A multi-task allocation algorithm optimization method for complex task scenarios

Abstract

The application provides a kind of multi-task allocation algorithm optimization method for complex task scene, comprising: based on task chain and OODA theory, considering node attribute, heterogeneity and edge directionality on the basis of task chain model, establish effective task net model;According to the effective task net model, a dynamic reconstruction strategy after random failure and deliberate attack failure is established;Initialize the effective task net model, analyze the failure of unmanned cluster nodes and edges, use adjacency matrix and arrival matrix to realize the task success rate of different task allocation algorithms;Through the numerical simulation results of the task success rate of different task allocation algorithms, the advantages and disadvantages of the allocation algorithm are evaluated, and the optimization analysis of the task allocation algorithm under different task scenes is realized.The application greatly improves the task success rate of the cluster in the complex task environment, the task success rate based on the effective task net is simulated and optimized for the task allocation and execution process, effectively supports the control decision of the cluster, and improves the task execution efficiency of the cluster.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of unmanned aerial vehicle clusters, and particularly relates to a multi-task allocation algorithm optimization method for complex task scenarios. BACKGROUND

[0002] An unmanned cluster is a new task execution force formed by multiple unmanned device systems and related support systems that are functionally connected and interacted, and cooperatively work and are organically integrated.

[0003] An unmanned cluster is composed of heterogeneous unmanned systems, including unmanned aerial vehicles, unmanned vehicles, unmanned underwater vehicles, etc. An unmanned cluster usually performs complex tasks, which can be decomposed into different subtasks, and different tasks require different systems to complete. An unmanned cluster is complex and highly cooperative, and task success rate is a core element reflecting whether it can run smoothly and normally. The task success rate of an unmanned cluster is not only related to the stability and performance of a single device, but also directly reflects the task execution efficiency of the entire cluster.

[0004] Due to the characteristics of node heterogeneity, link directionality and task network reconfigurability of an unmanned cluster, traditional complex network-based cluster task success rate evaluation methods cannot reflect its characteristics. At the same time, based on multi-agent or task execution simulation methods, the calculation efficiency is low, and it is difficult to realize the dynamic evaluation of the task success rate of the unmanned cluster for task-oriented rapid response.

[0005] In the evaluation of unmanned cluster task allocation algorithms, the task success rate index is a key evaluation standard to measure the performance of the algorithm under a specific task. This index mainly reflects the proportion of successfully completed allocated tasks in the execution process of the unmanned cluster. Simply put, the task success rate is usually represented by the ratio of the number of completed tasks to the total number of tasks.

[0006] Specifically, the following aspects need to be considered when evaluating the task success rate index: (1) Task completion definition: clearly define what constitutes a successful completion of a task. For example: accurate arrival at the target location, correct execution of specific operations, etc. (2) Time factor: consider whether the task is completed within the specified time, which is particularly important for tasks with high real-time requirements. (3) Task value: different values of tasks may affect the allocation results of the allocation algorithm, and thus affect the task success rate, so factors such as task value should be considered in the evaluation. (4) Internal and external factors: changes in external environment (such as weather, terrain, etc.) and physical and electromagnetic interference of the task target and the reliability of the device itself will also affect the task success rate. (5) Communication and coordination: the task success rate of an unmanned cluster is closely related to the communication efficiency and coordination ability between its nodes, and a good communication and coordination mechanism can improve the overall task success rate.

[0007] Through comprehensive evaluation of the above aspects, the task success rate index can effectively reflect the advantages and disadvantages of the task allocation algorithm of the unmanned cluster, and provide an important reference for algorithm improvement and optimization. However, the traditional task allocation strategy evaluation index mainly focuses on the running time and resource consumption of the algorithm, and often ignores the influence of the task success rate. SUMMARY

[0008] The present application aims to solve the problem of ignoring the task success rate in the prior art, and provides a multi-task allocation algorithm optimization method for complex task scenarios, so that the unmanned cluster can effectively utilize the cluster resources, reasonably allocate tasks, maximize the target value, greatly improve the task success rate of the cluster in the complex task environment, and based on the task success rate of the effective task network, simulate and optimize the task allocation and execution process, effectively support the control decision of the cluster, and improve the task execution efficiency of the cluster.

[0009] To achieve the above-mentioned purpose, the technical solution provided by the present application is:

[0010] A multi-task allocation algorithm optimization method for complex task scenarios, comprising:

[0011] Step one, based on the task chain and the OODA theory, considering the node attribute, heterogeneity and edge directionality on the basis of the task chain model, an effective task network model is established;

[0012] Step two, according to the effective task network model, a dynamic reconstruction strategy after random failure and deliberate attack failure is established;

[0013] Step three, initializing the effective task network model, analyzing the failure of the unmanned cluster nodes and edges, and using the adjacency matrix and the arrival matrix to realize the task success rate of different task allocation algorithms;

[0014] Step four, through the task success rate numerical simulation results of different task allocation algorithms, the advantages and disadvantages of the allocation algorithm are evaluated, and the task allocation algorithm optimization analysis under different task scenarios is realized.

[0015] As a further improvement of the present application, the expression of the effective task network model is:

[0016] eon={A,V,E}

[0017] In the formula, A represents the set of adjacency matrices in the task chain, V represents the set of nodes in the task chain, and E represents the set of edges in the task chain; specifically:

[0018] A={A TS ,A SD ,A DW ,A WT}

[0019] wherein A TS denotes the adjacency matrix set of the detection device discovering the task target and obtaining certain information data, A SD denotes the adjacency matrix set of the detection device uploading the detected information to the decision device, A DW denotes the adjacency matrix set of one decision device receiving the control instruction from another decision device, A WT denotes the adjacency matrix set of the influence device affecting the task target;

[0020] V = (S, D, W, T)

[0021] wherein S denotes the detection node, D denotes the decision node, W denotes the influence node, and T denotes the communication link;

[0022]

[0023] s i (i = 1, 2,..., I)

[0024] d j (j = 1, 2,..., J)

[0025] w m (m = 1, 2,..., M)

[0026] t n (n = 1, 2,..., N)

[0027] wherein denotes the edge set of the network between the nth target node and the ith detection node, denotes the edge set of the network between the ith detection node and the jth decision node, denotes the edge set of the network between the jth decision node and the mth influence node, denotes the edge set of the network between the mth influence node and the nth target node, i denotes the ith detection node, and I denotes the number of the ith detection node, j denotes the jth decision node, and J denotes the number of the jth decision node, m denotes the mth influence node, and M denotes the number of the mth influence node, n denotes the nth target node, and N denotes the number of the nth target node.

[0028] As a further improvement of the present application, the step two comprises:

[0029] Step (21) analyzes the performance of the unmanned cluster device, establishes an unmanned cluster node model, and sets node attribute indicators according to the effective task network model and the performance and characteristics of the unmanned cluster device.

[0030] Step (22) establishes an unmanned cluster node edge model and sets edge attribute indicators based on the unmanned cluster node model and considering device communication and resource limitations.

[0031] As a further improvement of the present application, the step (21) is specifically:

[0032] (211) For the detection node S mainly including the detection device, there is a significant performance difference between different detection devices; the attribute model of the detection node S is established as:

[0033]

[0034] In the formula, s i represents a node in the detection node S, I represents the number of the i-th detection node, represents the failure rate of the detection node s i , represents the repair rate of the detection node s i , represents the degree of the detection node s i , represents the x-axis coordinate of the detection node s i , represents the y-axis coordinate of the detection node s i , cluster k represents that the node s i is a member of the cluster K, represents the maximum communication distance of the node s i , represents the maximum detection distance;

[0035] (212) For the decision node D, the main performance indicators include response time, throughput and accuracy, the detection and influence tasks of the decision node jointly affect the generation and effect of the whole OODA cycle and the task chain; the communication ability of the decision node directly affects its ability to issue tasks to other nodes; the attribute model of the decision node D is established as follows:

[0036]

[0037] In the formula, d j represents the j-th decision node, J represents the number of the j-th decision node, represents the failure rate of the d j node, represents the repair rate of the d jRepair rate of nodes, represents d j Degree of nodes, represents the x-axis position coordinate of the jth decision node, represents the y-axis position coordinate of the jth decision node, cluster k represents that the jth decision node is a member of cluster K, represents the maximum communication distance of the jth decision node;

[0038] (213) Different target nodes T exhibit different characteristics, and the main performance indicators include: target value, detection difficulty and task completion difficulty; The attribute model of the target node is established as:

[0039]

[0040] In the formula, t n represents the nth target node, and N represents the number of the nth target node, represents t n Failure rate of nodes, represents t n Repair rate of nodes, represents the x-axis position coordinate of the nth target node, represents the y-axis position coordinate of the nth target node, cluster k represents that the nth target node is a member of cluster K, represents the target value.

[0041] As a further improvement of the present application, the step (22) is specifically:

[0042] Edge Considering the allocation result of the detection task and the detection probability, the unmanned cluster node edge model is established as follows:

[0043]

[0044] In the formula, represents the distance between the nth target node t n and the ith detection node s i , represents the result of the detection task allocation, represents the detection probability of the nth target node t n to the ith detection node s i ;

[0045] Unmanned cluster node edge model is established as follows:

[0046]

[0047] wherein, denotes the distance between the ith probe node s i and the jth decision node d j , denotes the communication task success rate between the ith probe node s i and the jth decision node d j , denotes the packet loss rate between the ith probe node s i and the jth decision node d j ;

[0048] Unmanned cluster node edge model is established as follows:

[0049]

[0050] wherein, denotes the distance between the jth decision node d j and the mth influence node w m , denotes the communication task success rate between the jth decision node d j and the mth influence node w m , denotes the packet loss rate between the jth decision node d j and the mth influence node w m ;

[0051] Unmanned cluster node edge model is established as follows:

[0052]

[0053] wherein, denotes the distance between the mth influence node w m and the nth target node t n , denotes the assignment result of the influence task, denotes the influence task completion probability of the mth influence node w m on the nth target node t n .

[0054] As a further improvement of the present application, the step three comprises:

[0055] Step (31) initializes the simulation times, simulation time, and node edge attributes of the effective task network model, and the connection relationship between part of the nodes to obtain the adjacency matrix A SD , A DW ;

[0056] Step (32) uses various task allocation algorithms to allocate influence tasks and detection tasks to the unmanned cluster considering the value of the task, the cost of the influence factor, and obtains an adjacency matrix A TS , WT ;

[0057] Step (33) If the node and the edge exist, that is, no failure occurs, then α() = 1; if the node and the edge do not exist, that is, failure occurs, then α() = 0, wherein α() represents the existence of the node and the edge; the node indicator function is as follows:

[0058]

[0059] In the formula, s i represents the i th detection node, d j represents the j th decision node, and w m represents the m th influence node.

[0060] The edge existence indicator function is as follows:

[0061]

[0062] In the formula, represents the edge set between the i th detection node and the j th decision node, represents the edge set between the n th target node and the i th detection node, represents the edge set between the j th decision node and the m th influence node, represents the edge set between the m th influence node and the n th target node.

[0063] Step (34) calculates the number of effective task chains N eol (t) using the adjacency matrix and the reach matrix, and integrates the task success criterion N eol (t) ≥ τ i to calculate the task success rate; wherein τ i represents the minimum number of effective task chains required for the success of task i.

[0064] As a further improvement of the present application, step (33) further comprises a random failure step and a deliberate attack step, specifically:

[0065] The random failure step comprises: determining the number of failed nodes and edges of the unmanned cluster according to the node failure rate and the edge failure rate of the corresponding unmanned cluster node model and the corresponding unmanned cluster node edge model and their distributions, randomly removing the failed nodes and edges of the unmanned cluster node model, and adding the failed nodes and edges to the failed node list and the failed edge list, respectively; wherein s i is a detection node, dj and the detection node w m Failure or death of the node w Failure rate Failure rate Is described by the Poisson distribution of the failure rate

[0066] The deliberate attack step includes: determining the attack mode, arranging the nodes in the unmanned cluster in descending order of node degree, then removing the corresponding number of failed nodes and edges, and adding the failed nodes to the failed node list; as a special case, CNA is used to study node failure and dynamic reconstruction strategy; set the node degree attack strategy to attack only the top 50% of nodes, set Indicates the degree of node s i ,d j ,w m ;

[0067] Specifically:

[0068] When the failed node list is not empty, the inter-cluster reconstruction I is used to adjust or reconstruct the connection and relationship between different clusters in the cluster, and the inter-cluster reconstruction I is a rule-based reconstruction strategy;

[0069] When the failed node list is not empty, the intra-cluster reconstruction II is used to modify the connection and relationship between nodes in the same cluster, optimize the internal structure and coordination of the unmanned cluster, and make the communication, resource allocation and task allocation between nodes more efficient;

[0070] When the failed node list is not empty, the repair and generation reconstruction III is used to repair or add new nodes or edges in the cluster, which involves deploying additional clusters or activating dormant nodes, and improves the overall capacity and capacity of the cluster;

[0071] When the task changes or the cluster structure changes greatly, the task reallocation IV is triggered, which involves the reallocation of tasks between nodes in the cluster, to optimize task allocation, maximize cluster efficiency, and ensure completion of priority task objectives.

[0072] As a further improvement of the application, step (34) is specifically:

[0073] Step (341) obtains the adjacency matrix A={A TS ,A SD ,A DW ,A WT} between different types of nodes through an effective task network model;

[0074] Step (342) verifies the existence of the adjacency matrix elements by considering the properties and existence of nodes and edges:

[0075]

[0076] If there is a connection between nodes, then Otherwise,

[0077] The existence probability of the adjacent matrix element is:

[0078]

[0079] Wherein, if the failure and generation of the node obeys the exponential distribution, the existence probability of the unmanned cluster node is:

[0080]

[0081] The existence probability of each edge of the unmanned cluster is expressed as:

[0082]

[0083] In the formula, Indicates the allocation result of the detection task, Indicates the effective communication between the ith detection node s i And the jth decision node d j , Indicates the effective communication between the mth influence node w m And the nth target node t n ; wherein:

[0084]

[0085]

[0086] All adjacent matrix elements are verified to step (343), otherwise step (342) is turned to continue verification;

[0087] Step (343) calculates the number of effective task chains:

[0088] Calculate the arrival matrix A TT , A TT =A TS ×A SD ×A DW ×A WT ;

[0089] Calculate the number of effective task chains N TT of the arrival matrix A eol , N eol =Trace(A TT ), A TS Indicates that the dimension is N×I, A SD Indicates that the dimension is I×J, A DW Indicates that the dimension is J×M, and A WT Indicates that the dimension is M×N;

[0090] Task success rate calculation i = 1, judge whether the task chain is a valid task chain

[0091] If N eol ≥ τ i Indicate that the task is effectively completed, the task success constitutes a valid task chain, and the number of successful tasks Num task-success = Num task-success + 1, i = i + 1

[0092] If N eol < τ i Then it indicates that the requirement of effectively completing the task is not met i = i + 1; if i = n t Then turn to the next step, otherwise turn to this step; n t Is the number of T nodes

[0093] According to the number of simulations, the task success rate is calculated

[0094] As a further improvement of the application, the dynamic reconstruction strategy of step two includes: rule-based dynamic reconstruction and task-based dynamic reconstruction

[0095] The task success rate index is adopted to evaluate the task allocation strategy of the unmanned cluster, which combines the actual completion process and the completion result of the task, and forms an effective completion task as the task success standard

[0096] The step three uses the adjacency matrix and the arrival matrix of the unmanned cluster to calculate the number of valid task chains.

[0097] The advantages of the application are:

[0098] 1. The application selects the optimal algorithm for different task allocation algorithms in different task scenarios, so that the unmanned cluster can effectively utilize the cluster resources, reasonably allocate tasks, maximize the target value, and greatly improve the task success rate of the cluster in a complex task environment.

[0099] 2. The method has universality and can be popularized to the optimization of various types of unmanned cluster task allocation strategies, and only needs to specify the device and environmental parameters to complete the modeling simulation, analysis and optimization of large-scale unmanned cluster task process, effectively support the control decision of the cluster, improve the task execution efficiency of the cluster, provide solid and reliable decision support for task action, and provide theoretical and methodological support for reasonable scheduling of equipment and resources.

[0100] Additional aspects and advantages of the application will be given in part in the following description, part will become apparent from the following description, or will be understood by practicing the application. Attached Figure Description

[0101] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which:

[0102] Figure 1 This invention provides a flowchart of a method for selecting the best among multiple task allocation algorithms for complex task scenarios;

[0103] Figure 2 The flowchart of the task allocation algorithm provided by this invention;

[0104] Figure 3 The node distribution diagram of the unmanned cluster provided by this invention;

[0105] Figure 4 : A model diagram of the effective task network model for unmanned swarms provided by this invention;

[0106] Figure 5 The impact diagram of arbitrary failures and deliberate attacks on unmanned swarms provided by this invention;

[0107] Figure 6 : A schematic diagram of a rule-based unmanned cluster reconstruction strategy provided by this invention;

[0108] Figure 7 : A schematic diagram of a task-based unmanned cluster reconfiguration strategy provided by this invention;

[0109] Figure 8 : Impact of different reconstruction strategies on the success rate of unmanned cluster tasks;

[0110] Figure 9 : Task success rate analysis chart of three unmanned cluster task allocation algorithms. Detailed Implementation

[0111] The embodiments of the present invention are described in detail below. These embodiments are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0112] To address the issue of neglecting task success rate in optimizing task allocation for unmanned clusters, and to address the problem of significant discrepancies between the chosen optimization metric and actual results due to the numerous task allocation algorithms available in various task scenarios, please refer to [link to relevant documentation]. Figure 1The embodiment of the application provides a multi-task allocation algorithm optimization method for a complex task scene, and the success rate of a task network is used to simulate and optimize the task allocation and execution process; the unmanned cluster of the embodiment of the application is composed of 240 nodes, wherein 40 nodes are detection nodes, 20 nodes are decision nodes, 120 nodes are influence nodes, and 60 nodes are target nodes; the task scene of the unmanned cluster is a square region of 200*200, and the distribution of the nodes in the task scene is as shown in the figure Figure 2 The specific implementation method of the embodiment of the application is as shown in the figure Figure 3 The method of the embodiment comprises the following steps:

[0113] Step one, based on the task chain and the OODA theory, the node attribute, heterogeneity and edge directionality are considered based on the task chain model, and an effective task network model is established; the embodiment of the application characterizes the heterogeneity, directionality, diversity and emergence of the unmanned cluster through the effective task network model.

[0114] The expression of the effective task network model of the embodiment of the application is as follows:

[0115] eon={A,V,E}

[0116] In the formula, A represents the set of adjacent matrices in the task chain, V represents the set of nodes in the task chain, and E represents the set of edges in the task chain; specifically,

[0117] A={A TS ,A SD ,A DW ,A WT}

[0118] In the formula, A TS represents the set of adjacent matrices in which the detection equipment discovers the task target and obtains certain information data, A SD represents the set of adjacent matrices in which the detection equipment uploads the detected information to the decision equipment, A DW represents the set of adjacent matrices in which one decision equipment receives the control instruction from another decision equipment, and A WT represents the set of adjacent matrices in which the influence equipment targets the task target.

[0119] V=(S,D,W,T)

[0120] In the formula, S represents the detection node, D represents the decision node, W represents the influence node, and T represents the communication link.

[0121]

[0122] s i(i = 1, 2,..., I)

[0123] d j (j = 1, 2,..., J)

[0124] w m (m = 1, 2,..., M)

[0125] t n (n = 1, 2,..., N)

[0126] In the formula, represents the edge set of the network between the nth target node and the ith detection node, represents the edge set of the network between the ith detection node and the jth decision node, represents the edge set of the network between the jth decision node and the mth influence node, represents the edge set of the network between the mth influence node and the nth target node, i represents the ith detection node, I represents the number of the ith detection node, j represents the jth decision node, J represents the number of the jth decision node, m represents the mth influence node, M represents the number of the mth influence node, n represents the nth target node, N represents the number of the nth target node.

[0127] In an ideal state, the nodes of the unmanned cluster can be connected to each other to form a fully connected network. However, in actual task scenarios, the nodes of different types of systems are limited by their own properties and running resources. According to the functions, the nodes are divided into four categories S, D, W and T, and let s i (i = 1, 2,..., I) represents the ith detection node, d j (i = 1, 2,..., J) represents the jth decision node, w m (m = 1, 2,..., M) represents the mth influence node, t n (n = 1, 2,..., N) represents the nth target node, I, J, M and N represent the number of nodes of different types; and an effective task network model is as shown in Figure 4 A = {A TS ,A SD ,A DW ,A WT} represents a set of adjacency matrices in the network, V represents a set of nodes in the network, wherein the nodes are divided into four categories V = (S, D, W, T), and E represents a set of edges in the network, wherein the edges are divided into five categories Due to resource sharing and information fusion, Figure 4In our unmanned cluster, nodes of the same type can exchange information and data in real time, so communication and interaction between nodes of the same type are not considered.

[0128] Step 2: Based on the effective task network model, establish a dynamic reconstruction strategy after random failures and deliberate attack failures. This embodiment of the invention considers the differences in performance attributes of unmanned cluster devices and the characteristics of communication task allocation to establish unmanned cluster node models and connection models; comprehensively considering internal and external influencing factors such as the environment of the unmanned cluster task (weather, temperature, humidity, etc.), the impact of task objectives, and the success rate of the device's own tasks, it classifies them into two categories: random failures and deliberate attacks, realizing the analysis of unmanned cluster node and connection failures.

[0129] The dynamic reconstruction strategy of this invention includes rule-based dynamic reconstruction and task-based dynamic reconstruction. For an effective task network model, this invention considers the attributes and resource constraints of unmanned cluster devices to establish an unmanned cluster node and connection model, and provides a given task simulation time.

[0130] Step two of this embodiment of the invention includes:

[0131] Step (21) Based on the effective task network model, considering the performance and characteristics of the unmanned cluster equipment, analyze the performance of the unmanned cluster equipment, establish the unmanned cluster node model, and set the node attribute indicators.

[0132] In this embodiment of the invention, for detection nodes, which mainly include detection equipment, there may be significant performance differences between different detection devices. Let... Represents node s i The maximum detection range, if the distance to the target exceeds If the detection node fails to detect the target, it will be unable to detect it. The detection probability of a detection node refers to the probability that the node successfully detects the target within its perception range. The attribute model of detection node S is established as follows:

[0133] (211) For the detection node S, which mainly includes detection equipment, there are significant performance differences among different detection equipment; the attribute model of the detection node S is established as follows:

[0134]

[0135] In the formula, s i Let I represent the number of nodes in probe node S, and let I represent the number of the i-th probe node. Indicates the probe node s i failure rate, Indicates the probe node s i Repair rate Indicates the probe node s i The degree, denotes the x-axis coordinate of the probe node s i denotes the y-axis coordinate of the probe node s i cluster k denotes that node s i is a member of cluster K, denotes the maximum communication distance of node s i denotes the maximum probe distance.

[0136] (212) For the decision node D, the main performance indicators include response time, throughput and accuracy, and the probe and influence task allocation results of the decision node jointly affect the generation and effect of the whole OODA cycle and task chain; the communication ability of the decision node directly affects its ability to issue tasks to other nodes, and the attribute model of the decision node D is established as follows:

[0137]

[0138] In the formula, d j denotes the jth decision node, and J denotes the number of the jth decision node, denotes the failure rate of d j node, denotes the repair rate of d j node, denotes the degree of d j node, denotes the x-axis position coordinate of the jth decision node, denotes the y-axis position coordinate of the jth decision node, cluster k denotes that the jth decision node is a member of cluster K, denotes the maximum communication distance of the jth decision node;

[0139] (213) Different target nodes T exhibit different characteristics, and the main performance indicators include target value, probe difficulty and task completion difficulty; the attribute model of the target node is established as follows:

[0140]

[0141] In the formula, t n denotes the nth target node, and N denotes the number of the nth target node, denotes the failure rate of t n node, denotes the repair rate of t n node, denotes the x-axis position coordinate of the nth target node, denotes the y-axis position coordinate of the nth target node, cluster​​k represents the nth target node is a member of the cluster K, represents the target value.

[0142] Step (22) is based on the unmanned cluster node model, considering the device communication and resource constraints, establishes an unmanned cluster node edge model, and sets edge attribute indexes.

[0143] The step (22) of the embodiment of the application is specifically:

[0144] Edge Considering the allocation result of the detection task and the detection probability, the unmanned cluster node edge model is established as follows:

[0145]

[0146] In the formula, represents the distance between the nth target node t n and the ith detection node s i , represents the result of the detection task allocation, represents the detection probability of the nth target node t n to the ith detection node s i ;

[0147] The unmanned cluster node edge model is established as follows:

[0148]

[0149] In the formula, represents the distance between the ith detection node s i and the jth decision node d j , represents the communication task success rate between the ith detection node s i and the jth decision node d j , represents the packet loss rate between the ith detection node s i and the jth decision node d j ;

[0150] The unmanned cluster node edge model is established as follows:

[0151]

[0152] In the formula, represents the distance between the jth decision node d j and the mth influence node w m , represents the distance between the jth decision node dj a communication task success rate of the mth influence node w m denotes a jth decision node d j a packet loss rate of the mth influence node w m

[0153] An unmanned cluster node edge model is established as follows:

[0154]

[0155] In the formula, denotes a distance between the mth influence node w m and the nth target node t n denotes an assignment result of an influence task, denotes an influence task completion probability of the mth influence node w m to the nth target node t n .

[0156] Embodiments of the present application cluster and group according to task and area distribution for the same unmanned cluster, and collaboratively complete different subtasks.

[0157] For an influence node (W), different types of W have different capabilities, and main performance indicators include influence accuracy, maximum task completion distance and completion probability. A property model of W is established as follows:

[0158]

[0159] In the formula, denotes two-dimensional position coordinates of w m , cluster k denotes that w m is a member of a cluster K, and denotes a maximum communication distance and a maximum task completion distance of w m .

[0160] Step three, initializing an effective task network model, analyzing unmanned cluster nodes and edges, and using an adjacency matrix and a reach matrix to realize a task success rate of different task assignment algorithms. Embodiments of the present application use an unmanned cluster adjacency matrix and a reach matrix to calculate an effective task chain quantity. Embodiments of the present application extend a traditional task network model, and consider comprehensive influences of internal and external factors such as random failure, deliberate attack, task assignment result, key performance indicators and task resource limitations in an unmanned cluster task execution process.

[0161] Step three of embodiments of the present application includes: ​​​

[0162] Step (31) initializes the simulation times of the effective task network model, the simulation time and the node edge attribute, and the connection relationship between part of the nodes to obtain the adjacency matrix A SD ,A DW ;

[0163] Step (32) considers the task value and influence cost factors, and uses various task allocation algorithms to allocate the influence task and the detection task of the unmanned cluster, and obtains the adjacency matrix A TS ,A WT ;

[0164] Step (33) If the node and the edge exist, that is, no failure occurs, then α()=1; if the node and the edge do not exist, that is, failure occurs, then α()=0, wherein α() represents the existence of the node and the edge; the node indicator function is as follows:

[0165]

[0166] In the formula, s i represents the i-th detection node, d j represents the j-th decision node, and w m represents the m-th influence node.

[0167] The edge existence indicator function is as follows:

[0168]

[0169] In the formula, represents the edge set between the i-th detection node and the j-th decision node, represents the edge set between the n-th target node and the i-th detection node, represents the edge set between the j-th decision node and the m-th influence node, represents the edge set between the m-th influence node and the n-th target node.

[0170] In the formula, the step (33) of the embodiment of the application further includes a random failure step and a deliberate attack step, specifically:

[0171] The random failure step includes: according to the node failure rate and the edge failure rate of the corresponding unmanned cluster node model and the corresponding unmanned cluster node edge model and their distributions, determining the number of node and edge failures of the unmanned cluster through Monte Carlo simulation, randomly removing the failed nodes of the unmanned cluster node model and their edges, and adding the failed nodes and edges to the failed node list and the failed edge list respectively; wherein the failure or death of the detection node s i , the detection node d j and the detection node w m is determined by the failure rate parameter failure rate failure rate It is described by the Poisson distribution.

[0172] The steps of a deliberate attack include: determining the attack pattern; sorting the nodes in the unmanned cluster in descending order of node degree; removing a corresponding number of failed nodes and their edges; adding the failed nodes to the failed node list; using CNA as a special case to study node failure and dynamic reconstruction strategies; and setting the node degree attack strategy to only attack the top 50% of nodes. Indicates the probe node s i The degree, Indicates the probe node d j The degree, Indicates the probe node w m The degree.

[0173] Specifically:

[0174] When the list of failed nodes is not empty, the process of adjusting or reconstructing the connections and relationships between different clusters within the cluster is adopted by inter-cluster reconstruction I. Inter-cluster reconstruction I is a rule-based reconstruction strategy.

[0175] When the list of failed nodes is not empty, the cluster reconstruction II is used to modify the connections and relationships between nodes in the same cluster, optimize the internal structure and coordination of the unmanned cluster, and make communication, resource allocation and task allocation between nodes more efficient.

[0176] When the list of failed nodes is not empty, Repair and Reconstruction III repairs or adds new nodes or edges within the cluster, involving the deployment of additional clusters or the activation of dormant nodes, thereby improving the overall capacity and capability of the cluster.

[0177] When tasks change or the cluster structure undergoes large-scale changes, Task Reassignment IV is triggered, which involves the redistribution of tasks among nodes within the cluster to optimize task allocation, maximize cluster efficiency, and ensure the completion of priority task objectives.

[0178] When tasks change or the cluster structure undergoes large-scale changes, Task Reassignment IV (a task-based reconfiguration strategy) is triggered: this involves the redistribution of tasks among nodes within the cluster. This can occur for various reasons, such as changes in task requirements, node failures, or the need to balance workloads. The purpose of task reassignment is to optimize task allocation, maximize cluster efficiency, and ensure the completion of priority task objectives. The simulation time exceeding halfway is selected as the trigger condition for task reassignment. For the impact of different dynamic reconfiguration strategies on unmanned clusters, please refer to [link to relevant documentation]. Figure 8The different strategies are compared with the unmanned cluster task success rate without strategy respectively, all the reconstruction strategies effectively improve the unmanned cluster task success rate, and with the increase of the number of reconstruction strategies adopted, the unmanned cluster task success rate is obviously improved.

[0179] Step (34) calculates the number of effective task chains N by using the adjacency matrix and the arrival matrix eol (t), and integrates the task success criterion N eol (t)≥τ i Calculate the task success rate; wherein: τ i represents the minimum number of effective task chains required for task i to be successful.

[0180] Step (34) of the embodiment of the application is specifically:

[0181] Step (341) obtains the adjacency matrix A={A TS ,A SD ,A DW ,A WT} between different types of nodes through the effective task network model.

[0182] Step (342) verifies the existence of the adjacency matrix element by considering the node, edge attribute and existence:

[0183]

[0184] If there is a connection relationship between the nodes, Otherwise,

[0185] The existence probability of the adjacency matrix element is:

[0186]

[0187] The verification of the existence of the element in the embodiment of the application includes the following sub-steps:

[0188] If the failure and generation of the node follow the exponential distribution, the existence probability of the unmanned cluster node is:

[0189]

[0190] The existence probability of each edge of the unmanned cluster is represented as:

[0191]

[0192] In the formula, represents the allocation result of the detection task, represents the i-th detection node s ieffective communication between the jth decision node d j , effective communication between the mth influence node w m and the nth target node t n ; wherein:

[0193]

[0194] All the elements of the adjacency matrix are verified, and the step (343) is turned to, otherwise the step (342) is continued to verify;

[0195] The step (343) calculates the number of effective task chains:

[0196] The reach matrix A TT is calculated TT =A TS ×A SD ×A DW ×A WT ;

[0197] The number of effective task chains N eol of the reach matrix A TT is calculated, N eol =Trace(A TT ), A TS represents the dimension of N×I, A SD represents the dimension of I×J, A DW represents the dimension of J×M, and A WT represents the dimension of M×N;

[0198] The task success rate is calculated i=1, and it is judged whether the task chain is an effective task chain

[0199] If N eol ≥τ i , it indicates that the task is effectively completed, the task success constitutes an effective task chain, and the number of successful tasks Num task-success =Num task-success +1, i=i+1;

[0200] If N eol <τ i , it indicates that the effective completion of the task is not satisfied i=i+1; if i=n t , the next step is turned to, otherwise the step is turned to; n t is the number of T nodes;

[0201] The task success rate is calculated according to the number of simulations

[0202] Step four, through the task success rate numerical simulation results of different task allocation algorithms, the advantages and disadvantages of the allocation algorithm are evaluated, and the optimal analysis of the task allocation algorithm under different task scenes is realized. The specific analysis of the unmanned cluster multiple task allocation algorithms in the embodiment of the present application can be referred to Figure 9 . Among them, the Hungarian algorithm (HA), the contract network algorithm (CNA) and the ant colony algorithm (ACA) are optimized. According to the task success rate, the ant colony algorithm is more stable, and the Hungarian algorithm and the contract network algorithm have higher task success rate in the initial stage of the task. Therefore, the ant colony algorithm is suitable for long-time task, and the Hungarian algorithm and the contract network algorithm are suitable for short-time task.

[0203] The embodiment of the present application adopts the task success rate index of the comprehensive task actual completion process and completion result to evaluate the unmanned cluster task allocation strategy; the task success rate index is closely combined with the effective task chain to form the effective completion task as the task success standard.

[0204] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited to this, any person skilled in the art can easily think of various equivalent modifications or replacements within the technical range disclosed by the present application, and these modifications or replacements should be covered within the protection scope of the present application.

Claims

1. A multi-task allocation algorithm optimization method for complex task scenarios, characterized in that, The application relates to a method for task allocation in an unmanned cluster, and belongs to the field of unmanned cluster technology. Step one, based on the task chain and the OODA theory, an effective task network model is established by considering node attributes, heterogeneity and edge directionality on the basis of a task chain model; wherein the expression of the effective task network model is: eon={A,V,E} In the formula, A represents a set of adjacency matrices in the task chain, V represents a set of nodes in the task chain, and E represents a set of edges in the task chain; specifically: A = {A TS ,A SD ,A DW ,A WT} wherein A TS represents the adjacency matrix set of the detection equipment discovering the task target and obtaining certain information data, A SD represents the adjacency matrix set of the detection equipment uploading the detected information to the decision equipment, A DW represents the adjacency matrix set of one decision equipment receiving the control instruction from another decision equipment, A WT represents the adjacency matrix set of the influence equipment affecting the task target; V=(S,D,W,T) In the formula, S represents a detection node, D represents a decision node, W represents an influence node, and T represents a communication link; s i (i = 1, 2,..., I) d j (j = 1, 2,..., J) w m (m = 1, 2,..., M) t n (n = 1, 2,..., N) wherein, denotes the set of edges of the network between the nth target node and the ith probe node, denotes the set of edges of the network between the ith probe node and the jth decision node, denotes the set of edges of the network between the jth decision node and the mth impact node, denotes the set of edges of the network between the mth impact node and the nth target node, s i denotes the ith probe node, I denotes the number of ith probe nodes, d j denotes the jth decision node, J denotes the number of jth decision nodes, w m denotes the mth impact node, M denotes the number of mth impact nodes, t n denotes the nth target node, N denotes the number of nth target nodes; Step two, a dynamic reconstruction strategy is established according to the effective task network model, wherein the dynamic reconstruction strategy comprises an inter-cluster reconstruction strategy, an intra-cluster reconstruction strategy or a node self-repairing strategy; Step three, the effective task network model is initialized, and the failure of unmanned cluster nodes and edges is analyzed, and the task success rate of different task allocation algorithms is realized by using an adjacency matrix and an arrival matrix; wherein the step three comprises: Step (31) initializes the simulation times, simulation time and node edge properties of the effective task network model, and the connection relationship between part of the nodes to obtain the adjacency matrix A SD , A DW ; Step (32) uses various task allocation algorithms to allocate the influence tasks and the detection tasks to the unmanned clusters considering the task value and the influence cost factors, and an adjacency matrix A is obtained TS , A WT ; Step (33) if the node and the edge exist, that is, no failure occurs, then alpha (·) = 1; if the node and the edge do not exist, that is, failure occurs, then alpha (·) = 0, wherein alpha (·) represents the existence of the node and the edge; specifically, the failure corresponding to the dynamic reconstruction strategy in the step two comprises random failure and deliberate attack failure, wherein: When the list of failed nodes is not empty, (1) the inter-cluster reconstruction I is adopted to adjust or reconstruct the connection and relationship between different clusters in the cluster, and the inter-cluster reconstruction I is a rule-based reconstruction strategy; (2) the intra-cluster reconstruction II is adopted to modify the connection and relationship between nodes in the same cluster, and the intra-cluster reconstruction II optimizes the internal structure and coordination of the unmanned cluster, so that the communication, resource allocation and task allocation among the nodes are more efficient; (3) the repair and generation reconstruction III is adopted to repair or add new nodes or edges in the cluster, which involves deploying additional clusters or activating dormant nodes, and improves the overall capacity and capacity of the cluster; When the task changes or the cluster structure changes on a large scale, the task redistribution IV is triggered, which involves the redistribution of tasks among the nodes in the cluster, so as to optimize the task allocation, maximize the efficiency of the cluster and ensure the completion of the priority task target; Step (34) calculates the number of effective task chains N by using the adjacency matrix and the arrival matrix eol (t), and integrates the task success criterion N eol (t) ≥ τ i calculates the task success rate; wherein: τ i represents the minimum number of effective task chains required for task i to be successful; Step four, the advantages and disadvantages of the allocation algorithm are evaluated through the numerical simulation results of the task success rate of different task allocation algorithms, and the optimal analysis of the task allocation algorithm in different task scenarios is realized.

2. The method of claim 1, wherein, The step two comprises: Step (21) according to the effective task network model, the performance and characteristics of the unmanned cluster equipment are considered, the performance of the unmanned cluster equipment is analyzed, the unmanned cluster node model is established, and the node attribute index is set; Step (22) based on the unmanned cluster node model, the device communication and resource limitation are considered, the unmanned cluster node edge model is established, the edge attribute index is set, and the dynamic reconstruction strategy is established.

3. The method of claim 1, wherein, The step (21) specifically comprises: (211) for the detection node s mainly comprising a detection device, there are significant performance differences between different detection devices; the attribute model of the detection node s is established as: wherein s i denotes the number of nodes in the probing node s, denotes the failure rate of the probing node s i , denotes the repair rate of the probing node s i , denotes the degree of the probing node s i , denotes the x-coordinate of the probing node s i , denotes the y-coordinate of the probing node s i , k denotes that the node s i is a member of the cluster K, denotes the maximum communication distance of the node s i , denotes the maximum probing distance; (212)For the decision node D, the main performance indicators include response time, throughput and accuracy, and the detection and influence task allocation results of the decision node jointly affect the generation and effect of the whole OODA cycle and task chain; the communication ability of the decision node directly affects its ability to issue tasks to other nodes, and the attribute model of the decision node D is established as follows: where d j represents the jth decision node, J represents the number of the jth decision node, represents the d j failure rate of the node, represents the d j repair rate of the node, represents the d j degree of the node, represents the x-axis position coordinate of the jth decision node, represents the y-axis position coordinate of the jth decision node, cluster k represents that the jth decision node is a member of the cluster K, represents the maximum communication distance of the jth decision node; (213)Different target nodes T show different characteristics, and the main performance indicators include target value, detection difficulty and task completion difficulty; The attribute model of the target node is established as: where t n represents the nth target node, N represents the number of the nth target node, represents t n the failure rate of the node, represents t n the repair rate of the node, represents the x-axis position coordinate of the nth target node, represents the y-axis position coordinate of the nth target node, cluster k represents that the nth target node is a member of the cluster K, represents the target value.

4. The method of claim 1 or 3, wherein, The step (22) is specifically: edge Considering the assignment result of the detection task and the detection probability, the unmanned cluster node edge model is established as follows: wherein denotes the nth target node t n the distance to the ith probe node s i denotes the result of the probe task assignment, denotes the nth target node t n the probe probability for the ith probe node s i .​ Unmanned cluster node edge model Is established as follows: wherein denotes the distance between the ith probing node s i and the jth decision node d j , denotes the communication task success rate between the ith probing node s i and the jth decision node d j , denotes the packet loss rate between the ith probing node s i and the jth decision node d j ; Unmanned cluster node edge model Is established as follows: wherein, denotes the jth decision node d j distance to the mth influence node w m denotes the jth decision node d j communication task success rate to the mth influence node w m denotes the jth decision node d j packet loss rate to the mth influence node w m ​​​ Unmanned cluster node edge model Is established as follows: In the formula, denotes the mth influence node w m and the distance between the nth target node t n , denotes the assignment result of the influence task, denotes the mth influence node w m on the nth target node t n influence task completion probability.

5. The method of claim 1, wherein, In the step (33), based on the existence of nodes and edges a(·), the node indicator function is as follows: where s i represents the ith probe node, d j represents the jth decision node, w m represents the mth impact node; The edge existence indicator function is as follows: wherein denotes the set of edges of the network between the ith probing node and the jth decision node, denotes the set of edges of the network between the nth target node and the ith probing node, denotes the set of edges of the network between the jth decision node and the mth influencing node, denotes the set of edges of the network between the mth influencing node and the nth target node.

6. The method of claim 1, wherein, The step (33) further includes a random failure step and a deliberate attack step, specifically: The random failure step comprises: according to the node of the corresponding unmanned cluster node model and the edge failure rate and its distribution of the corresponding unmanned cluster node edge model, determining the number of node and edge failures of the unmanned cluster through Monte Carlo simulation, randomly removing the failed nodes of the unmanned cluster node model and the edges thereof, and adding the failed nodes and edges to the failed node list and the failed edge list respectively; wherein the failure or death of the probe nodes s i , the probe nodes d j , and the probe nodes w m is described by a Poisson distribution with a failure rate failure rate failure rate . The deliberate attack step includes: determining an attack mode, arranging nodes in the unmanned cluster in descending order of node degree, then removing a corresponding number of failed nodes and edges thereof, and adding the failed nodes to a failed node list; using CNA as a special case, studying node failure and dynamic reconstruction strategies; setting the node degree attack strategy to attack only the top 50% of nodes, setting The degree of node s i , d j , w m .

7. The method of claim 1 or 6, wherein, The step (34) is specifically: Step (341) derives the adjacency matrix A = {A} between different types of nodes using the effective task network model. TS A SD A DW A WT }; The step (342) considers the properties and existence of nodes and edges, and verifies the existence of the adjacent matrix elements: If there is a connection between the nodes, then Otherwise, The probability of the existence of the adjacency matrix element is: Wherein, if the failure and generation of the node obey the exponential distribution, the probability of the existence of the unmanned cluster node is: The existence probability of each edge of the unmanned cluster is expressed as: wherein represents the assignment result of the detection task, represents the effective communication between the ith detection node s i and the jth decision node d j , represents the effective communication between the mth influence node w m and the nth target node t n ; wherein: After all the adjacent matrix elements are verified, turn to step (343), otherwise turn to step (342) to continue verification; The step (343) calculates the number of effective task chains: Computing the reachability matrix A TT , A TT = A TS x A SD x A DW x A WT ; Compute the reachability matrix A TT The number of effective task chains N eol , N eol = Trace(A TT ), A TS is of dimension N x I, A SD is of dimension I x J, A DW is of dimension J x M, A WT is of dimension M x N; Task success rate calculation i = 1, determine whether the task chain is a valid task chain If N eol ≥τ i indicates that the task is completed effectively, the successful task forms an effective task chain, and the number of successful tasks Num task-success =Num task-success +1,i=i+1; If N eol < τ i then it is indicated that the requirements for effective completion of the task are not met for i = i + 1; if i = n t then the next step is turned to, otherwise the step is turned to; n t is the number of T nodes. According to the number of simulations, the success rate of the task is calculated 8. The method of claim 1, wherein, The dynamic reconstruction strategy of the step two includes: rule-based dynamic reconstruction and task-based dynamic reconstruction; The task success rate indicator is adopted to evaluate the unmanned cluster task allocation strategy by combining the actual completion process and the completion result of the task; the task success rate indicator is closely combined with the effective task chain to form the effective completion task as the task success standard; The step three uses the adjacent matrix and the arrival matrix of the unmanned cluster to calculate the number of effective task chains.

Images