Adaptive resource management method for intelligent empowerment center

Abstract

This invention provides an adaptive resource management method for intelligent empowerment centers, relating to the field of resource data processing technology. By automatically auditing the resource allocation and usage recorded on the blockchain, it can promptly detect abnormal access and potential security threats. Sensors and IoT modules communicate with device resources using a hybrid communication method. Furthermore, after verifying resource allocation requests, detailed logs are generated and stored in specific blockchain blocks, providing a basis for subsequent resource management analysis, auditing, and troubleshooting. This significantly enhances the security and traceability of resource allocation, ensuring stable business operations. A warning threshold is set using machine learning algorithms to predict resource demand. When the threshold is approached or exceeded, management personnel are notified in advance, and an emergency allocation plan is activated. This effectively avoids business interruptions due to resource shortages, ensures stable resource supply, and further improves the stability of resource management throughout the intelligent empowerment center.

Description

Technical Field

[0001] This invention relates to the field of resource data processing technology, specifically to an adaptive resource management method for intelligent empowerment centers. Background Technology

[0002] According to a shareable cloud resource analysis and management system disclosed in Chinese Publication No. CN118885298A, the system includes a dynamic permission management module, a modular cloud resource management module, a data analysis module, an edge computing resource management module, and an AI intelligent recommendation module. The system dynamically divides user groups and assigns access permissions to achieve resource monitoring and scheduling. Simultaneously, it employs AI intelligent recommendations to optimize resource allocation. Specifically, the system includes resource monitoring and scheduling modules that monitor and dynamically adjust resource allocation in real time, as well as edge node management to support edge computing. This invention improves the real-time performance of data sharing, effectively analyzes user needs, and achieves efficient resource utilization by integrating a multi-layered resource storage module with both high-speed and low-speed storage media.

[0003] The aforementioned patent documents and prior art have the following technical problems when used:

[0004] Problem 1: Traditional resource management lacks effective security and auditing mechanisms for resource allocation and usage records, making it prone to security issues such as data tampering and unauthorized access. When resource demand changes, it lacks effective early warning and emergency measures, which can easily lead to insufficient resource supply and affect business continuity.

[0005] Problem 2: In the resource status data collection and processing stage, there are outliers and noise in the data, which affects the accuracy of resource allocation decisions. In addition, the transmission and processing of a large amount of sensor data will put a high burden on the cloud computing platform, reduce the efficiency of resource allocation, and lack an effective coordination mechanism between various resource management modules, making it impossible to make comprehensive optimization and adjustment based on the overall business objectives and the current resource status, resulting in low resource utilization efficiency.

[0006] Thirdly, traditional resource management decisions often lack a comprehensive consideration of the relationships between resources, business, and users, as well as the effective use of historical experience, leading to unreasonable resource allocation and an inability to meet the needs of complex business scenarios. Summary of the Invention

[0007] Technical problems to be solved

[0008] To address the shortcomings of existing technologies, this invention provides an adaptive resource management method for intelligent empowerment centers, solving the following problems:

[0009] 1. Addressing the issues of insufficient security and low stability in resource allocation and usage;

[0010] 2. Addressing the issues of low allocation efficiency and low resource management collaboration efficiency caused by insufficient resource data quality;

[0011] 3. To address the problem of insufficient intelligence in resource decision-making due to incomplete consideration in resource management decisions.

[0012] Technical solution

[0013] To achieve the above objectives, the present invention provides the following technical solution: an adaptive resource management method for an intelligent empowerment center, wherein the resource management method includes the following steps:

[0014] SP1: Resource Security and Allocation Management Based on Artificial Intelligence and Blockchain: Utilizing machine learning algorithms, including supervised learning algorithms, to analyze the relationship between historical resource usage data and business types, and deep learning algorithms to mine complex patterns, resource usage patterns are learned and analyzed to predict resource demand. Simultaneously, the distributed ledger and cryptographic algorithms of blockchain are used to record resource allocation and usage. Resource allocation rules are automatically executed through smart contracts. First, the Long Short-Term Memory (LSTM) network algorithm is run to predict resource demand trends based on historical resource usage data, such as time series of server CPU and memory usage. Subsequently, when changes in resource demand trigger resource allocation requests, the Practical Byzantine Fault Tolerance (PBFT) consensus algorithm of blockchain is used to allow management nodes to initiate requests. Other nodes verify and reach consensus according to blockchain rules, ensuring that resource allocation is legal and reasonable.

[0015] SP2: Dynamic allocation and optimization of resources integrating IoT and cloud computing: Sensors and IoT modules are installed on the equipment and resources of the smart empowerment center to collect resource status data, such as server hardware parameters and network equipment bandwidth utilization. The elastic computing and virtualization technologies of cloud computing are used to achieve rapid resource allocation. Fuzzy logic algorithms are used to fuzzify the complex resource status data collected by IoT according to preset fuzzy rules, and resource adjustment suggestions are output. Then, combined with the elastic scaling algorithm of cloud computing, the allocation of cloud computing resources is automatically adjusted according to the fuzzification suggestions and real-time business needs to ensure smooth business operation.

[0016] Sp3: Resource intelligent decision-making and management combining knowledge graphs and reinforcement learning. It uses deep learning-based algorithms such as TransE and TransR to construct a resource knowledge graph, linking resources (such as servers, storage devices, etc.), business (application systems, processes), and user information, and presenting the relationships between them. At the same time, it uses reinforcement learning algorithms to allow intelligent agents to explore resource allocation strategies in the knowledge graph environment. Using reinforcement learning algorithms such as Deep Q-Network (DQN), the intelligent agent searches for resource allocation actions based on the current resource status and business needs in the knowledge graph environment. The Q value is updated based on the business execution results (such as execution time, resource utilization) as reward feedback, thereby optimizing the resource allocation decision-making strategy.

[0017] SP4: Comprehensive Coordination and Optimization Mechanism: Establish a unified resource management coordination center to aggregate resource data, decision suggestions, and execution results from different modules (based on artificial intelligence and blockchain, integrating IoT and cloud computing, and combining knowledge graphs and reinforcement learning). Based on overall business objectives and current resource status, comprehensively optimize and adjust the resource management strategies of each module to ensure that resources are allocated and used rationally and efficiently within the intelligent empowerment center, maintaining system stability, security, and sustainable development. When a specific situation occurs in a module (such as when an AI and blockchain-based module detects a security risk), the coordination center promptly notifies other modules to collaboratively address complex resource management issues.

[0018] Preferably, in the resource security and allocation management based on artificial intelligence and blockchain, the data source used by the machine learning algorithm covers resource usage records involved in all business processes within the intelligent empowerment center, including usage details of computing resources, storage resources, and network resources under different time periods and business scenarios. The resource allocation and usage recorded on the blockchain are audited regularly to check for abnormal access or potential security threats. The audit process is automatically executed through a dedicated audit smart contract, and the audit results are fed back to the resource management system in real time.

[0019] Preferably, in the dynamic allocation and optimization of resources integrating IoT and cloud computing, the communication between sensors and IoT modules and the devices and resources of the intelligent empowerment center adopts a hybrid wireless and wired communication method. For the collected resource status data, before processing by fuzzy logic algorithm, data cleaning and preprocessing are performed to remove outliers and noisy data.

[0020] Preferably, in the intelligent resource decision-making and management combining knowledge graphs and reinforcement learning, the knowledge graph update mechanism combines real-time updates with periodic updates. When resource information, business processes, or user permissions change, a partial update of the knowledge graph is triggered in real time, and the entire knowledge graph is fully updated at regular intervals to ensure the accuracy and completeness of the knowledge graph. When exploring resource allocation strategies, the agent in the reinforcement learning algorithm will consider the historical allocation success rate of resources and the importance level of the business.

[0021] Preferably, in the resource security and allocation management based on artificial intelligence and blockchain, after a resource allocation request is verified by the Practical Byzantine Fault Tolerance (PBFT) consensus algorithm of the blockchain, a detailed resource allocation log is generated, which records information such as the request initiation time, the type and quantity of resources involved, and the business or user to which the resource is allocated. These logs are stored in specific blocks of the blockchain and can be used for subsequent resource management analysis, auditing and troubleshooting.

[0022] Preferably, in the dynamic allocation and optimization of resources integrating the Internet of Things and cloud computing, when the cloud computing platform performs rapid resource allocation, in addition to considering the current business needs and the resource status data collected by the Internet of Things, it also needs to combine the resource maintenance plan and upgrade plan. For example, if a server is about to undergo maintenance, important businesses should be avoided when allocating resources to that server. At the same time, when there is a resource upgrade plan, business migration should be reasonably arranged to reduce the impact on business.

[0023] Preferably, in the comprehensive coordination and optimization mechanism, a two-way communication channel is established between the resource management coordination center and each module. On the one hand, the coordination center receives resource data, decision suggestions and execution results from each module. On the other hand, the coordination center sends optimization instructions and global resource management strategy update information to each module to ensure that each module can respond in a timely manner and adjust its own resource management behavior.

[0024] Preferably, in the resource security and allocation management based on artificial intelligence and blockchain, different levels of early warning thresholds are set for resource demand predicted by machine learning algorithms. When the predicted value of resource demand approaches or exceeds the threshold, an early warning notification is sent to the resource management personnel in advance, and an emergency resource allocation plan is activated to ensure a stable supply of resources and business continuity.

[0025] Preferably, in the dynamic allocation and optimization of resources integrating IoT and cloud computing, IoT modules and sensors are managed in regional groups according to the business characteristics and resource equipment distribution of the intelligent empowerment center. Sensor data in each region is preprocessed and preliminarily analyzed locally, and only key information and abnormal data are transmitted to the cloud computing platform, thereby reducing the amount of data transmission and the processing burden of the cloud computing platform, and improving the efficiency of dynamic resource allocation.

[0026] Beneficial effects

[0027] This invention provides an adaptive resource management method for intelligent empowerment centers. It has the following beneficial effects:

[0028] 1. This invention employs automatic auditing of resource allocation and usage records on the blockchain, enabling timely detection of abnormal access and potential security threats. Simultaneously, after verifying resource allocation requests, detailed logs are generated and stored in specific blockchain blocks, providing a basis for subsequent resource management analysis, auditing, and troubleshooting. This significantly enhances the security and traceability of resource allocation, ensuring stable business operations. Furthermore, by setting early warning thresholds for predicting resource demand using machine learning algorithms, management personnel are notified in advance and emergency allocation plans are activated when these thresholds are approached or exceeded. This effectively avoids business interruptions due to resource shortages, ensures stable resource supply, and further improves the stability of resource management throughout the intelligent empowerment center.

[0029] 2. This invention employs a hybrid communication method between sensors, IoT modules, and device resources. Furthermore, the collected data is pre-cleaned and pre-processed to improve data quality, providing a more accurate basis for subsequent resource allocation decisions. Simultaneously, IoT modules and sensors are regionally grouped and managed according to business characteristics and resource device distribution, reducing the data transmission volume and processing burden of the cloud computing platform. This significantly improves the efficiency of dynamic resource allocation. When allocating resources, the cloud computing platform combines resource maintenance and upgrade plans to avoid assigning important services to servers under maintenance. During upgrades, business migration is rationally arranged, making resource allocation more reasonable, reducing negative impacts on business operations, and ensuring smooth business operation.

[0030] 3. This invention employs a knowledge graph mechanism that combines real-time and periodic updates, ensuring its accuracy and completeness. It better presents the relationships between resources, businesses, and users. When exploring resource allocation strategies, the intelligent agent considers the historical allocation success rate of resources and the importance level of businesses, making resource allocation decisions more intelligent and more aligned with actual business needs, thereby improving the accuracy of resource utilization. A two-way communication channel is established between the resource management coordination center and each module, enabling timely receipt and transmission of information and comprehensive optimization and adjustment of resource management strategies for each module. This allows the entire resource management system to work more collaboratively, making optimal decisions based on overall business goals and the current resource status, thus enhancing the intelligence level of resource management. Attached Figure Description

[0031] Figure 1 This is a flowchart of the resource management method of the present invention;

[0032] Figure 2 Line graphs showing the input and output sequences of the Long Short-Term Memory (LSTM) network algorithm of this invention;

[0033] Figure 3 This is a diagram illustrating the fuzzy reasoning process of the present invention;

[0034] Figure 4 This is a line graph showing the change in resource utilization rate over time according to the present invention.

[0035] Figure 5 This is a line graph showing the change of the Q value of the present invention as a function of the training process. Detailed Implementation

[0036] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. Specific Implementation Example 1:

[0038] like Figure 1-5 As shown, the adaptive resource management method of the intelligent empowerment center includes the following steps:

[0039] SP1: Resource Security and Allocation Management Based on Artificial Intelligence and Blockchain: Utilizing machine learning algorithms, including supervised learning algorithms, to analyze the relationship between historical resource usage data and business types, and deep learning algorithms to mine complex patterns, resource usage patterns are learned and analyzed to predict resource demand. Simultaneously, the distributed ledger and cryptographic algorithms of blockchain are used to record resource allocation and usage. These mechanisms ensure the transparency, immutability, and security of resource allocation and usage records. Every allocation, use, and adjustment of resources is recorded on the blockchain, forming a traceable chain to prevent illegal occupation and tampering. Resource allocation rules are automatically executed through smart contracts. First, a Long Short-Term Memory (LSTM) algorithm is run, based on historical resource usage data, such as time series data of server CPU and memory usage, to predict resource demand trends. LSTM effectively handles long-term dependencies in time series data, predicting future resource demand trends based on past resource usage. For example, based on the server load at different times of the day over the past week, the server load at the same time of the next day can be predicted. Then, when changes in resource demand trigger resource allocation requests, the Practical Byzantine Fault Tolerance (PBFT) consensus algorithm of the blockchain allows the management node to initiate a request. Other nodes verify and reach a consensus according to the blockchain rules, ensuring that the resource allocation is legal and reasonable. When changes in resource demand are predicted, the management node of the smart empowerment center initiates a resource allocation request. Other nodes verify and confirm the request according to the rules and records on the blockchain, ensuring the rationality and legality of the resource allocation.

[0040] SP2: Dynamic allocation and optimization of resources integrating IoT and cloud computing: Sensors and IoT modules are installed on the equipment and resources of the smart empowerment center to collect resource status data, such as server hardware parameters and network equipment bandwidth utilization. The elastic computing and virtualization technologies of cloud computing are used to achieve rapid resource allocation. Based on the real-time data collected by IoT and business needs, the cloud computing platform can quickly create, destroy or adjust virtual machine instances, storage volumes and other resources to meet the needs of different businesses. Fuzzy logic algorithms are used to fuzzify the complex resource status data collected by IoT according to preset fuzzy rules and output resource adjustment suggestions. Then, combined with the elastic scaling algorithm of cloud computing, the allocation of cloud computing resources is automatically adjusted according to the fuzzification suggestions and real-time business needs to ensure smooth business operation.

[0041] Sp3: Resource intelligent decision-making and management combining knowledge graphs and reinforcement learning. It employs deep learning-based algorithms such as TransE and TransR to construct a resource knowledge graph, linking resources (e.g., servers, storage devices), business processes (application systems, workflows), and user information, presenting the relationships between them. The knowledge graph helps quickly discover relationships between resources and dependencies between business processes and resources, providing comprehensive knowledge support for resource management. Simultaneously, reinforcement learning algorithms are used to allow agents to explore resource allocation strategies within the knowledge graph environment. Utilizing deep Q-networks (DQN) and other reinforcement learning algorithms, the agent searches for resource allocation actions based on the current resource status and business needs within the knowledge graph environment. The Q-value is updated based on business execution results (e.g., execution time, resource utilization) as reward feedback, optimizing resource allocation decision-making strategies. When a new business task arrives, the agent selects appropriate server and storage resources for allocation based on the business-resource relationships in the knowledge graph, and adjusts subsequent resource allocation strategies based on feedback information such as task execution time and resource utilization.

[0042] SP4: Comprehensive Coordination and Optimization Mechanism: Establish a unified resource management coordination center to aggregate resource data, decision suggestions, and execution results from different modules (based on artificial intelligence and blockchain, integrating IoT and cloud computing, and combining knowledge graphs and reinforcement learning). Based on overall business objectives and resource status, comprehensively optimize and adjust the resource management strategies of each module to ensure the rational and efficient allocation and use of resources within the intelligent empowerment center, maintaining system stability, security, and sustainable development. When a specific situation occurs in a module (such as the AI ​​and blockchain module detecting security risks), the coordination center promptly notifies other modules to collaboratively address complex resource management issues. When the AI ​​and blockchain module detects resource allocation security risks, the coordination center can notify the IoT and cloud computing module to suspend certain resource allocation operations and coordinate with the knowledge graph and reinforcement learning module to reassess the current business's resource needs and allocation strategies, jointly addressing complex situations in resource management.

[0043] The entire resource management methodology further includes the following processes:

[0044] In resource security and allocation management based on artificial intelligence and blockchain, the data source used by machine learning algorithms covers resource usage records involved in all business processes within the intelligent empowerment center, including details of computing, storage, and network resource usage under different time periods and business scenarios. The resource allocation and usage recorded on the blockchain are audited regularly to check for abnormal access or potential security threats. The audit process is automatically executed through a dedicated audit smart contract, and the audit results are fed back to the resource management system in real time. In this AI- and blockchain-based resource security and allocation management, once a resource allocation request is verified through the blockchain's Practical Byzantine Fault Tolerance (PBFT) consensus algorithm, a detailed resource allocation log is generated, recording information such as the request initiation time, the type and quantity of resources involved, and the business or user allocated to it. These logs are stored in specific blocks on the blockchain and can be used for subsequent resource management analysis, auditing, and troubleshooting. For resource demand predicted by machine learning algorithms, different levels of early warning thresholds are set. When the predicted resource demand approaches or exceeds the threshold, an early warning notification is sent to resource management personnel in advance, and an emergency resource allocation plan is activated to ensure stable resource supply and business continuity.

[0045] In the dynamic allocation and optimization of resources integrating IoT and cloud computing, sensors and IoT modules communicate with the devices and resources of the smart empowerment center using a hybrid wireless and wired communication method. Before processing the collected resource status data using fuzzy logic algorithms, data cleaning and preprocessing are performed to remove outliers and noisy data. In this dynamic allocation and optimization, the cloud computing platform, when rapidly allocating resources, considers not only current business needs and the resource status data collected by IoT, but also resource maintenance and upgrade plans. For example, if a server is about to undergo maintenance, important business operations should be avoided when allocating resources to that server. Furthermore, when there is a resource upgrade plan, business migration should be rationally arranged to minimize the impact on business operations. Based on the business characteristics and resource device distribution of the smart empowerment center, IoT modules and sensors are managed in regional groups. Sensor data within each region undergoes local preprocessing and preliminary analysis, and only key information and abnormal data are transmitted to the cloud computing platform, reducing data transmission volume and the processing burden on the cloud computing platform, thereby improving the efficiency of dynamic resource allocation.

[0046] In resource intelligent decision-making and management that combines knowledge graphs and reinforcement learning, the knowledge graph update mechanism combines real-time updates with periodic updates. When resource information, business processes, or user permissions change, a partial update of the knowledge graph is triggered in real time. At regular intervals, the entire knowledge graph is fully updated to ensure its accuracy and completeness. When exploring resource allocation strategies, the agent in the reinforcement learning algorithm considers the historical allocation success rate of resources and the importance level of the business.

[0047] In the integrated coordination and optimization mechanism, a two-way communication channel is established between the resource management coordination center and each module. On the one hand, the coordination center receives resource data, decision suggestions and execution results from each module. On the other hand, the coordination center sends optimization instructions and global resource management strategy update information to each module to ensure that each module can respond in a timely manner and adjust its own resource management behavior.

[0048] The above methods enable adaptive resource management in the intelligent empowerment center. Blockchain auditing and log recording ensure the security and traceability of resource allocation, and resource demand early warning and emergency plans ensure stable resource supply, thus solving security and supply issues. Hybrid communication, data cleaning and preprocessing, and regionalized sensor grouping management improve data quality, reduce cloud computing burden, and improve resource allocation efficiency. At the same time, cloud computing combined with resource planning makes allocation more reasonable. Knowledge graph update mechanism and intelligent agents considering historical and business factors improve the intelligence and accuracy of resource decision-making. Furthermore, the two-way communication between the resource management coordination center and each module enables comprehensive optimization and adjustment, improving overall resource utilization efficiency and effectively overcoming many technical challenges in existing resource management. Specific Implementation Example 2:

[0050] like Figure 1-5 As shown, based on the content of the above specific embodiments, the following content is further disclosed:

[0051] The algorithm details for each step in resource management are as follows:

[0052] Long Short-Term Memory (LSTM) Algorithm

[0053] Mathematical formula:

[0054] i t =σ(W xi x t +W hi h t-1 +b i )

[0055] f t =σ(W xf x t +W hf ht-1 +b f )

[0056] o t =σ(W xo x t +W ho h t-1 +b o )

[0057]

[0058] h t =o t ⊙tanh(c t )

[0059] Where: t represents the time step, i t f t o t Let x represent the values ​​of the input gate, forget gate, and output gate at time step, respectively. σ is the sigmoid function, which compresses the values ​​between 0 and 1. W and b are the weight matrix and bias term, respectively. In the subscript, x represents the input, h represents the hidden state, and i, f, o, and c correspond to the input gate, forget gate, output gate, and cell state, respectively. t h represents the input at time step. t-1 This represents the hidden state of the previous time step. c represents the candidate cell state at time step t. t denoted by , ⊙ represents the cell state at time step t, tanh represents the hyperbolic tangent function;

[0060] Implementation steps:

[0061] First, based on the current input x t The hidden state h of the previous time step t-1 Calculate the input gate i respectively t Forgotten Gate t Output gate o t and candidate cell status

[0062] Then, through the forget gate f t Controlling the cell state c in the previous time step t-1 The degree of forgetting, and simultaneously through input gate i t Controlling new candidate cell states The degree of inclusion is used to obtain the cell state c at the current time step. t ;

[0063] Finally, through the output gate o t Controlling cell state c t For hidden state h tThe degree of influence, output the hidden state h at the current time step. t .

[0064] Operating logic: LSTM controls the flow of information through input gates, forget gates, and output gates. The input gate determines which new information needs to be stored in the cell state, the forget gate determines which old information needs to be forgotten, and the output gate determines which information needs to be output. This effectively handles long-term dependencies in time series data.

[0065] For resource demand forecasting, LSTM effectively handles long-term dependencies in time-series data such as server CPU and memory usage. Based on past resource usage patterns, it accurately predicts future resource demand trends, providing a reliable basis for resource allocation.

[0066] Practical Byzantine Fault Tolerance (PBFT) Consensus Algorithm

[0067] Implementation steps: The client sends a request to the master node, the master node broadcasts the request to all replica nodes, the replica nodes execute the request and send a reply to other nodes, each node collects the replies from other nodes and determines whether a consensus has been reached according to certain rules. If a consensus is reached, the node executes the request and replies to the client.

[0068] Operating Logic: The PBFT algorithm achieves consensus through communication and negotiation among multiple replica nodes. When the management node initiates a resource allocation request, other nodes verify and reach a consensus according to blockchain rules, ensuring the legality and rationality of resource allocation. During the resource allocation process, the PBFT consensus algorithm ensures the legality and rationality of resource allocation. Every allocation, use, and adjustment of resources is recorded on the blockchain, forming a traceable chain to prevent illegal occupation and tampering of resources.

[0069] Fuzzy logic algorithm

[0070] Suppose we have input variables x1, x2, ..., xn n Examples include server hardware parameters and network device bandwidth utilization. The output variable is y, which provides resource adjustment suggestions.

[0071] Define fuzzy set A i，j Indicates the input variable x i The j-th fuzzy state, B k This represents the k-th fuzzy state of the output variable y;

[0072] Fuzzy rules are usually expressed as: if x1 is And x2 is Eye...and x n yes So y is B k ;

[0073] Implementation steps

[0074] Fuzzification: For each input variable x i Determine its position in each fuzzy set A i，j In fuzzy sets, membership degrees can be calculated. For example, if the fuzzy sets of input variable X1 are "low", "medium", and "high", membership functions, such as triangular membership functions and trapezoidal membership functions, can be used to calculate the degree to which it belongs to each fuzzy set. This indicates that x1 belongs to the fuzzy set a. 1，j Membership degree;

[0075] Fuzzy reasoning: Calculate the excitation intensity of each rule based on fuzzy rules. For the rule "If x1 is...",... And x2 is And...and x n yes Then y is b k Its excitation intensity is:

[0076]

[0077] Deblurring: Using the centroid method, calculate the precise value of the output variable y. Let y k It outputs the fuzzy set B. k If the representative value is the center value of the fuzzy set, then the precise value of the output variable Y is:

[0078]

[0079] Operational logic: First, the input resource status data is fuzzified, converting precise numerical values ​​into membership degrees of fuzzy concepts. Then, inference is performed based on preset fuzzy rules to determine the activation intensity of each rule. Finally, defuzzification transforms the fuzzy output into precise resource adjustment suggestions. This approach can handle complex resource status data, especially those with uncertainty and fuzziness. Through fuzzification and fuzzy inference, human experience and knowledge can be transformed into fuzzy rules, better adapting to actual conditions. The output resource adjustment suggestions are more flexible and reasonable, improving the efficiency and accuracy of dynamic resource allocation.

[0080] Elastic scaling algorithms in cloud computing

[0081] Let the total amount of resources be R. total The current amount of resources used is R. used Resource utilization rate

[0082] Define the upper limit threshold as U max The lower threshold is U minWhen resource utilization exceeds the upper limit threshold, the amount of resource ΔR needs to be increased. up When resource utilization falls below the lower threshold, the amount of resources ΔR needs to be reduced. down ;

[0083] Implementation steps: Regularly or in real time monitor resource usage and calculate resource utilization rate U;

[0084] If U>U max Then calculate the increased resource amount ΔR. up =R total ×(UU max Based on the calculation results, new virtual machine instances, storage volumes, and other resources are created.

[0085] If U min Then calculate the amount of resources reduced, ΔR. down =R total ×(U min -U), based on the calculation results, destroy or adjust some virtual machine instances, storage volumes, and other resources;

[0086] If U m in≤U≤U max If so, no resource adjustment will be made;

[0087] Operational Logic: By continuously monitoring resource utilization and comparing it with preset thresholds, the system determines whether to elastically scale resources based on the comparison results. When resource utilization is too high, resources are added to meet business needs; when resource utilization is too low, resources are reduced to lower costs. It can automatically adjust resource allocation according to changes in business needs, improving resource utilization. During peak business periods, it ensures system performance and stability; during off-peak periods, it avoids resource waste and reduces operating costs. Simultaneously, by combining real-time data collected through the Internet of Things (IoT) with business requirements, it can more accurately allocate resources, ensuring smooth business operation.

[0088] Intelligent resource decision-making and management combining knowledge graphs and reinforcement learning

[0089] Knowledge graph construction algorithms such as TransE and TransR

[0090] TransE:

[0091] f r (h, t) = ||h + rt||2

[0092] Where h represents the head entity, t represents the tail entity, r represents the relation, and ||.||2 represents the L2 norm;

[0093] TransR:

[0094] h​r =M r h, t r =M r t, f r (h, t) = ||h r +rt r ||2

[0095] Among them, M r It is a projection matrix specific to the relationship;

[0096] Implementation steps:

[0097] For TransE: Given a set of triples (head entity, relation, tail entity), learn vector representations of entities and relations by minimizing the objective function;

[0098] For TransR: First, entities are projected into a relation-specific space, and then vector representations of entities and relations are learned by minimizing the objective function;

[0099] Operational Logic: Algorithms such as TransE and TransR construct resource knowledge graphs by learning vector representations of entities and relationships. These graphs connect resources (such as servers and storage devices), business processes (application systems and workflows), and user information, presenting the relationships between them. This helps to quickly discover relationships between resources and dependencies between business processes and resources, providing comprehensive knowledge support for resource management.

[0100] Reinforcement learning algorithm (Deep Q-Network, DQN)

[0101] Mathematical formula:

[0102] Q(s,a)=Q(s,a)+α[r+γmax a′ Q(s′,a′)-Q(s,a)]

[0103] Where Q(s, a) represents the Q-value of taking action α in state s, α is the learning rate, r is the reward, γ is the discount factor, and s ′ The next state is a. ′ It is the optimal action in the next state.

[0104] Implementation steps:

[0105] Initialize the Q-network and the target Q-network.

[0106] For each time step:

[0107] Choose action: Based on the current state s, use the ∈-greedy strategy to choose action a;

[0108] Perform action: Perform action a in the environment, and receive reward r and the next state s. ′;

[0109] Storage experience: (s, a, r, s) ′ Stored in the experience replay pool.

[0110] Training the Q-network: Randomly sample a batch of experiences from the experience replay pool, calculate the target Q value, and update the parameters of the Q-network using gradient descent.

[0111] Regularly update the target Q network: copy the parameters of the Q network into the target Q network.

[0112] Operational Logic: The agent explores resource allocation strategies within the knowledge graph environment. Based on the current resource status and business requirements, it searches for resource allocation actions. It updates its Q-value using business execution results (such as execution time and resource utilization) as reward feedback, optimizing resource allocation decision-making strategies. Through continuous learning and optimization, the agent can select appropriate servers and storage resources for allocation based on the relationship between business and resources in the knowledge graph. Furthermore, it adjusts subsequent resource allocation strategies based on feedback information such as task execution time and resource utilization, improving the efficiency and accuracy of resource allocation. Specific Implementation Example 3:

[0114] like Figure 1-5 As shown, based on the content of the above specific embodiments, the following content is further disclosed:

[0115] When performing dynamic adaptive adjustments for resource management, the dynamic flow of data is as follows:

[0116] The data sources, data flow stages and processing methods, data output and destination for resource security and allocation management based on artificial intelligence and blockchain are as follows:

[0117] Data source: Covers resource usage records for all business processes within the intelligent empowerment center, including details of computing, storage, and network resource usage in different time periods and business scenarios;

[0118] Data flow and processing methods: Machine learning algorithms, including supervised learning algorithms, are used to analyze the relationship between historical resource usage data and business types, and deep learning algorithms are used to mine complex patterns. Resource usage patterns are learned and analyzed to predict resource demand. Specifically, the Long Short-Term Memory (LSTM) network algorithm is first run to predict resource demand trends based on historical resource usage data (such as time series of server CPU and memory usage). At the same time, the distributed ledger and encryption algorithms of blockchain are used to record resource allocation and usage. Every allocation, use, and adjustment of resources is recorded on the blockchain, forming a traceable chain to prevent illegal occupation and tampering of resources. Resource allocation rules are automatically executed through smart contracts. When changes in resource demand trigger resource allocation requests, the Practical Byzantine Fault Tolerance (PBFT) consensus algorithm of the blockchain allows the management node to initiate a request. Other nodes verify and reach a consensus according to the blockchain rules to ensure that resource allocation is legal and reasonable.

[0119] Data Output and Destination: The system outputs resource demand forecast results. When a change in resource demand is predicted, the management node of the intelligent empowerment center initiates a resource allocation request, generating detailed resource allocation logs. These logs record information including the request initiation time, the type and quantity of resources involved, and the business or user allocated to them. These logs are stored in specific blocks on the blockchain and can be used for subsequent resource management analysis, auditing, and troubleshooting. The resource allocation and usage records on the blockchain are audited regularly to check for abnormal access or potential security threats. Audit results are fed back to the resource management system in real time. For resource demands predicted by machine learning algorithms, different levels of early warning thresholds are set. When the predicted resource demand approaches or exceeds the threshold, an early warning notification is sent to resource management personnel, and an emergency resource allocation plan is activated.

[0120] The data sources, data flow stages and processing methods, data output and destination for the dynamic allocation and optimization of resources integrating the Internet of Things and cloud computing are as follows:

[0121] Data source: Sensors and IoT modules are installed on the equipment and resources of the smart empowerment center to collect resource status data, such as server hardware parameters and network equipment bandwidth utilization.

[0122] Data Flow and Processing: Sensors and IoT modules collect data from the devices and resources of the Smart Empowerment Center using a hybrid wireless and wired communication method. Before processing the collected resource status data using fuzzy logic algorithms, data cleaning and preprocessing are performed to remove outliers and noise. Based on the business characteristics and resource device distribution of the Smart Empowerment Center, IoT modules and sensors are managed in regional groups. Sensor data within each region undergoes local preprocessing and preliminary analysis. Only key information and abnormal data are transmitted to the cloud computing platform, reducing data transmission volume and the platform's processing burden. When rapidly allocating resources, the cloud computing platform considers not only current business needs and the resource status data collected by the IoT, but also resource maintenance and upgrade plans. Fuzzy logic algorithms are used to fuzzify the complex resource status data collected by the IoT according to preset fuzzy rules, outputting resource adjustment suggestions. Then, combined with the elastic scaling algorithm of cloud computing, the allocation of cloud computing resources is automatically adjusted based on the fuzzification suggestions and real-time business needs to ensure smooth business operation.

[0123] Data output and destination: Output resource adjustment suggestions and cloud computing resource allocation results to meet the needs of different businesses and ensure smooth business operation.

[0124] The data sources, data flow stages and processing methods, data outputs and destinations for resource-based intelligent decision-making and management combining knowledge graphs and reinforcement learning are shown below:

[0125] Data source: Resource knowledge graphs are constructed using deep learning-based algorithms such as TransE and TransR, linking resources (such as servers, storage devices, etc.), business (application systems, processes), and user information.

[0126] Data flow stages and processing methods: The knowledge graph update mechanism combines real-time and periodic updates. When resource information, business processes, or user permissions change, a partial update of the knowledge graph is triggered in real time. At regular intervals, the entire knowledge graph is fully updated to ensure its accuracy and completeness. Reinforcement learning algorithms are used to allow agents to explore resource allocation strategies in the knowledge graph environment. Reinforcement learning algorithms such as Deep Q-Network (DQN) are used to allow agents to search for resource allocation actions based on the current resource status and business needs in the knowledge graph environment. The Q-value is updated based on the business execution results (such as execution time and resource utilization) as reward feedback, thereby optimizing the resource allocation decision strategy.

[0127] Data output and destination: When a new business task arrives, the agent selects appropriate server and storage resources for allocation based on the relationship between business and resources in the knowledge graph, and adjusts the subsequent resource allocation strategy based on feedback information such as task execution time and resource utilization.

[0128] The data sources, data flow stages and processing methods, data outputs and destinations in the integrated coordination and optimization mechanism steps are shown below:

[0129] Data sources: We receive resource data, decision suggestions, and execution results from three modules: artificial intelligence and blockchain, the integration of the Internet of Things and cloud computing, and the combination of knowledge graphs and reinforcement learning.

[0130] Data flow and processing methods: A unified resource management coordination center is established to aggregate resource data, decision-making suggestions, and execution results from different modules. Based on overall business objectives and current resource status, the resource management strategies of each module are comprehensively optimized and adjusted. When a specific situation arises in a particular module, the coordination center promptly notifies other modules to collaboratively address complex resource management issues.

[0131] Data output and destination: Send optimization instructions and global resource management strategy update information to each module to ensure that each module can respond in a timely manner and adjust its own resource management behavior, so as to ensure that resources are allocated and used reasonably and efficiently within the intelligent empowerment center, and maintain the system's stability, security and sustainable development. Specific Implementation Example 4:

[0133] like Figure 1-5 As shown, based on the content of the above specific embodiments, the following content is further disclosed:

[0134] In practical use, the hardware modules corresponding to each of the above steps are as follows:

[0135] Resource security and allocation management based on artificial intelligence and blockchain has the following hardware architecture:

[0136] server

[0137] Function: Store and process resource usage records for all business processes within the intelligent empowerment center, run machine learning algorithms to predict resource demand, and participate as a node in the blockchain network in the verification and consensus process of resource allocation.

[0138] Performance requirements: It needs to have high computing power and storage capacity to process large amounts of historical resource usage data and run complex machine learning algorithms. At the same time, in order to participate in the blockchain network, it needs to have a stable network connection and a certain degree of encrypted computing capability.

[0139] Blockchain node devices

[0140] Function: To form a blockchain network, record resource allocation and usage, and ensure the transparency, immutability, and security of these records. It automatically executes resource allocation rules through smart contracts, verifying and reaching consensus on resource allocation requests.

[0141] Performance requirements: Reliable storage capacity and network connectivity are required to ensure the integrity and consistency of blockchain data. At the same time, a certain level of cryptographic computing capability is required to ensure the security of the blockchain.

[0142] The dynamic allocation and optimization of resources integrating the Internet of Things and cloud computing has the following hardware structure:

[0143] Sensors and IoT modules

[0144] Function: Installed on devices and resources in the intelligent empowerment center, it collects resource status data, such as server hardware parameters and network device bandwidth utilization.

[0145] Performance requirements: High-precision data acquisition and stable communication capabilities are required to ensure the accuracy and reliability of the collected data. Additionally, to adapt to different devices and environments, it needs to be compact and low-power.

[0146] Cloud computing server

[0147] Function: Leveraging elastic computing and virtualization technologies, resources can be rapidly allocated. Based on real-time data collected from the Internet of Things and business needs, resources such as virtual machine instances and storage volumes can be quickly created, destroyed, or adjusted to meet the needs of different businesses.

[0148] Performance requirements: It needs to possess powerful computing capabilities and storage capacity to support large-scale cloud computing resource allocation. Simultaneously, it needs to have high reliability and high availability to ensure business continuity.

[0149] The resource intelligent decision-making and management system that combines knowledge graphs and reinforcement learning has the following hardware architecture:

[0150] server

[0151] Function: Store and process resource knowledge graph data, and run reinforcement learning algorithms to make resource allocation decisions.

[0152] Performance requirements: High computing power and storage capacity are required to process complex knowledge graph data and run reinforcement learning algorithms. A stable network connection is also necessary for data interaction with other modules.

[0153] The integrated coordination and optimization mechanism has the following hardware structure:

[0154] Resource Management Coordination Center Server

[0155] Function: To establish a unified resource management and coordination center, to aggregate resource data, decision-making suggestions and execution results from different modules, and to comprehensively optimize and adjust the resource management strategies of each module.

[0156] Performance requirements: The system needs powerful computing capabilities and storage capacity to handle large amounts of resource data and perform complex optimization calculations. Simultaneously, it needs high reliability and availability to ensure the stable operation of the resource management and coordination center.

[0157] In summary, the intelligent empowerment center adaptive resource management method requires the collaborative work of various hardware devices in practical use, including servers, blockchain node devices, sensors and IoT modules, cloud computing servers, etc. These hardware devices need to have different performance requirements to meet the needs of each step in the method. At the same time, in order to ensure the stable operation of the system and efficient resource management, these hardware devices need to be configured and managed in a reasonable manner.

[0158] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising a reference structure" does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0159] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An adaptive resource management method for intelligent empowerment centers, characterized by: The resource management method includes the following steps: SP1: Resource Security and Allocation Management Based on Artificial Intelligence and Blockchain: Utilizing machine learning algorithms to learn and analyze resource usage patterns to predict resource demand, while using blockchain's distributed ledger and cryptographic algorithms to record resource allocation and usage. Resource allocation rules are automatically executed through smart contracts. First, the Long Short-Term Memory (LSTM) algorithm is run to predict resource demand trends based on historical resource usage data. Then, when changes in resource demand trigger resource allocation requests, the Practical Byzantine Fault Tolerance (PBFT) consensus algorithm of the blockchain allows management nodes to initiate requests. Other nodes verify and reach consensus according to blockchain rules, ensuring that resource allocation is legal and reasonable. SP2: Dynamic allocation and optimization of resources integrating IoT and cloud computing: Sensors and IoT modules are installed on the equipment and resources of the smart empowerment center to collect resource status data. The elastic computing and virtualization technologies of cloud computing are used to realize rapid resource allocation. Fuzzy logic algorithms are used to fuzzify the complex resource status data collected by IoT according to preset fuzzy rules, and resource adjustment suggestions are output. Then, combined with the elastic scaling algorithm of cloud computing, the allocation of cloud computing resources is automatically adjusted according to the fuzzification suggestions and real-time business needs to ensure smooth business operation. Sp3: Resource intelligent decision-making and management combining knowledge graphs and reinforcement learning. It uses deep learning-based algorithms such as TransE and TransR to construct a resource knowledge graph, linking resources, business, and user information and presenting the relationships between them. At the same time, it uses reinforcement learning algorithms to allow agents to explore resource allocation strategies in the knowledge graph environment. Using reinforcement learning algorithms such as Deep Q-Network (DQN), the agent searches for resource allocation actions based on the current resource status and business needs in the knowledge graph environment. The Q value is updated based on the business execution results as reward feedback to optimize the resource allocation decision strategy. SP4: Comprehensive Coordination and Optimization Mechanism: Establish a unified resource management coordination center to aggregate resource data, decision suggestions, and execution results from different modules. Based on overall business objectives and current resource status, comprehensively optimize and adjust the resource management strategies of each module to ensure that resources are allocated and used rationally and efficiently within the intelligent empowerment center, maintaining system stability, security, and sustainable development. When a specific situation occurs in a certain module, the coordination center promptly notifies other modules to collaboratively address complex resource management issues.

2. The adaptive resource management method for the intelligent empowerment center according to claim 1, characterized in that: In the aforementioned resource security and allocation management based on artificial intelligence and blockchain, the data source used by the machine learning algorithm covers resource usage records involved in all business processes within the intelligent empowerment center, including usage details of computing resources, storage resources, and network resources under different time periods and business scenarios. The resource allocation and usage recorded on the blockchain are audited regularly to check for abnormal access or potential security threats. The audit process is automatically executed through a dedicated audit smart contract, and the audit results are fed back to the resource management system in real time.

3. The adaptive resource management method for the intelligent empowerment center according to claim 1, characterized in that: In the dynamic allocation and optimization of resources that integrates the Internet of Things and cloud computing, the communication between sensors and IoT modules and the devices and resources of the intelligent empowerment center adopts a hybrid wireless and wired communication method. Before processing the collected resource status data with fuzzy logic algorithms, data cleaning and preprocessing are performed to remove outliers and noisy data.

4. The adaptive resource management method for the intelligent empowerment center according to claim 1, characterized in that: In the resource intelligent decision-making and management that combines knowledge graphs and reinforcement learning, the knowledge graph update mechanism combines real-time updates with periodic updates. When resource information, business processes, or user permissions change, a partial update of the knowledge graph is triggered in real time. The entire knowledge graph is fully updated at regular intervals to ensure its accuracy and completeness. When exploring resource allocation strategies, the agent in the reinforcement learning algorithm considers the historical allocation success rate of resources and the importance level of the business.

5. The adaptive resource management method for the intelligent empowerment center according to claim 1, characterized in that: In the aforementioned resource security and allocation management based on artificial intelligence and blockchain, after a resource allocation request is verified by the Practical Byzantine Fault Tolerance (PBFT) consensus algorithm of the blockchain, a detailed resource allocation log is generated, recording information such as the request initiation time, the type and quantity of resources involved, and the business or user to which the resource is allocated. These logs are stored in specific blocks of the blockchain and can be used for subsequent resource management analysis, auditing, and troubleshooting.

6. The adaptive resource management method for the intelligent empowerment center according to claim 1, characterized in that: In the dynamic allocation and optimization of resources that integrates the Internet of Things and cloud computing, when the cloud computing platform performs rapid resource allocation, in addition to considering current business needs and resource status data collected by the Internet of Things, it also needs to combine resource maintenance and upgrade plans.

7. The adaptive resource management method for the intelligent empowerment center according to claim 1, characterized in that: In the aforementioned integrated coordination and optimization mechanism, a two-way communication channel is established between the resource management coordination center and each module. On the one hand, the coordination center receives resource data, decision suggestions, and execution results from each module. On the other hand, the coordination center sends optimization instructions and global resource management strategy update information to each module to ensure that each module can respond in a timely manner and adjust its own resource management behavior.

8. The adaptive resource management method for the intelligent empowerment center according to claim 1, characterized in that: In the resource security and allocation management based on artificial intelligence and blockchain, different levels of early warning thresholds are set for resource demand predicted by machine learning algorithms. When the predicted value of resource demand approaches or exceeds the threshold, an early warning notification is sent to resource management personnel in advance, and an emergency resource allocation plan is activated to ensure stable resource supply and business continuity.

9. The adaptive resource management method for the intelligent empowerment center according to claim 1, characterized in that: In the dynamic allocation and optimization of resources that integrates the Internet of Things (IoT) and cloud computing, IoT modules and sensors are managed in regional groups according to the business characteristics and resource equipment distribution of the intelligent empowerment center. Sensor data in each region is preprocessed and preliminarily analyzed locally, and only key information and abnormal data are transmitted to the cloud computing platform, thereby reducing the amount of data transmission and the processing burden of the cloud computing platform and improving the efficiency of dynamic resource allocation.

Images