Emergency command converged communication system resource dynamic scheduling method based on artificial intelligence

Abstract

The invention provides an emergency command converged communication system resource dynamic scheduling method based on artificial intelligence. The method comprises the following steps: S1, acquiring a plurality of sensors and log data of an emergency command system; constructing a dynamic graph neural network optimization communication topology, and selecting an optimal communication path; combining resource prediction and the optimized communication topology to formulate a global optimal resource scheduling strategy, and generating an optimal resource allocation scheme for each task; the resource pool comprises available computing resources and communication bandwidth; dynamically adjusting resource allocation and a communication path in a task execution process, and generating task execution feedback of a corresponding task; the system state is monitored in real time, and resource prediction, topology adjustment and scheduling strategies are optimized based on execution error feedback of task execution feedback. According to the method, efficient and intelligent dynamic resource scheduling is realized in an emergency command converged communication system through resource prediction of self-supervised learning, self-adaptive topological optimization of a variable topological graph neural network and global scheduling of reinforcement learning driving.

Description

Technical Field

[0001] The present invention belongs to the field of emergency command, and particularly relates to a method for dynamically scheduling resources of an emergency command integrated communication system based on artificial intelligence. Background Art

[0002] In a modern emergency command system, the integrated communication system plays a crucial role, capable of integrating various communication means (such as satellite communication, shortwave radio, private network, public network, etc.) to achieve cross-regional and cross-departmental information sharing and unified scheduling. However, in emergencies (such as earthquakes, floods, terrorist attacks, military exercises, etc.), there are still many technical bottlenecks in the resource scheduling of traditional integrated communication systems, mainly reflected in the following aspects:

[0003] Firstly, it is difficult to predict resource requirements in advance, resulting in scheduling lags or resource waste. The occurrence of emergency events is often accompanied by drastic changes in communication requirements, and traditional static configuration or rule-driven resource scheduling methods are difficult to adapt to complex emergencies. For example, in sudden natural disasters, the communication requirements in the affected areas surge, and it is often difficult for the scheduling center to perceive these requirements in a timely manner, resulting in a lag in the allocation of communication resources and thus affecting the rescue efficiency. In addition, current mainstream resource scheduling methods mostly adopt an experience-based rule approach, such as presetting a fixed communication topology or allocating resources according to predefined scheduling priorities. Although this approach can play a certain role under normal circumstances, when the scale, type, and scope of influence of the event are unpredictable, the preset rules often cannot adapt to new requirements, easily leading to unbalanced resource scheduling, with communication resources being tense in some areas while resources are idle in other areas.

[0004] Secondly, the traditional communication topology structure is rigid and cannot adapt to dynamic changes in emergencies. The physical and logical topology structures of the integrated communication system may change drastically in emergencies. For example, some communication nodes may fail due to damage or overload, and some communication links may become unavailable due to interference. Traditional networks usually adopt a fixed topology or rely only on static backup links, making it difficult to flexibly adjust the network structure in emergencies. For example, in a battlefield or disaster area, some communication base stations may be damaged, and the existing system is difficult to quickly reconfigure the optimal communication path, resulting in some rescue units or command nodes losing contact, seriously affecting the effectiveness of scheduling and command. Even if some systems support a certain degree of link switching and load balancing, most still rely on static preset rules and lack intelligent optimization capabilities, resulting in a slow process of adjusting communication links and affecting real-time performance.

[0005] Furthermore, the resource scheduling lacks the ability of global optimization and it is difficult to balance communication stability and efficient resource utilization. Traditional converged communication systems usually allocate resources according to a single dimension (such as bandwidth, computing resources), without comprehensively considering various factors such as the priorities of different tasks, bandwidth requirements, and computing loads, resulting in poor global optimality of scheduling decisions. For example, in the same emergency mission, some units may require high bandwidth for video transmission, while others mainly conduct low-bandwidth voice scheduling. However, traditional scheduling systems are difficult to intelligently adjust resource allocation, leading to high-priority tasks not being fully supported, while low-priority tasks occupy too many resources. In addition, with the wide application of edge computing and cloud computing, the complexity of resource scheduling further increases, and existing scheduling systems are difficult to achieve intelligent coordination between edge computing nodes and cloud resources, resulting in inefficient utilization of computing resources in emergency situations.

[0006] In summary, the current converged communication system faces core problems such as insufficient resource demand prediction ability, rigidity of communication topology and difficulty in adaptive adjustment, and lack of global optimization in resource scheduling in emergency command and dispatch, which seriously affect the response speed and stability of the emergency command system. Therefore, there is an urgent need for a dynamic scheduling method that can intelligently predict resource demands, adaptively adjust the communication topology, and have the ability of global optimization to improve the reliability and scheduling efficiency of the emergency command system. Summary of the Invention

[0007] The purpose of the present invention is to propose a dynamic resource scheduling method for an emergency command converged communication system based on artificial intelligence. Through self-supervised learning-based resource prediction, adaptive topology optimization of variable topology graph neural networks, and global scheduling driven by reinforcement learning, efficient, stable, and intelligent dynamic resource scheduling is achieved in the emergency command converged communication system.

[0008] To achieve the above purpose, the present invention provides a dynamic resource scheduling method for an emergency command converged communication system based on artificial intelligence, and the method includes the following steps:

[0009] S1. Obtain multiple sensor and log data of the emergency command system, use contrastive learning based on task similarity to construct unlabeled learning tasks, and combine with a time series Transformer to automatically learn the resource consumption patterns in historical data and generate resource predictions for future time periods;

[0010] S2. Construct a dynamic graph neural network to optimize the communication topology, dynamically and adaptively adjust the communication topology according to the resource prediction combined with reinforcement learning, and select the optimal communication path;

[0011] S3. Obtain a resource pool, formulate a globally optimal resource scheduling strategy in combination with the resource prediction and the optimized communication topology, and generate an optimal resource allocation plan for each task; the resource pool includes available computing resources and communication bandwidth;

[0012] S4. Calculate the optimal resource allocation scheme for each task and the optimized communication topology based on the node status information of the emergency command system, and intelligently match the computing tasks with the computing nodes, and dynamically adjust the resource allocation and communication paths during the task execution process to generate the task execution feedback for the corresponding tasks; the node status information includes: maximum computing capacity, current load, maximum bandwidth, and current bandwidth occupancy;

[0013] S5. Monitor the system status in real time, and optimize the resource prediction, topology adjustment, and scheduling strategy based on the execution error feedback of the task execution feedback.

[0014] Preferably, the data sources of the multiple sensors and log data include: network monitoring devices, computing node management systems, and device health monitoring systems;

[0015] The S1 further includes: adopting a sliding window method to slice the data of consecutive T w time steps to form a historical sequence; normalizing data with different dimensions so that they can be learned on the same scale.

[0016] Preferably, the contrastive learning based on task similarity constructs unlabeled learning tasks, including:

[0017] Using a transformer model to calculate the historical resource usage patterns of tasks, generating task feature vectors to represent the characteristics of tasks, and extracting the key resource usage patterns of tasks

[0018] Based on the normalized cross-entropy analysis, make the feature vectors of similar tasks closer and the feature vectors of different tasks farther away;

[0019] Among them, the construction of unlabeled learning tasks, combined with the time series Transformer, automatically learns the resource consumption patterns in historical data and generates resource predictions for future time periods, including:

[0020] Predict future resource requirements based on historical data, use the time series Transformer for modeling, use the task feature vector as the input, and output the predicted values of bandwidth requirements, computing resources, and device health status for future time steps; the time series Transformer uses weighted mean square error to calculate the prediction error to ensure that the prediction error of key resources has a greater impact.

[0021] Preferably, the S2 specifically includes:

[0022] Construct a dynamic communication graph Initialize the node status, where the dynamic communication graph includes: a node set representing communication devices; an edge set ε = {eij} represents the communication link between nodes; weight matrix Record the link quality;

[0023] Based on the dynamic communication graph Adopt a dynamic graph neural network to model the communication topology, iteratively propagate link information, and enable each node to perceive the global state; through multiple rounds of iterative calculations, each node dynamically adjusts its own connections according to the states of its neighbor nodes, calculates the retention or deletion probability of each link, decides whether to retain or delete the link, and establishes new links between high-demand nodes to generate an optimized communication topology;

[0024] On the optimized communication topology, select the lowest-delay path based on the bandwidth requirements of the task, and ensure that the links in the path have sufficient bandwidth margin to generate the path cost of the communication topology.

[0025] Preferably, the calculation of the retention or deletion probability of each link and the decision on whether to retain or delete the link are specifically as follows:

[0026]

[0027] When p ij < τ, delete the link;

[0028] When p ij > 1 - τ, add a link between high-demand nodes to form a new network topology G t ';

[0029] Among them, p ij is the retention or deletion probability of each link, w ij is the weight of edge e ij , w j′j' is the weight of edge e i'j' , i', j' are edges i' and j', and τ is a set threshold;

[0030] The weight w ij of edge e ij is calculated as:

[0031]

[0032] Among them, d ij is the physical distance of the link, r ij is the link reliability, l ij is the link delay; Combined with the predicted bandwidth requirements, guide the topology to optimize towards high-demand nodes; α1, α2, α3, α4 are hyperparameters to balance various factors.

[0033] Preferably, the resource scheduling strategy satisfies the task priority; the task priority is obtained through weighted linear analysis based on the task urgency, the proportion of the task's bandwidth requirement, the scheduling priority of the task in case of resource shortage, and the quality of the adjacent link of the current task's communication topology.

[0034] Preferably, the resource scheduling strategy is optimized by combining the task priority with the constraint that the available computing resources and communication bandwidth do not exceed the maximum capacity and the goal of maximizing resource utilization; the scheduling loss function of the resource scheduling strategy is:

[0035]

[0036] where and represent the matching degree between the allocated resources and the requirements; constrains the bandwidth imbalance of adjacent tasks to ensure that the link will not be overloaded due to a single task; λ is the regularization coefficient, which controls the trade-off between resource balance and priority.

[0037] Preferably, the S4 includes:

[0038] Based on the computing resources and bandwidth required by the task, the maximum computing resources and bandwidth capacity of the computing node, the current computing resources and bandwidth occupancy of the node, and the link score associated with the task, calculate the fitness of the task to be executed on the computing node; where the link score associated with the task is the weight of the corresponding edge.

[0039] The task selects the computing node with the highest fitness to execute.

[0040] During the execution of the task, the load of each computing node changes over time. If an overload occurs, the task allocation strategy needs to be adjusted; if the communication link on which the task depends becomes congested, the communication path of the task needs to be dynamically adjusted.

[0041] After each task is executed, record the actual consumption values of its computing resources and communication bandwidth, compare them with the predicted values, calculate the error. If the error exceeds the set preset threshold, return the resource prediction value of adjustment step S1 to improve the accuracy of the next round of task execution; if the error continues to exceed the standard, adjust the resource scheduling strategy of step S3 to improve the fitness calculation of task execution and optimize the matching rule between computing tasks and computing nodes.

[0042] Preferably, the path selection and switching rule for adjusting the task allocation strategy when an overload occurs is:

[0043] Set the maximum load threshold τ of the computing node C , if a certain computing node 's current computing load exceeds this threshold, then the task allocation needs to be adjusted:

[0044]

[0045] Among them, represents the set of tasks currently assigned to the node ; is the computing resource required for task T i ;

[0046] If the load exceeds the standard, some tasks will be preferentially migrated to the node v k with surplus computing resources, or the computing priority of some tasks will be reduced to make them wait for execution;

[0047] The path selection and switching rules for dynamically adjusting the communication path of the task when the communication link on which the task depends is congested are as follows:

[0048] If the link bandwidth on the new path is sufficient, directly switch;

[0049] If the bandwidth is insufficient, reduce the bandwidth occupancy of low-priority tasks to give priority to high-priority tasks.

[0050] Preferably, the execution error is the bandwidth allocation error and the computing resource allocation error. If the bandwidth allocation error or the computing resource allocation error exceeds the set threshold, it indicates that the resource prediction is inaccurate and the prediction parameters of the resource prediction need to be optimized;

[0051] The S5 further includes:

[0052] If the link bandwidth utilization rate on which the task depends exceeds the preset threshold, reduce the weight of this link, try to find a new optimal path, and adjust the communication topology:

[0053] Calculate the fitness m i of task T i ' on the new path P ij '. If m ij '> S i , then migrate the task.

[0054] If the fitness m ij is lower than the threshold, adjust the bandwidth occupancy of task T i to reduce the impact on network resources

[0055] The beneficial technical effects of the present invention are at least as follows:

[0056] First, to address the problem that resource requirements are difficult to predict in advance, the present invention uses self-supervised learning (SSL), combines time series Transformer and contrastive learning, and automatically learns the resource requirement patterns of different tasks from historical emergency event data. Without a large amount of manually labeled data, it can accurately predict future bandwidth, computing resources, and device usage requirements. Through this innovation, the system can perceive upcoming resource bottlenecks in advance and provide data support for subsequent scheduling, thus avoiding resource waste or communication interruptions caused by scheduling lags.

[0057] Second, to address the problem that traditional communication topologies are rigid and cannot adapt to sudden changes, the present invention uses a dynamic topology graph neural network (DT-GNN). By constructing a dynamic graph model, it analyzes the communication network structure in real time and combines graph reinforcement learning (Graph RL) to autonomously learn the optimal link adjustment strategy. When the links are blocked due to damage or overload of some communication nodes, DT-GNN can adaptively adjust the communication topology according to the resource requirement prediction results, automatically select the optimal communication path, and ensure that the command and dispatch are not interrupted. In addition, by combining the collaborative optimization of edge computing and cloud computing, the present invention can quickly perform topology adjustment under low-latency conditions and improve the real-time response ability of the system.

[0058] Finally, to address the problem of lack of global optimization in resource scheduling, based on resource prediction and topology optimization, the present invention uses reinforcement learning combined with a multi-objective optimization method, comprehensively considers factors such as task priority, resource consumption, computing, and communication load, and intelligently adjusts the computing resource and bandwidth allocation strategies. Through this mechanism, the present invention can not only improve the utilization rate of resources, but also ensure that the resource requirements of different tasks are reasonably met, and avoid the situation of resource shortage for critical tasks and resource waste for low-priority tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] The present invention will be further described with reference to the accompanying drawings. However, the embodiments in the drawings do not constitute any limitation to the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on the following drawings without creative efforts.

[0060] Figure 1 It is a flowchart of a resource dynamic scheduling method for an emergency command fusion communication system based on artificial intelligence disclosed in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0061] Embodiments of the present invention will be described in detail below. Examples of the embodiments are shown in the accompanying drawings, where like or similar reference numerals denote like or similar elements or elements having like or similar functions throughout. The embodiments described below by referring to the drawings are exemplary only for explaining the present invention and should not be construed as limiting the present invention.

[0062] As Figure 1 shown, the method for dynamic resource scheduling of an emergency command fusion communication system based on artificial intelligence provided by an embodiment of the present invention includes the following steps:

[0063] S1. Obtain multiple sensor and log data of the emergency command system, and use contrastive learning based on task similarity to construct an unlabeled learning task. Combine with a time series Transformer to automatically learn the resource consumption patterns in historical data and generate resource predictions for future time periods.

[0064] Specifically, obtain the data sources: multiple sensors and log data of the emergency command system, including:

[0065] Network monitoring device: Collect the bandwidth usage B i , representing the network resource occupancy of task T i at different times.

[0066] Computing node management system: Record the computing resource occupancy C i , reflecting the computing consumption of task T i during operation.

[0067] Device health monitoring system: Provide device status data H i , such as CPU load, storage occupancy, channel interference situation, etc.

[0068] Furthermore, data preprocessing: Adopt a sliding window method to slice the data of continuous T w time steps to form a historical sequence to ensure that the model can learn time-dependent relationships. Normalize data with different dimensions so that they can be learned on the same scale, improving the model convergence speed.

[0069] Furthermore, construction of self-supervised learning objectives: Use contrastive learning based on task similarity to construct an unlabeled learning task:

[0070] Calculate the historical resource usage pattern z i of task T i , representing the characteristics of the task, and extract the key resource usage patterns of the task:

[0071] z i = f θ (Wi ) (1)

[0072] Among them, f θ is a transducer model with parameters that can automatically learn the resource consumption patterns of tasks.

[0073] Make the feature vectors z of similar tasks i , closer, and the feature vectors of different tasks are far away:

[0074]

[0075] Among them, sim(z i , z j ) represents the cosine similarity, and τ is the temperature coefficient.

[0076] Furthermore, construct a resource demand prediction model: predict future resource demands based on historical data and use a time series Transformer for modeling.

[0077] Input: The task feature vector z i , and extract the trend from the historical resource usage patterns of the task.

[0078] Output: The predicted value of the bandwidth demand for the next T p time steps

[0079] Calculate the resources and the device health status

[0080] Use the weighted mean square error to calculate the prediction error, ensuring that the prediction error of key resources has a greater impact:

[0081]

[0082] Among them, w B , w C , w H are the loss weights for bandwidth, computing resources, and device health status respectively.

[0083] Furthermore, the optimization objective: jointly train by comprehensively considering task similarity (self-supervised learning) and prediction error (supervised learning), enabling the model to learn from unlabeled data and accurately predict future resource demands.

[0084] Final output: The predicted resource demands for the future time period

[0085] S2. Build a dynamic graph neural network to optimize the communication topology. Dynamically and adaptively adjust the communication topology according to resource prediction combined with reinforcement learning, and select the optimal communication path.

[0086] Specifically, in the emergency command integrated communication system, emergencies may cause some communication nodes to fail and link congestion, affecting the stability of command and dispatch. To ensure that the communication network can dynamically adapt to changes in resource requirements, this step is based on the resource prediction results of Step 1 Optimize the communication topology to adapt to future bandwidth, computing resource requirements, and device health status. This step uses a dynamic graph neural network (DT-GNN) to optimize the communication topology, adjusts the link weights in combination with task requirements, and dynamically reconstructs the network structure to ensure the stability and efficiency of the communication link. Finally, output the optimized communication topology G t ', for subsequent resource scheduling.

[0087] Furthermore, build the communication network topology and initialize the node state:

[0088] Input data: Use the resource prediction of Step 1 Combined with the current communication network structure G t For dynamic adjustment.

[0089] Topology definition: Build a dynamic communication graph Where:

[0090] Node set Represents communication devices such as command centers, base stations, and terminals.

[0091] Edge set ε = {e ij} Represents the communication link between nodes.

[0092] Weight matrix Records the link quality, including bandwidth, delay, packet loss rate, etc.

[0093] Initialize the node state: Each node v i Has an initial state h i , including bandwidth demand Computing resources And device health status Ensure that the topology adjustment can adapt to changes in resource requirements. Set the weight w ij Of the edge e ij , and the calculation method is as follows:

[0094]

[0095] Where, d ij Is the physical distance of the link, r ij Is the link reliability, l ij Is the link delay; Combined with the predicted bandwidth requirements in step 1, guide the topology to optimize towards high-demand nodes; α1, α2, α3, α4 are hyperparameters to balance various factors.

[0096] Furthermore, information dissemination and update: Use a dynamic graph neural network (DT-GNN) to model the communication topology, and iteratively disseminate link information so that each node can perceive the global state.

[0097] Topology adjustment calculation: Through multiple rounds of iterative calculations, each node dynamically adjusts its own connections according to the states of its neighbor nodes, ensuring that high-reliability links are preferentially retained and low-quality links are dynamically removed.

[0098] Calculate the probability p of each link's retention or deletion ij , which is used to decide whether to retain or delete the link, and establish new links between high-demand nodes:

[0099]

[0100] When p ij < τ (set threshold), delete the link;

[0101] When p ij > 1 - τ, add links between high-demand nodes to form a new network topology G t '.

[0102] Furthermore, calculate the optimal communication path: On the optimized topology G t ', based on the bandwidth requirements of the task select the path with the lowest latency and ensure that the links in the path have sufficient bandwidth margin.

[0103] Path cost calculation: Define the optimization objective, minimize the path latency, and at the same time increase the bandwidth margin. The formula is as follows:

[0104]

[0105] Among them, is the maximum bandwidth capacity of the link, and β is used as an adjustment factor to ensure that high-demand tasks can preferentially obtain low-latency paths.

[0106] Output the optimized topology information: Output the new communication topology G t ' for resource scheduling in step 3 to ensure the efficient matching of computing resources and communication resources; output the score w of each communication link ij , which is used to preferentially select high-quality communication links in subsequent scheduling processes to improve the reliability of the emergency command system.

[0107] S3. Obtain the resource pool, and formulate a globally optimal resource scheduling strategy by combining resource prediction and the optimized communication topology, and generate an optimal resource allocation plan for each task; the resource pool includes available computing resources and communication bandwidth.

[0108] Specifically, in the emergency command integrated communication system, the goal of resource scheduling is to ensure the reasonable allocation of computing resources and communication bandwidth to meet the dynamic requirements of tasks under emergencies. Since emergencies may cause instability in some communication links, excessive load on computing nodes, or imbalance in resource allocation, relying solely on traditional static scheduling schemes cannot guarantee the efficiency and stability of the system. This step is based on the optimized communication topology G t ' and link weight w ij , and combines the resource demand prediction in Step 1 to formulate a globally optimal resource scheduling strategy. This strategy needs to consider simultaneously the dynamic topological changes of communication links, task priorities, and the global optimality of resource allocation to ensure that the command system can still operate efficiently in case of emergencies.

[0109] Construct a resource scheduling optimization model:

[0110] Input data: The input of this step includes:

[0111] The prediction result of Step 1 for guiding the decision-making of the demand side for resource allocation.

[0112] The optimized topology G t ' and its link weight w ij for calculating the availability of resources and link stability.

[0113] Task resource status representation: Define the demand vector of task T i where: is the bandwidth resource required by task T i ; is the computing resource demand; represents the device health status, which determines whether the task is scheduled preferentially.

[0114] Resource pool definition: Set the available computing resources C total and communication bandwidth B total to ensure that the allocation plan does not exceed the system capacity.

[0115] Furthermore, calculate the task priority: Since the urgency of the task, the device health status, and the bandwidth demand have different impacts on the scheduling decision, define the task priority score p i as the key basis for resource allocation:

[0116] ​​​

[0117] Among them, u i is the task urgency level, provided by the command system. Reflects the proportion of bandwidth requirements of the task, ensuring that high-demand tasks are preferentially allocated resources. Affects the scheduling priority of tasks in case of resource shortage. Additional item Calculate task T i The quality of the adjacent links in the communication topology where it is located, ensuring that the task preferentially selects a path with high stability. γ1, γ2, γ3, γ4 are weight factors to control the influence of different factors.

[0118] Furthermore, the goal is to maximize the utilization efficiency of bandwidth and computing resources while ensuring the task priority order, and avoid resource waste or overload.

[0119] Optimization strategy: Based on task priority p i , calculate the optimal resource allocation plan

[0120] Set constraint conditions to ensure that the total amount of resource allocation does not exceed the system capacity:

[0121]

[0122] Adopt a dynamic adjustment mechanism based on link weight w ij . If the reliability of the communication link mainly relied on by task T i decreases, reduce the bandwidth allocation for this task and preferentially allocate it to other tasks with stable paths.

[0123] Calculate the scheduling loss function: Adopt the goal of maximizing resource utilization rate and optimize it in combination with task priority:

[0124]

[0125] Among them, and represent the matching degree between the allocated resources and the requirements. Restrict the bandwidth imbalance of adjacent tasks to ensure that the link will not be overloaded due to a single task. λ is the regularization coefficient to control the trade-off between resource balance and priority.

[0126] Final output: Generate the optimal resource allocation plan for each task T i and pass the result to the next step to ensure that the task can be efficiently executed based on the allocated resources.

[0127] At the same time, provide feedback on the resource usage status to ensure that the system load balance can be adjusted in subsequent steps and the resource utilization efficiency can be optimized.

[0128] Adjust the scheduling strategy of computing resources. In the case of overload of computing nodes, automatically allocate computing tasks to computing units with lower loads to ensure the operational stability of the entire system.

[0129] S4. Based on the node status information of the emergency command system, intelligently match computing tasks with computing nodes for the optimal resource allocation plan and optimized communication topology of each task, and dynamically adjust resource allocation and communication paths during the task execution process to generate task execution feedback for the corresponding tasks; the node status information includes: maximum computing capacity, current load, maximum bandwidth, and current bandwidth occupancy.

[0130] Specifically, in the emergency command converged communication system, the execution of tasks not only depends on reasonable resource scheduling (the output of step 3), but also needs to ensure a high degree of matching between computing tasks and the communication network to ensure the stable and efficient operation of the system during the execution process. The goal of this step is based on the optimized resource scheduling plan in step 3 Combined with the optimized communication topology G t ', intelligently match computing tasks with computing nodes, and dynamically adjust resource allocation and communication paths during the task execution process to cope with possible computing load overload or network status fluctuations.

[0131] Obtain input data:

[0132] The resource scheduling result of step 3 That is, the optimal bandwidth and computing resource allocation plan for task T i .

[0133] The optimized topology G t ' of step 2, which is used to select the execution nodes of computing tasks to ensure the stability of the communication path.

[0134] Computing node status information: computing node set Contains the computing capabilities of each computing node v j Current load Maximum bandwidth And current bandwidth occupancy

[0135] Furthermore, select the task execution node:

[0136] The fitness m i of computing task T j executed on computing node v ij , comprehensively considering computing resources, bandwidth, and communication link quality:

[0137]

[0138] Among them, and are the computing resources and bandwidth required for task T i respectively. and are the maximum computing resources and bandwidth capacities of computing node v j respectively. and represent the current computing resources and bandwidth occupancy of node v j respectively, ensuring that a computing node with a lower load is selected for task scheduling. w ij is the link score associated with task T i , ensuring that a stable communication path is selected. δ1, δ2, δ3 are adjustment coefficients to balance computing power, bandwidth adaptability, and communication stability.

[0139] Task T i selects the computing node with the highest adaptability for execution, that is:

[0140]

[0141] If the adaptability of all computing nodes is lower than the threshold τ m , it indicates that the current system computing resources are insufficient, and the task execution priority needs to be adjusted, and the execution of low-priority tasks is delayed to ensure that high-priority tasks can be scheduled.

[0142] Furthermore, computing node load monitoring: During task execution, the load of each computing node v j changes over time. If an overload occurs, the task allocation strategy needs to be adjusted.

[0143] Task load adjustment rule: Set the maximum load threshold τ C of the computing node. If the current computing load of a certain computing node exceeds this threshold, the task allocation needs to be adjusted:

[0144]

[0145] Among them represents the task set currently allocated to node .

[0146] If the load exceeds the standard, part of the tasks are preferentially migrated to the computing node v k with rich computing resources, or the computing priority of part of the tasks is reduced to make them wait for execution.

[0147] Furthermore, dynamic optimization of the communication path: During task execution, if the communication link on which task T i depends becomes congested (bandwidth utilization rate exceeds τ B),then it is necessary to dynamically adjust the communication path of the task.

[0148] Path selection and switching rules: In the optimized topology G t ', find a new low-latency path P i ', and reallocate the bandwidth to ensure that the task execution is not interrupted.

[0149] Bandwidth adjustment strategy: If the link bandwidth on the new path P i ' is sufficient, directly switch. If the bandwidth is insufficient, reduce the bandwidth occupancy of low-priority tasks so that high-priority tasks can be executed first.

[0150] Furthermore, task completion status recording: After each task is executed, record the actual consumption values of its computing resources and communication bandwidth and the predicted values for comparison to calculate the error:

[0151]

[0152] If the error exceeds the set threshold τ Δ , then adjust the resource prediction model in step 1 to improve the accuracy of the next round of task execution.

[0153] Error feedback mechanism: If the error continues to exceed the standard, then adjust the resource scheduling strategy in step 3 to improve the adaptability calculation of task execution and optimize the matching rules between computing tasks and computing nodes.

[0154] This step combines computing resource matching, dynamic load balancing, communication link switching, and error feedback mechanism during task execution to implement an adaptive task execution optimization scheme. Through intelligent task scheduling and resource adaptation, this step ensures that the emergency command integrated communication system can operate stably and efficiently in complex environments and improves the reliability of overall task execution.

[0155] S5. Real-time monitor the system status and optimize the resource prediction, topology adjustment, and scheduling strategy based on the execution error feedback from task execution.

[0156] Specifically, in the emergency command integrated communication system, during task execution, resource scheduling imbalance or task execution failure may occur due to environmental changes, computing resource load fluctuations, or communication link congestion. Therefore, it is necessary to real-time monitor the system status and optimize the resource prediction, topology adjustment, and scheduling strategy based on the execution error feedback. This step is based on the actual status of task execution in step 4 and the resource allocation scheme in step 3 for comparison, calculate the error, and optimize the resource prediction model through online adjustment to improve the dynamic adaptability of the system.

[0157] Furthermore, obtain input data:

[0158] Task execution feedback of Step 4 i.e., task T i The actually consumed bandwidth and computing resources.

[0159] Resource scheduling result of Step 3 As the allocation benchmark before task execution.

[0160] Optimized topology G of Step 2 t ', used to evaluate the communication link status.

[0161] Error calculation: Calculate the bandwidth and computing resource allocation errors:

[0162]

[0163] If ΔB i or ΔC i exceeds the set threshold τ Δ , it indicates that the resource prediction is inaccurate and the resource prediction model needs to be optimized.

[0164] Link quality change evaluation: Analyze the link bandwidth utilization U i on which task T ij depends. If it exceeds the set threshold τ B , the communication topology or resource scheduling strategy needs to be adjusted.

[0165] Furthermore, the optimization goal: Improve the resource demand prediction model of Step 1 to adapt to the real-time changing system load and reduce the resource allocation error.

[0166] Update the prediction model parameters: Define the error correction term Δθ, which is used to adjust the resource prediction model parameters θ to reduce the prediction deviation of future tasks:

[0167]

[0168] where η is the learning rate, controlling the update step size. represents the gradient of the resource prediction model f θ with respect to the input data X i .

[0169] Dynamically adjust the prediction weights: For tasks with high errors, increase the weights of historical data to make their impact on model adjustment greater and improve the prediction accuracy.

[0170] Furthermore, link bandwidth utilization monitoring: If the bandwidth utilization of the link e i on which task T ij depends exceeds τ B , then reduce the weight w ij of this link and try to find a new optimal path.

[0171] Furthermore, the topology adjustment strategy: Calculate task T i on the new path P i 's fitness m ij '. If m ij '> m ij , then migrate the task. If the fitness m ij is lower than the threshold, then adjust the bandwidth occupancy of task T i to reduce the impact on network resources.

[0172] Furthermore, the dynamic adjustment of task priority: Based on the historical execution error, adjust the task scheduling priority to ensure that high-error tasks obtain resources first, so as to reduce future scheduling deviation.

[0173] Load balancing optimization: If the load of a certain computing node v j exceeds τ C , then adjust task T i to a node v k with more abundant computing resources. If the overall computing resources of the system are tight, then reduce the computing resource allocation of low-priority tasks so that critical tasks can be executed first.

[0174] This step realizes resource prediction, communication topology adjustment and scheduling strategy optimization through real-time monitoring and error feedback optimization, ensuring that the emergency command system can dynamically adapt to environmental changes and improve the stability of task execution and resource utilization efficiency.

[0175] In summary, through self-supervised learning-based resource prediction, adaptive topology optimization of variable topology graph neural networks, and reinforcement learning-driven global scheduling, the present invention realizes efficient, stable and intelligent dynamic resource scheduling in the emergency command fusion communication system. Compared with traditional methods, the present invention reduces scheduling lag, improves communication stability, optimizes resource allocation efficiency, and can significantly enhance the reliability of the emergency command system and the ability to respond to emergencies.

[0176] The above describes specific embodiments of this specification, and other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in a different order than in the embodiments and still achieve the desired results. Additionally, the processes depicted in the drawings do not necessarily have to be performed in the specific order or continuous order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0177] The systems, devices, modules or units described in the above embodiments can be specifically implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer can be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0178] For the convenience of description, when describing the above devices, various units are described separately according to their functions. Of course, when implementing this specification, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0179] Those skilled in the art should understand that the embodiments of this specification can be provided as a method, a system, or a computer program product. Therefore, the embodiments of this specification can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the embodiments of this specification can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0180] This specification is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of this specification. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for implementing the functions specified in Figure 1 one or more flows and / or Figure 1 blocks or multiple blocks.

[0181] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device, and the instruction device implements the functions specified in Figure 1 one or more flows and / or Figure 1 blocks or multiple blocks.

[0182] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one process Figure 1 one process or multiple processes and / or boxes Figure 1 steps of the functions specified in one box or multiple boxes.

[0183] In a typical configuration, a computing device includes one or more processors (CPUs), an input / output interface, a network interface, and memory.

[0184] The memory may include non-permanent memory in the form of computer-readable media, random access memory (RAM) and / or non-volatile memory such as read-only memory (ROM) or flash memory (flash RAM). The memory is an example of computer-readable media.

[0185] Computer-readable media includes permanent and non-permanent, removable and non-removable media and can store information by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette tapes, magnetic tape disk storage or other magnetic storage devices, or any other non-transmission media that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media such as modulated data signals and carrier waves.

[0186] It should also be noted that the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, commodity or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, commodity or device. Without further limitation, an element defined by the statement "comprising one..." does not exclude the presence of additional identical elements in the process, method, commodity or device comprising the element.

[0187] This specification can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. This specification can also be practiced in a distributed computing environment where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0188] Each embodiment in this specification is described in a progressive manner. For the same or similar parts among the embodiments, reference can be made to each other. Each embodiment focuses on the differences from other embodiments. In particular, for system embodiments, since they are basically similar to method embodiments, the description is relatively simple. For the relevant parts, reference can be made to the partial description of the method embodiments.

[0189] Finally, it should be noted that what is disclosed in an embodiment of a lithium battery pack chip equalization control platform of the present invention is only a preferred embodiment of the present invention, which is only used to illustrate the technical solution of the present invention and is not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments or equivalently replace some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for dynamic resource scheduling of emergency command fusion communication system based on artificial intelligence, characterized in that: The method comprises the following steps: S1. Obtain multiple sensor and log data from the emergency command system, use contrastive learning based on task similarity, construct unlabeled learning tasks, and combine with time series Transformer to automatically learn resource consumption patterns in historical data and generate resource forecasts for future time periods; S2. Build a dynamic graph neural network to optimize the communication topology, dynamically and adaptively adjust the communication topology based on resource prediction combined with reinforcement learning, and select the optimal communication path; S3. Obtain a resource pool, formulate a globally optimal resource scheduling strategy based on resource prediction and optimized communication topology, and generate an optimal resource allocation plan for each task; the resource pool includes available computing resources and communication bandwidth; S4. Intelligently match computing tasks and computing nodes for the optimal resource allocation scheme and optimized communication topology for each task based on the node status information of the emergency command system, dynamically adjust resource allocation and communication paths during task execution, and generate task execution feedback for the corresponding task; the node status information includes: maximum computing power, current load, maximum bandwidth and current bandwidth occupancy; S5. Monitor the system status in real time and optimize resource prediction, topology adjustment and scheduling strategies based on the execution error feedback of task execution feedback.

2. The method for dynamic resource scheduling of emergency command fusion communication system based on artificial intelligence according to claim 1 is characterized in that: The data sources of the multiple sensors and log data include: network monitoring equipment, computing node management system and equipment health monitoring system; The S1 also includes: using a sliding window method to w The data of each time step is sliced ​​to form a historical sequence; the data of different dimensions are normalized so that they can be learned on the same scale.

3. The method for dynamic resource scheduling of emergency command fusion communication system based on artificial intelligence according to claim 1 is characterized in that: The contrastive learning based on task similarity constructs an unlabeled learning task, including: Use the transformer model to calculate the historical resource usage pattern of the task, generate a task feature vector to represent the characteristics of the task, and extract the key resource usage pattern of the task; Based on normalized cross entropy analysis, feature vectors of similar tasks are made closer and feature vectors of different tasks are made farther apart; The construction of unlabeled learning tasks, combined with time series Transformer, automatically learns resource consumption patterns in historical data and generates resource forecasts for future time periods, including: Based on historical data, future resource requirements are predicted, and a time series Transformer is used for modeling. The task feature vector is used as input, and the predicted bandwidth requirement, computing resources, and equipment health status of the future time step are output. The time series Transformer uses a weighted mean square error to calculate the prediction error, ensuring that the prediction error of key resources has a greater impact.

4. The method for dynamic resource scheduling of emergency command fusion communication system based on artificial intelligence according to claim 1 is characterized in that: The S2 specifically includes: Building a dynamic communication graph Initialize the node state, where the dynamic communication graph Includes: Node Set represents the communication equipment; the edge set ε={e ij } represents the communication link between nodes; weight matrix Record link quality; Based on dynamic communication graph The communication topology is modeled using a dynamic graph neural network, and link information is iteratively propagated to enable each node to perceive the global state. Through multiple rounds of iterative calculations, each node dynamically adjusts its own connection according to the state of neighboring nodes, and calculates the probability of each link staying or leaving, decides whether to keep or delete the link, and establishes new links between high-demand nodes to generate an optimized communication topology. On the optimized communication topology, the path with the lowest latency is selected based on the bandwidth requirement of the task, and it is ensured that the links in the path have sufficient bandwidth margin to generate the path cost of the communication topology.

5. The method for dynamic resource scheduling of emergency command fusion communication system based on artificial intelligence according to claim 4 is characterized in that: The calculation of the probability of each link being retained or deleted is performed to determine whether to retain or delete the link. Specifically, When p ij <τ, delete the link; When p ij >1-τ, add links between high-demand nodes to form a new network topology G t '; Among them, p ij is the probability of leaving or staying on each link, w ij For edge e ij The weight, w i'j' For edge e i'j' The weight of , i', j' is the edge i' and j', τ is the set threshold; The edge ij The weight w ij Calculated as: Among them, d ij is the physical distance of the link, r ij is the link reliability, l ij is the link delay; Combined with the predicted bandwidth demand, the topology is guided to optimize towards high-demand nodes; α1, α2, α3, and α4 are hyperparameters that balance various factors.

6. The method for dynamic resource scheduling of emergency command fusion communication system based on artificial intelligence according to claim 1 is characterized in that: The resource scheduling strategy satisfies the task priority; the task priority is obtained by performing a weighted linear analysis based on the task urgency, the bandwidth demand ratio of the task, the scheduling priority of the task under resource constraints, and the quality of the adjacent links in the communication topology where the current task is located.

7. The method for dynamic resource scheduling of emergency command fusion communication system based on artificial intelligence according to claim 6 is characterized in that: The resource scheduling strategy is based on the constraint that the available computing resources and communication bandwidth do not exceed the maximum capacity, with the goal of maximizing resource utilization, and is optimized in combination with task priority; the scheduling loss function of the resource scheduling strategy is: in, and Indicates the matching degree between allocated resources and demand; Constrain the bandwidth imbalance of adjacent tasks to ensure that the link will not be overloaded by a single task; λ is the regularization coefficient, which controls the trade-off between resource balance and priority.

8. The method for dynamic resource scheduling of emergency command fusion communication system based on artificial intelligence according to claim 5 is characterized in that: The S4 comprises: Based on the computing resources and bandwidth required for the task, the maximum computing resources and bandwidth capacity of the computing node, the current computing resources and bandwidth occupancy of the node, and the link score associated with the task, the fitness of the task to be executed on the computing node is calculated; wherein the link score associated with the task is the weight of the corresponding edge; The task selects the computing node with the highest fitness for execution; At the same time, during the task execution process, the load of each computing node changes over time. If an overload occurs, the task allocation strategy needs to be adjusted; if the communication link that the task depends on is congested, the communication path of the task needs to be dynamically adjusted; After each task is executed, the actual consumption values ​​of its computing resources and communication bandwidth are recorded, compared with the predicted values, and the error is calculated. If the error exceeds the preset threshold, the resource prediction value of step S1 is returned to improve the accuracy of the next round of task execution; if the error continues to exceed the standard, the resource scheduling strategy of step S3 is adjusted to improve the fitness calculation of task execution and optimize the matching rules between computing tasks and computing nodes.

9. The method for dynamic resource scheduling of emergency command fusion communication system based on artificial intelligence according to claim 8 is characterized in that: If an overload occurs, the path selection and switching rules of the task allocation strategy need to be adjusted as follows: Set the maximum load threshold τ of the computing node C , if a computing node If the current computing load exceeds this threshold, the task allocation needs to be adjusted: in, Indicates the node currently assigned A collection of tasks; For task T i The computing resources required; If the load exceeds the limit, some tasks will be migrated to nodes with sufficient computing resources. k , or lower the computing priority of some tasks to make them wait for execution; If the communication link that the task depends on is congested, the path selection and switching rules of the communication path of the task need to be dynamically adjusted as follows: If the link bandwidth on the new path is sufficient, switch directly; If the bandwidth is insufficient, the bandwidth usage of low-priority tasks will be reduced so that high-priority tasks can be executed first.

10. The method for dynamic resource scheduling of emergency command fusion communication system based on artificial intelligence according to claim 1, characterized in that: The execution error is a bandwidth allocation error and a computing resource allocation error. If the bandwidth allocation error or the computing resource allocation error exceeds a set threshold, it indicates that the resource prediction is inaccurate and the prediction parameters of the resource prediction need to be optimized. The S5 further includes: If the bandwidth utilization of the link that the task depends on exceeds the preset threshold, the weight of the link is reduced, and an attempt is made to find a new optimal path and adjust the communication topology: Computational task T i On the new path P i 'The fitness m ij ', if m ij '>S i , then migrate the task; If the fitness m ij If it is lower than the threshold, adjust the task T i bandwidth usage and reduce the impact on network resources.

Images