resident enterprise in each quantity percentage set is equal to the enterprise number of each resident enterprise in the smart park, including: the sequence number of each quantity percentage corresponding to each resident enterprise in each quantity percentage set starts with 1 and ends with the value corresponding to the total number of enterprises of each resident enterprise. Among them, the various quantity ratios obtained when allocating different number ratios of servers to each resident enterprise in the smart park are taken as a quantity ratio set. The process of traversing various allocations to obtain each quantity ratio set also includes: when the quantity ratio corresponding to any sequence number in the quantity ratio set changes, it is regarded as a new quantity ratio set that needs to be traversed. The curve showing the numerical change of the number of training iterations of the deep neural network and the curve showing the numerical change of the total number of servers provided in the smart park's computer room have the same curvature. This includes: taking the curve showing the numerical change of the number of training iterations of the deep neural network as the first curve and the curve showing the numerical change of the total number of servers provided in the smart park's computer room as the second curve, wherein the curvature at uniformly spaced points on the first curve is equal to the curvature at uniformly spaced points on the first curve.

[0039] Those skilled in the art will understand that various modifications and alterations can be made without departing from the scope and spirit of this invention. Therefore, it should be understood that the above embodiments are for illustrative purposes only and are not intended to limit the scope. Because the scope of this invention is defined by the claims rather than the foregoing description, any changes and modifications falling within the scope and boundaries of the claims, or their equivalents, are subject to the claims.

Claims

1. A smart park policy intelligent matching calculation method based on a large model, characterized in that, The method includes: The various quantities of servers obtained when allocating different quantities of servers to each enterprise in the smart park are collected as a set of quantities. The various allocations are then iterated to obtain the sets of quantities. Obtain the operational status information of each enterprise in the smart park for each past time interval before the current moment, with each time interval having an equal duration; To build an intelligent status prediction model for smart parks, based on a customized structural design of the total number of servers provided in the computer rooms of smart parks; The intelligent state prediction model uses the proportion of each quantity in each quantity proportion set, the operational status information of each enterprise's operation program in each past time interval before the current time, the duration of each time interval, and the enterprise-related data of each enterprise to intelligently predict the proportion of each delay time in the current time interval under the configuration of the quantity proportion set. The current time interval starts from the current time. The intelligent matching quantity percentage set is determined based on the percentage of delay time corresponding to each obtained quantity percentage set.

2. The intelligent matching calculation method for smart park policies based on a large model as described in claim 1, characterized in that: The intelligent matching quantity percentage set is determined based on the delay time percentage corresponding to each acquired quantity percentage set. This includes: acquiring the delay time percentage corresponding to each quantity percentage set; taking the quantity percentage set whose mean square error of the corresponding delay time percentage is greater than a set mean square error threshold as a candidate quantity percentage set to obtain multiple candidate quantity percentage sets; taking the candidate quantity percentage set with the smallest mean of the corresponding delay time percentage as the intelligent matching quantity percentage set; and providing each quantity percentage in the intelligent matching quantity percentage set to each resident enterprise. Among them, the number of servers obtained when allocating different numbers of servers to each enterprise in the smart park is a set of number percentages. The various allocations are traversed to obtain each set of number percentages, including: the cumulative value of each number percentage in each set of number percentages is equal to 1, and each enterprise in the smart park only uses the servers provided in the smart park's computer room, and each server has the same maximum computing power, memory capacity and communication bandwidth per unit time. Among them, obtaining the operational status information of each enterprise's operating program in the smart park for each past time interval before the current moment includes: the operational status information of each enterprise's operating program in the smart park for each past time interval before the current moment is the peak computing volume, peak memory consumption, peak communication bandwidth, server allocation ratio, and latency ratio of the enterprise's operating program in the past time interval. The enterprise-related data for each resident enterprise includes the number of employees of that enterprise, the maximum computing power of the office computer per unit time, the memory capacity of the office computer, and the communication bandwidth of the office computer. All office computers used by each resident enterprise are of the same model.

3. The intelligent matching calculation method for smart park policies based on a large model as described in claim 2, characterized in that: The intelligent state prediction model for smart parks, based on a customized structure design considering the total number of servers provided in the smart park's data center, comprises: a deep neural network trained multiple times, with the number of training iterations of the deep neural network varying with the total number of servers provided in the smart park's data center. The intelligent state prediction model is a deep neural network that has been trained multiple times, and the trend of the number of training times of the deep neural network with the total number of servers provided in the smart park's computer room includes: the curve of the number of training times of the deep neural network has the same curve curvature as the curve of the total number of servers provided in the smart park's computer room. The intelligent state prediction model is a deep neural network that has been trained multiple times. The trend of the number of training times of the deep neural network with the total number of servers provided in the smart park's computer room also includes: the deep neural network has multiple hidden layers, a single output layer and a single input layer. The multiple hidden layers are located between the single input layer and the single output layer. The number of hidden layers of the deep neural network is positively correlated with the number of enterprises in the smart park. In each training iteration of the deep neural network, the percentage of delay time of a known resident company in the smart park within a certain past time interval, configured with a certain set of quantity percentages, is used as the single output of the deep neural network. The percentage of the resident company obtained in the certain set of quantity percentages, the operational status information of the resident company's business operation program corresponding to each past time interval before the certain past time interval, the duration of each time interval, and the enterprise association data corresponding to the resident company are used as the inputs of the deep neural network to complete this training.

4. The intelligent matching calculation method for smart park policies based on a large model as described in claim 3, characterized in that, Before acquiring the operational status information of each enterprise in the smart park for each past time interval before the current moment, and before the duration of each time interval is equal, the method further includes: The number of time intervals for each past time interval before the current moment is determined based on the number of companies located in the smart park. The determination of the number of time intervals for each past time interval before the current moment based on the number of enterprises in the smart park includes: the number of time intervals for each past time interval before the current moment is proportional to the number of enterprises in the smart park.

5. The intelligent matching calculation method for smart park policies based on a large model as described in claim 3, characterized in that, Before iterating through various allocations to obtain sets of server percentages for each company in the smart park, the method further includes: The data includes the total number of servers provided in the computer room of the smart park and the number of enterprises that have settled in the smart park.

6. The intelligent matching calculation method for smart park policies based on a large model as described in claim 3, characterized in that, After determining the intelligent matching quantity proportion set based on the delay time proportions corresponding to each acquired quantity proportion set, the method further includes: The determined set of intelligent matching quantity proportions is used as the optimal server allocation policy for the smart park in the current time interval and is wirelessly transmitted to the remote allocation policy service network element through the wireless communication network. The allocation policy service network element that wirelessly transmits the determined set of intelligent matching quantity proportions as the optimal server allocation policy for the smart park in the current time interval to the remote end via the wireless communication network includes: the wireless communication network is based on time division duplex communication mode or frequency division duplex communication mode.

7. The intelligent matching calculation method for smart park policies based on a large model as described in claim 3, characterized in that, After determining the intelligent matching quantity proportion set based on the delay time proportions corresponding to each acquired quantity proportion set, the method further includes: The determined set of intelligent matching quantity proportions is used as the optimal server allocation policy for the smart park in the current time interval, and the optimal server allocation policy for the smart park in the current time interval is displayed on-site in the central control room of the smart park.

8. The intelligent matching calculation method for smart park policies based on a large model as described in any one of claims 3-7, characterized in that: The various percentages obtained when allocating different numbers of servers to each enterprise in the smart park are collected as a set of percentages. The set of percentages obtained by traversing various allocations also includes: the ratio of the number of servers allocated to each enterprise in the smart park to the total number of servers provided in the smart park's data center is the percentage of the number of servers allocated to that enterprise. Among them, the operational status information of each enterprise's operating program in the smart park for each past time interval before the current moment includes the peak computing volume, peak memory consumption, peak communication bandwidth, server allocation ratio, and latency ratio of the enterprise's operating program in the past time interval. The latency ratio of the enterprise's operating program in the past time interval is the proportion of the total time of operational latency of the enterprise's operating program in the past time interval to the duration of the past time interval. The percentage of delay time in the past time interval of the enterprise operation program is the proportion of the total time of operational delay in the past time interval to the duration of the past time interval. This includes taking the cumulative value of each delay duration corresponding to each operational delay in the past time interval as the total time of operational delay in the past time interval.

9. The intelligent matching calculation method for smart park policies based on a large model as described in any one of claims 3-7, characterized in that: The intelligent state prediction model is based on the proportion of each quantity in each quantity proportion set, the operational status information of each enterprise's business operation program in each past time interval before the current time, the duration of each time interval, and the enterprise association data corresponding to each enterprise. It intelligently predicts the proportion of delay time that each enterprise's business operation program will have in the current time interval under the configuration of the quantity proportion set. The method is as follows: For each enterprise as the target enterprise, the proportion of the target enterprise in each quantity proportion set, the operational status information of the target enterprise's business operation program in each past time interval before the current time, the duration of each time interval, and the enterprise association data corresponding to the target enterprise are synchronously input into the intelligent state prediction model. The method of using an intelligent state prediction model to intelligently predict the percentage of delay time of each enterprise's business operation program in the current time interval under the configuration of the number percentage set, based on the percentage of each quantity in each quantity percentage set, the operational status information corresponding to each enterprise's business operation program in each past time interval before the current time, the duration of each time interval, and the enterprise association data corresponding to each enterprise, also includes: running the intelligent state prediction model to obtain the percentage of delay time of the target enterprise's business operation program in the current time interval under the configuration of the number percentage set. Specifically, for each resident enterprise as a target resident enterprise, the following steps are taken: the target resident enterprise's quantity percentage in each quantity percentage set, the target resident enterprise's operational procedures for each past time interval before the current moment, the duration of each time interval, and the enterprise-related data corresponding to the target resident enterprise are synchronously input into the intelligent status prediction model. Among them, the proportion of the target enterprise in each proportion set, the operational status information of the target enterprise's enterprise operation program in each past time interval before the current moment, the duration of each time interval, the enterprise association data corresponding to the target enterprise, and the proportion of delay time of the target enterprise's enterprise operation program in the current time interval under the configuration of the proportion set are all numerically normalized representations.

10. A smart park policy intelligent matching calculation system based on a large model, characterized in that, The system includes: The first parsing device is used to obtain a set of quantity proportions when allocating different quantities of servers to each enterprise in the smart park, and to traverse various allocations to obtain a set of quantity proportions. The second analysis device is used to obtain the operational status information of each enterprise in the smart park in each past time interval before the current moment, with each time interval having an equal duration. The third analytical device is used to build an intelligent state prediction model for smart parks based on the total number of servers provided in the computer room of the smart park and the customized structural design. The state prediction device is connected to the first analysis device, the second analysis device, and the third analysis device respectively. It is used to use an intelligent state prediction model to intelligently predict the proportion of delay time that the enterprise operation program of each enterprise in the current time interval will appear under the configuration of the quantity proportion set, based on the proportion of each quantity in each quantity proportion set, the operation status information corresponding to each enterprise operation program of each enterprise in the past time interval before the current time, the duration of each time interval, and the enterprise association data corresponding to each enterprise in the current time interval. The current time interval starts from the current time. The intelligent matching device, connected to the state prediction device, is used to determine the intelligent matching quantity proportion set based on the proportion of each delay time corresponding to each acquired quantity proportion set.

Images