A continuous reputation driving-based unmanned aerial vehicle cluster resource collaborative allocation method and system

Abstract

The application discloses a kind of continuous reputation driving-based unmanned aerial vehicle cluster resource collaborative allocation method and system, comprising: obtaining the state information of each node of unmanned aerial vehicle cluster, constructs time-varying directed interaction topology;Calculate the basic net income of node;Introduce channel fading tolerance threshold, filter environmental noise, calculate dynamic continuous reputation score, and obtain the comprehensive total income of driving strategy update in combination with basic net income;Implement the asymmetric learning mechanism based on concurrent communication in-degree bottleneck, update the resource input proportion of node, prevent core node overload;Identify and protect the key backbone link that maintains global connectivity;Based on comprehensive total income and reputation score, drive the self-organizing evolution of machine group topology under the premise of protecting backbone link.The present application realizes fine-grained resource fair scheduling and accurate identification of selfish behavior, eliminates the decision shock risk of core relay node, and enhances the survivability and global connectivity of machine group topology in dynamic environment.

Description

Technical Field

[0001] This invention relates to the field of distributed computing technology, and in particular to a method and system for collaborative allocation of unmanned aerial vehicle (UAV) swarm resources based on continuous reputation. Background Technology

[0002] With the rapid development of drone and artificial intelligence technologies, drones have been widely applied in fields such as smart agriculture, disaster relief, environmental monitoring, and battlefield reconnaissance. In these applications, individual drones, limited by their onboard computing power, storage capacity, and energy resources, struggle to independently complete complex sensing and computation tasks. Therefore, drone swarm technology, based on the collaborative work of multiple drones, has gradually become an important development direction. In a drone swarm system, distributed processing of large-scale sensing data can be achieved through information interaction and collaborative processing among multiple nodes, thereby improving overall computing efficiency and task execution capabilities. Currently, research on drone swarm collaborative computing technology mainly focuses on task scheduling, resource allocation, and data processing strategies. Related methods include rule-based task allocation methods, optimization algorithm-based resource scheduling methods, and machine learning-based adaptive decision-making methods.

[0003] However, with the continuous expansion of UAV swarm scale and the increasing complexity of application tasks, traditional methods relying on centralized computing or fixed allocation strategies are gradually becoming insufficient to meet practical needs. On the one hand, cloud computing-based centralized processing models suffer from problems such as large communication latency and strong network dependence; on the other hand, resource allocation methods based on static rules lack the ability to dynamically respond to changes in node states, making it difficult to adapt to the heterogeneous nature of UAV node resources and the dynamic changes in network structure. Therefore, how to achieve adaptive adjustment of resource allocation strategies and multi-node collaborative optimization in the dynamic network environment of UAV swarms, combining the resource states and historical collaborative behaviors of each node, has become an important research direction in the field of collaborative computing for UAV swarms.

[0004] Existing technologies still have the following main drawbacks in achieving the above objectives:

[0005] 1. The lack of a continuous collaborative incentive mechanism leads to limited granularity in UAV resource allocation and susceptibility to channel noise: Existing mechanisms mostly employ binary discrete game models, which cannot effectively model and incentivize UAVs to continuously and finely adjust their computing power and bandwidth lending ratios. Furthermore, due to the lack of tolerance design for physical channel fading, existing systems struggle to distinguish between "performance fluctuations caused by complex high-altitude channels" and "subjective betrayal by UAVs to save power," easily leading to misjudgments of reputation and hindering the formation of stable and fair collaborative guidance.

[0006] 2. Ignoring concurrent communication load bottlenecks can easily lead to decision-making oscillations and overload crashes in core relay drones: Existing scheduling strategies typically assume sufficient node processing capacity. In dense formation scenarios, core relay drones face extremely high concurrent communication in-degrees when receiving a large number of coordination requests simultaneously. Due to the lack of effective load suppression mechanisms, the decision-making logic of core nodes is easily overwhelmed by massive redundant signaling, leading to network card bandwidth exhaustion or CPU overload, causing decision-making oscillations or even system paralysis, severely affecting the convergence stability of large-scale clusters.

[0007] 3. The dynamic topology reconfiguration mechanism is prone to blindness, easily leading to swarm network outages and global connectivity collapse: Existing network reconfiguration technologies, when dealing with selfish nodes or local performance degradation, often blindly disconnect and reconnect based on instantaneous local gains, lacking awareness of the robustness of the global topology. This mechanism, in the process of expelling inferior nodes, is highly susceptible to mistakenly severing the "backbone support links" that maintain the entire swarm's communication connection, resulting in network fragmentation, creating "flying islands," and seriously threatening the global resilience and mission continuity of the UAV swarm. Summary of the Invention

[0008] To address the shortcomings of existing technologies, this invention provides a method and system for collaborative resource allocation in unmanned aerial vehicle (UAV) swarms based on continuous reputation.

[0009] To achieve the above-mentioned objectives, the technical solution adopted by the present invention is as follows:

[0010] A method for collaborative resource allocation in a drone swarm based on continuous reputation, applied to a collaborative computing network system consisting of multiple drone nodes, includes the following steps:

[0011] S101: Obtain the initial physical parameters and operating status of each node in the UAV cluster, and construct a time-varying directed interaction topology that characterizes the computing power offloading or communication relay request link between UAV nodes.

[0012] S102, based on game modeling of continuous action space, calculate the basic net income of each UAV node; wherein, the cooperative behavior of UAV nodes is mapped to the continuous resource input ratio in the interval [0,1], and the basic net income is the algebraic sum of the system cooperative gain of receiving computing power support from neighboring nodes and the endogenous energy consumption cost generated by its own resource input.

[0013] S103, within a preset sliding time window, based on the change in the resource investment ratio and the preset channel fading tolerance threshold, calculate the dynamic continuous reputation score of each UAV node, and superimpose the dynamic continuous reputation score with the basic net income to obtain the comprehensive total income of each UAV node.

[0014] S104. Based on the total comprehensive benefit, and according to the concurrent communication in-degree of each UAV node, calculate the nonlinear suppression operator and implement asymmetric learning to update the resource investment ratio of each UAV node.

[0015] S105, based on the Strongly Connected Components (SCC) decomposition algorithm, identifies and extracts the set of key communication backbone links that maintain the global communication connectivity of the UAV cluster.

[0016] S106, based on the comprehensive total revenue and the dynamic continuous reputation score, drive the topology self-organizing evolution of the drone swarm: when disconnecting from low-reputation neighbor nodes, apply immune disconnection protection to the links in the set of key communication backbone links, and guide drone nodes to establish new connections with potential high-reputation collaborators.

[0017] Further, in step S101, the construction of the time-varying directed interaction topology includes:

[0018] The drone swarm network is abstracted as a directed graph that evolves over time. ,in For a collection of drone nodes, Let be the set of request links at time t;

[0019] Using binary adjacency matrix Characterizing the physical topology, when drone nodes Towards When sending a collaboration request, ,otherwise ;

[0020] Calculate the concurrent communication out-degree of each drone node. Concurrent communication in-degree .

[0021] Furthermore, in step S102, the drone node exist The resource allocation ratio at any given time is Its basic net income function The calculation formula is:

[0022]

[0023] in, These are interaction variables defined based on directed adjacency relationships. When the elements of the directed adjacency matrix... or hour, ;otherwise, ; Indicates drone node The undirected physical interaction degree, whose value is ; This represents the endogenous energy consumption cost per unit of resource input; This represents the system cooperative gain resulting from receiving computing power support from neighboring nodes; the system parameters satisfy... .

[0024] Further, in step S103, the unmanned aerial vehicle (UAV) nodes are calculated. Dynamic continuous credit scoring include:

[0025] Define physical channel tolerance threshold And based on the change in the proportion of continuous resource input. With the threshold Update the consistency count value based on the comparison results. ;

[0026] According to the formula Calculate the dynamic continuous reputation score, where c is the credit incentive strength coefficient; i (t) represents the time consistency count value;

[0027] According to the formula Calculate the total comprehensive revenue. .

[0028] Further, in step S104, the implementation of asymmetric learning to update the resource allocation ratio includes:

[0029] Based on the aforementioned total comprehensive benefits and The difference Drone nodes in a computer cluster Mimicking neighbor nodes The basic transition probability of the strategy ;

[0030] Based on drone nodes concurrent communication in-degree Calculate the nonlinear suppression operator : ,in Based on the baseline signaling sensitivity, This refers to hardware load suppression parameters;

[0031] Based on the aforementioned basic transition probability Nonlinear suppression operator And adjacency relationships, calculate the comprehensive influence weight between nodes. ;

[0032] Based on the continuous-time dynamics equation Update drone nodes resource input ratio .

[0033] Furthermore, in step S105, identifying the set of key communication backbone links specifically involves: before each topology evolution cycle, analyzing the current directed communication graph. Run the Strongly Connected Components (SCC) decomposition algorithm to extract and lock the minimum set of key edges that maintains the globally strongly connected state. , as a set of protective backbone links that are subject to forced immune disconnection operations.

[0034] Further, in step S106, the driving topology self-organizing evolution includes:

[0035] Calculate the relative performance satisfaction index of each UAV node within its local communication neighborhood. );

[0036] like Below the preset tolerance limit If so, the corresponding drone node will be activated and enter the evolution queue;

[0037] For unmanned aerial vehicle nodes in the evolutionary queue With probability Actively disconnect it from neighboring nodes The unprotected links between them, Sensitivity to the disruption of chains in response to selfish behavior, Neighboring nodes The proportion of resources invested;

[0038] Guide the drone node that has lost its link proportional to potential collaborators resource input ratio The probability of the potential collaborator Establish a new connection, in which This is the preference coefficient.

[0039] This invention also discloses a continuous reputation-driven UAV swarm resource collaborative allocation system for executing the aforementioned UAV swarm resource collaborative allocation method, comprising:

[0040] The environmental perception and topology construction module is used to obtain the initial physical parameters and operating status of each node in the UAV cluster, and to construct a time-varying directed interactive topology.

[0041] The continuous action game and payout calculation module is used for game modeling based on the continuous action space and to calculate the basic net payout of each drone node.

[0042] The tolerance filtering and continuous reputation evaluation module is used to calculate the dynamic continuous reputation score and overall revenue of each UAV node based on the change in the resource investment ratio and the preset channel fading tolerance threshold.

[0043] The load suppression and asymmetric learning module is used to update the resource investment ratio of each UAV node through an asymmetric learning mechanism based on the total comprehensive revenue and the concurrent communication in-degree of each node.

[0044] The SCC backbone link identification and protection module is used to identify and extract the set of key communication backbone links that maintain the global communication connectivity of the UAV cluster based on the strongly connected component SCC decomposition algorithm.

[0045] A reputation-driven topology self-organizing evolution module is used to drive the self-organizing evolution of the cluster topology based on the comprehensive total revenue and the dynamic continuous reputation score, while protecting critical backbone links.

[0046] The present invention also discloses an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the program to implement the steps of the above-described unmanned aerial vehicle (UAV) swarm resource collaborative allocation method.

[0047] The present invention also discloses a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-described method for collaborative allocation of resources in a drone swarm.

[0048] Compared with the prior art, the advantages of the present invention are as follows:

[0049] 1. This invention breaks through the limitations of traditional binary discrete incentive mechanisms. By modeling UAV cooperative strategies in a continuous action space using game theory, it achieves fine-grained and continuous allocation of resources such as computing power and bandwidth, better reflecting the real physical constraints of limited UAV battery capacity. More importantly, this invention innovatively introduces a continuous reputation evaluation mechanism based on channel fading tolerance thresholds. This mechanism can effectively filter out random performance fluctuations caused by complex electromagnetic environments at high altitudes, thereby accurately distinguishing between "decreased cooperative efficiency due to poor objective channel quality" and "subjective betrayal behavior by UAVs to save power." This fundamentally avoids the reputation misjudgment problem that traditional models are prone to in noisy environments, providing high-reliability incentive guidance for long-term, stable, and fair cooperation among clusters.

[0050] 2. To address the bottleneck issue of network card or CPU overload caused by core nodes in dense formations simultaneously processing massive collaborative requests (high-concurrency communication in-degree), this invention innovatively designs an asymmetric learning mechanism based on hardware processing capabilities. This mechanism introduces a nonlinear suppression operator that decays exponentially with the concurrent communication in-degree, automatically and dynamically downweighting or masking a large number of low-value or redundant external collaborative request signaling when the in-degree approaches its processing limit. This protects the independence and stability of the core UAV decision-making unit, preventing it from being overwhelmed by signaling and causing oscillations, and significantly accelerates the convergence speed of the global resource allocation strategy for the entire cluster when facing high-concurrency tasks, greatly improving the system's resilience and operational robustness.

[0051] 3. This invention overcomes the fatal flaw of existing topology reconfiguration technologies, which can lead to network disintegration due to "blindly disconnecting links." Before driving drones to disconnect inferior connections based on reputation, the system pre-runs a Strongly Connected Component (SCC) decomposition algorithm (such as the Tarjan algorithm) to identify and lock the minimum set of critical backbone links necessary to maintain global connectivity of the entire network, and applies immune disconnection protection to them. Under this constraint, based on node satisfaction and reputation scores, the system drives drones to disconnect from selfish nodes and actively seek high-reputation collaborators to establish new links. This mechanism ensures that the topology evolution process is "guided" and "protected," enabling the swarm to continuously optimize cooperation efficiency and gather resources towards high-reputation nodes, while fundamentally preventing the generation of "flying islands," thus guaranteeing the global resilience of large-scale drone swarms in complex dynamic environments and the continuous, uninterrupted execution of tasks. Attached Figure Description

[0052] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0053] Figure 1 This is a flowchart of a method for collaborative allocation of drone swarm resources based on continuous reputation in an embodiment of the present invention. Detailed Implementation

[0054] 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, not all, of the embodiments of the present invention. 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.

[0055] This invention provides a resource collaborative allocation method for UAV swarms based on continuous reputation, in order to solve the problems in the current field of edge computing collaboration for UAV swarms, such as the difficulty in modeling continuous computing power and bandwidth resource incentives, decision oscillations caused by high load of core relay UAVs, and blind topology dynamic reconstruction leading to swarm network outages and disconnections.

[0056] To achieve the above objectives, this embodiment adopts the following technical solution:

[0057] Step 101: Obtain the initial physical parameters and operational status of each node in the UAV swarm, and construct a time-varying directed interactive topology. The initial physical parameters and operational status refer to the set of static configurations and real-time dynamic data for each UAV performing collaborative reconnaissance or computational relay tasks. Static configurations include the UAV's maximum CPU computing power and maximum communication bandwidth; real-time dynamic data includes remaining battery power, concurrent communication in-degree, and the signal-to-noise ratio (SNR) with neighboring UAVs. These data collectively constitute the basic environmental information for the UAV agent to make resource decisions.

[0058] Step 102: A game-theoretic modeling method based on continuous action space is used to calculate the basic net payoff of the UAV nodes. The cooperative behavior of the UAVs is mapped to the proportion of continuous resource input within the interval [0, 1], achieving a fine-grained representation of computing power and bandwidth allocation. In this model, the basic payoff function of the UAVs integrates the system-level cooperative gain obtained by receiving computing power support from neighboring UAVs, as well as the intrinsic battery energy consumption cost generated by its own execution of computing tasks.

[0059] Step 103 introduces channel fading tolerance features to obtain a dynamic continuous reputation score and update the overall total revenue. This step samples the resource allocation ratio of the UAV within a preset sliding time window and sets a communication tolerance threshold to filter performance fluctuations caused by high-altitude physical wireless channel fading. Based on time consistency, the effective action increment is integrated and accumulated to accurately identify selfish UAVs that "stealthily betray" their mission to save power, thereby generating a continuous reputation score. This score is then superimposed with the basic net revenue to obtain the overall total revenue driving the system evolution.

[0060] Step 104: Implement an asymmetric learning mechanism based on the bottleneck of concurrent communication in-degree to update the drone resource allocation ratio. For high-in-degree drones at the center of the cluster, calculate a nonlinear suppression operator by monitoring their concurrent communication in-degree in real time and combining it with the hardware processing bandwidth threshold. During the strategy evolution process, this operator is used to dynamically reduce and shield external redundant coordination requests, preventing core relay drones from experiencing decision oscillations due to signaling overwhelm, and ensuring the stable convergence of the drone resource scheduling strategy.

[0061] Step 105: Identify key communication backbone links based on the Strongly Connected Component (SCC) decomposition algorithm. Before driving the swarm topology reconstruction, conduct a global topology robustness analysis to extract the minimum set of critical edges that maintain global communication connectivity for the UAV swarm from all network links, and mark them as protected backbone links. In subsequent evolution, predict and limit the disconnection of core links to avoid the UAV formation from becoming a "computation island" unable to communicate due to blind link breaks.

[0062] Step 106: Combining the overall total revenue and continuous reputation evaluation results, the system drives the self-organizing evolution of the drone swarm topology. Under the premise of applying immune disconnection protection to critical communication backbone links, the system activates the evolution queue based on the drones' satisfaction with the current local resource allocation. With a specific probability, it proactively disconnects inferior cooperative links with low-reputation, selfish drones and prioritizes establishing new connections with potentially high-reputation drones, ultimately driving the entire drone swarm to converge to a resource scheduling steady state of high cooperation and high resilience.

[0063] Specifically, in step 101, the process of obtaining the initial physical parameters and operating status of each node in the UAV cluster and constructing a time-varying directed interactive topology is as follows:

[0064] The dynamic drone swarm network is abstracted as a directed graph that evolves over time. .in, Indicates inclusion A collection of drone nodes; Indicates in This refers to the set of unidirectional computing power offloading or communication relay request links established between unmanned aerial vehicles (UAVs) at any given time. The underlying physical topology consists of a binary adjacency matrix. Characterization: When drones Towards When sending a cooperation request, the corresponding adjacency matrix element ,otherwise The system periodically collects the concurrent communication output of each drone. and concurrent communication in-degree This provides physical input for subsequent asymmetric signaling suppression.

[0065] Specifically, in step 102, the process of calculating the basic net profit of the drone node using a game-theoretic modeling method based on continuous action space is as follows:

[0066] Define drones exist The strategy state at any given time is the resource allocation ratio. .in, This indicates that the drone is in a completely selfish state, refusing to lend computing power in order to conserve battery power; This represents full-load sharing of its own computing power and bandwidth. To avoid redundant energy consumption calculations caused by bidirectional communication in directed request networks, a symmetric interaction indicator variable is defined. :when or hour, ;otherwise Computational drones Undirected physical interaction degree Furthermore, drones Basic net income function The calculation formula is:

[0067]

[0068] in, This refers to the internal battery energy consumption cost generated per unit of computing power or bandwidth invested by the drone. Indicates receiving a message from a neighbor's drone. The system-level gain obtained after support; parameters must meet the following requirements. This creates a game-theoretic dilemma.

[0069] Specifically, in step 103, the process of introducing channel fading tolerance features, obtaining dynamic continuous reputation scores, and updating the overall total revenue is as follows:

[0070] Considering that random fading in the high-altitude wireless channel of UAVs can cause jitter in the observed amount of shared resources, a physical channel tolerance threshold is introduced. Drones Time consistency count value The update rule is: when the change in continuous action... At that time, it is identified as physical layer noise, and the reputation is accumulated. When the decline in resources exceeds If the issue is not related to the channel, it is determined that the drone employed a "stealthy betrayal" strategy to conserve power, and the counter value is reset to [value missing]. (Set upper limit) Make (Prevent overflow).

[0071] Subsequently, computational drones Dynamic continuous credit scoring :

[0072]

[0073] in, The credit incentive strength coefficient, Map the continuous policy interval to This allows for the rewarding of high levels of collaboration and the punishment of selfish behavior. Therefore, the overall total benefit of using drones to guide the next strategy update is established as follows: .

[0074] Specifically, in step 104, the implementation of the asymmetric learning mechanism based on the concurrent communication in-degree bottleneck to update the UAV resource allocation ratio is as follows:

[0075] Based on the overall total revenue output in step 103, the drone Mimicking neighbor nodes The base transition probability of a superior strategy Comply with Fermi update rules:

[0076]

[0077] in, Choose sensitivity for the strategy. This represents the difference in the total revenue between the two drones. To prevent the relay drones at the center of the cluster from overloading the network card or CPU due to excessive concurrent requests, a nonlinear suppression operator is defined:

[0078] :

[0079] in, Based on the baseline signaling sensitivity, This is a hardware load suppression parameter. When the in-degree... When approaching the bottleneck, It exhibits exponential decay. By integrating active detection and passive disturbance characteristics, the comprehensive influence weight among UAVs is calculated:

[0080] :

[0081] Ultimately, drones The proportion of continuous resource input evolves according to the following continuous-time dynamic equation:

[0082]

[0083] Specifically, in step 105, the process of identifying key communication backbone links based on the Strongly Connected Component (SCC) decomposition algorithm is as follows:

[0084] To prevent formation disintegration caused by drones breaking links during the process of seeking advantages and avoiding disadvantages, the system background runs a Tarjan algorithm based on depth-first search to analyze the current directed communication graph before each topology evolution cycle. Perform strongly connected component identification.

[0085] Extract the minimum set of critical edges that maintains a globally strong connectivity state from the entire network link set, denoted as the protected backbone link set. :

[0086]

[0087] In subsequent actions, it belongs to The air links are subject to hard constraints, forcibly preventing disconnection operations. The remaining links are divided into flexible link sets. It allows for dynamic refactoring.

[0088] Specifically, in step 106, the process of driving the self-organizing evolution of the cluster topology by combining the overall total revenue and the continuous reputation evaluation results is as follows:

[0089] First, calculate the relative performance satisfaction index of the UAV within its local communication neighborhood. :

[0090]

[0091] in, and They are nodes The maximum and minimum combined total revenue among the observed neighboring nodes. If Below the set tolerance limit If so, the drone is activated and enters the evolution queue.

[0092] For flexible link sets Each outbound collaborative link in drones will be based on probability Actively disconnect from poorly behaving neighbors Connection:

[0093]

[0094] in, This refers to the sensitivity to disruptions caused by selfish behavior. The lower the resource investment from the other party, the higher the probability of disruption.

[0095] After disconnection, the drone broadcasts probe signals to non-neighboring nodes to search for potential high-quality collaborators. (must meet) (Greater than the local average level). And adopts a proportionality to... ( New links are established based on the probability of preference coefficients to guide clustered collaboration among high-reputation, high-contribution drones and complete the resilient topology reconstruction of the edge network.

[0096] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0097] In another embodiment, a continuous reputation-driven UAV swarm resource collaborative allocation system is provided, which corresponds one-to-one with the continuous reputation-driven UAV swarm resource collaborative allocation method in the above embodiments. The system specifically includes:

[0098] Environmental perception and topology construction module: Used to periodically acquire the initial physical parameters and operating status of each node in the UAV cluster (including maximum CPU computing power, remaining battery power, concurrent communication in-degree and wireless signal-to-noise ratio SNR), and abstract the computing power offloading and communication relay requests between UAVs into a time-varying directed interactive topology matrix.

[0099] Continuous Action Game Theory and Profit Calculation Module: This module uses a game theory modeling method based on continuous action space to map the cooperative behavior of UAVs to the proportion of continuous resource input within the interval [0,1]; and calculates the basic net profit of UAV nodes by comprehensively considering the system-level gain from receiving external support and the internal battery energy consumption cost.

[0100] Tolerance Filtering and Continuous Reputation Evaluation Module: This module samples UAV strategies within a preset time window and introduces a physical channel fading tolerance threshold to filter jitter caused by environmental noise. It calculates a dynamic continuous reputation score based on behavioral temporal consistency, identifies hidden betrayal nodes, and superimposes the reputation score with the basic net income to output the comprehensive total income.

[0101] Load suppression and asymmetric learning module: Used to monitor the in-degree of concurrent communication of the core UAV in real time, and trigger a nonlinear suppression operator when the in-degree approaches the hardware processing bottleneck; use this operator to dynamically reduce the weight of external collaborative requests to prevent signaling overwhelming, and generate random numbers locally to complete the stable update and evolution of the resource investment ratio.

[0102] SCC Backbone Link Identification and Protection Module: Before the topology reconfiguration command is issued, the Tarjan depth-first search algorithm is used to identify strongly connected components (SCCs) in the current UAV directed graph, extract and lock the minimum critical backbone link set that maintains the connectivity of the entire network, and force immune disconnection operations to prevent the cluster from losing connection.

[0103] Reputation-driven topology self-organizing evolution module: Used to activate the evolution queue based on the drone's satisfaction with local resource allocation, control the drone to actively disconnect from inferior cooperative links with low-reputation neighbors with a specific probability, and prioritize establishing new connections with potential high-reputation, high-quality drones in the entire network, outputting the optimal resource scheduling topology that ensures global connectivity.

[0104] Specific limitations regarding the continuous reputation-driven UAV swarm resource collaborative allocation system can be found in the above description of the limitations of a continuous reputation-driven UAV swarm resource collaborative allocation method, and will not be repeated here. Each module in the above system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0105] In another embodiment of the present invention, a terminal device is provided, comprising a processor and a memory. The memory stores a computer program, which includes program instructions. The processor executes the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions to achieve corresponding method flows or corresponding functions. The processor described in this embodiment can be used in the operation of a resource collaborative allocation method for unmanned aerial vehicle (UAV) swarms based on continuous reputation.

[0106] In another embodiment of the present invention, a storage medium is provided, specifically a computer-readable storage medium (Memory). This computer-readable storage medium is a memory device in a terminal device used to store programs and data. It is understood that the computer-readable storage medium here can include both the built-in storage medium in the terminal device and extended storage media supported by the terminal device. The computer-readable storage medium provides storage space that stores the terminal's operating system. Furthermore, this storage space also stores one or more instructions suitable for loading and execution by a processor. These instructions can be one or more computer programs (including program code). It should be noted that the computer-readable storage medium here can be high-speed RAM or non-volatile memory, such as at least one disk storage device.

[0107] One or more instructions stored in a computer-readable storage medium can be loaded and executed by a processor to implement the corresponding steps of the resource collaborative allocation method for UAV swarms based on continuous reputation in the above embodiments; one or more instructions in the computer-readable storage medium are loaded and executed by a processor.

[0108] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0109] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0110] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for collaborative resource allocation in a drone swarm based on continuous reputation, applied to a collaborative computing network system consisting of multiple drone nodes, characterized in that... Includes the following steps: S101: Obtain the initial physical parameters and operating status of each node in the UAV cluster, and construct a time-varying directed interaction topology that characterizes the computing power offloading or communication relay request link between UAV nodes. S102, based on game modeling of continuous action space, calculate the basic net income of each UAV node; wherein, the cooperative behavior of UAV nodes is mapped to the continuous resource input ratio in the interval [0,1], and the basic net income is the algebraic sum of the system cooperative gain of receiving computing power support from neighboring nodes and the endogenous energy consumption cost generated by its own resource input. S103, within a preset sliding time window, based on the change in the resource investment ratio and the preset channel fading tolerance threshold, calculate the dynamic continuous reputation score of each UAV node, and superimpose the dynamic continuous reputation score with the basic net income to obtain the comprehensive total income of each UAV node. S104. Based on the total comprehensive benefit, and according to the concurrent communication in-degree of each UAV node, calculate the nonlinear suppression operator and implement asymmetric learning to update the resource investment ratio of each UAV node. S105, based on the Strongly Connected Components (SCC) decomposition algorithm, identifies and extracts the set of key communication backbone links that maintain the global communication connectivity of the UAV cluster. S106, based on the comprehensive total revenue and the dynamic continuous reputation score, drive the topology self-organizing evolution of the drone swarm: when disconnecting from low-reputation neighbor nodes, apply immune disconnection protection to the links in the set of key communication backbone links, and guide drone nodes to establish new connections with potential high-reputation collaborators.

2. The method according to claim 1, characterized in that, In step S101, constructing the time-varying directed interaction topology includes: The drone swarm network is abstracted as a directed graph that evolves over time. ,in For a collection of drone nodes, Let be the set of request links at time t; Using binary adjacency matrix Characterizing the physical topology, when drone nodes Towards When sending a collaboration request, ,otherwise ; Calculate the concurrent communication out-degree of each drone node. Concurrent communication in-degree .

3. The method according to claim 2, characterized in that, In step S102, the drone node exist The resource allocation ratio at any given time is Its basic net income function The calculation formula is: ， in, These are interaction variables defined based on directed adjacency relationships. When the elements of the directed adjacency matrix... or hour, ;otherwise, ; Indicates drone node The undirected physical interaction degree, whose value is ; This represents the endogenous energy consumption cost per unit of resource input; This represents the system cooperative gain resulting from receiving computing power support from neighboring nodes; the system parameters satisfy... .

4. The method according to claim 3, characterized in that, In step S103, the number of UAV nodes is calculated. Dynamic continuous credit scoring include: Define physical channel tolerance threshold And based on the change in the proportion of continuous resource input. With the threshold Update the consistency count value based on the comparison results. ; According to the formula Calculate the dynamic continuous reputation score, where c is the credit incentive strength coefficient; i (t) represents the time consistency count value; According to the formula Calculate the total comprehensive revenue. .

5. The method according to claim 4, characterized in that, In step S104, the implementation of asymmetric learning to update the resource allocation ratio includes: Based on the aforementioned total revenue and The difference Drone nodes in a computer cluster Mimicking neighbor nodes The basic transition probability of the strategy ; Based on drone nodes concurrent communication in-degree Calculate the nonlinear suppression operator : ,in Based on the baseline signaling sensitivity, This refers to hardware load suppression parameters; Based on the aforementioned basic transition probability Nonlinear suppression operator And adjacency relationships, calculate the comprehensive influence weight between nodes. ; Based on the continuous-time dynamics equation Update drone nodes resource input ratio .

6. The method according to claim 1, characterized in that, In step S105, identifying the set of key communication backbone links specifically involves: before each topology evolution cycle, analyzing the current directed communication graph. Run the Strongly Connected Components (SCC) decomposition algorithm to extract and lock the minimum set of key edges that maintains the globally strongly connected state. , as a set of protective backbone links that are subject to forced immune disconnection operations.

7. The method according to claim 1, characterized in that, In step S106, the driving topology self-organizing evolution includes: Calculate the relative performance satisfaction index of each UAV node within its local communication neighborhood. ); like Below the preset tolerance limit If so, the corresponding drone node will be activated and enter the evolution queue; For unmanned aerial vehicle nodes in the evolutionary queue With probability Actively disconnect it from neighboring nodes The unprotected links between them, Sensitivity to the disruption of the chain of selfish behavior, For neighboring nodes The proportion of resources invested; Guide the drone node that has lost its link proportional to potential collaborators resource input ratio The probability of the potential collaborator Establish a new connection, in which This is the preference coefficient.

8. A resource collaborative allocation system for unmanned aerial vehicle (UAV) swarms based on continuous reputation-driven mechanisms, characterized in that, For performing the method according to any one of claims 1 to 7, comprising: The environmental perception and topology construction module is used to obtain the initial physical parameters and operating status of each node in the UAV cluster, and to construct a time-varying directed interactive topology. The continuous action game and payout calculation module is used for game modeling based on the continuous action space and to calculate the basic net payout of each drone node. The tolerance filtering and continuous reputation evaluation module is used to calculate the dynamic continuous reputation score and overall revenue of each UAV node based on the change in the resource investment ratio and the preset channel fading tolerance threshold. The load suppression and asymmetric learning module is used to update the resource investment ratio of each UAV node through an asymmetric learning mechanism based on the total comprehensive revenue and the concurrent communication in-degree of each node. The SCC backbone link identification and protection module is used to identify and extract the set of key communication backbone links that maintain the global communication connectivity of the UAV cluster based on the strongly connected component SCC decomposition algorithm. A reputation-driven topology self-organizing evolution module is used to drive the self-organizing evolution of the cluster topology based on the comprehensive total revenue and the dynamic continuous reputation score, while protecting critical backbone links.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program that, when executed by a processor, implements the method as claimed in any one of claims 1 to 7.

Images