Open cluster formation optimization tracking method based on communication compression and related device

Abstract

The invention discloses an open cluster formation optimization tracking method based on communication compression and a related device, and relates to the technical field of formation optimization tracking of an open cluster system.The method comprises the steps that formation optimization tracking operation is executed for each participating agent, and control input of each participating agent at the current moment is obtained; obtaining the state of each participating agent at the next moment according to the control input of each participating agent at the current moment; wherein the formation optimization tracking operation comprises the following steps: calculating a weighted average value of the target agent at the current moment according to communication compression information and weighting factors of all neighbor agents corresponding to the target agent at the current moment; and calculating the control input of the target agent at the current moment according to the state of the target agent at the current moment and the weighted average value. According to the invention, time-varying formation optimization tracking of the open cluster system can be realized.

Description

Technical Field

[0001] The present application relates to the technical field of formation optimization and tracking of an open cluster system, and in particular to an open cluster formation optimization and tracking method based on communication compression and related devices. Background Art

[0002] With the widespread application of cluster systems in the execution of complex tasks, multi-agent collaborative formation control has gradually become a research hotspot. The core goal of formation control is to achieve coordinated movement among agents, so that they can form and maintain a specific desired formation configuration and complete the path tracking task at the same time. Traditional formation control methods are mostly based on the construction of a distributed framework based on consistency theory, and global coordination is achieved through local information interaction. However, the formation tracking problem in a dynamic environment puts forward higher requirements: on the one hand, the formation trajectory needs to be collaboratively optimized by comprehensively considering the performance constraints of each agent; on the other hand, in the absence of a fixed leader, the agent needs to use neighbor information to autonomously calculate the trajectory in real time. In recent years, for the problem of time-varying formation optimization tracking, research has gradually introduced optimization theories such as subgradient methods, aiming to solve the global optimal trajectory through distributed strategies. Despite this, existing methods still have key limitations, especially when dealing with open cluster systems with dynamically changing numbers of agents, their applicability and efficiency face significant challenges.

[0003] Existing formation optimization and tracking methods usually assume that the cluster size is fixed, but in practical applications (such as dynamic increase and decrease of unmanned combat systems and real-time adjustment of logistics fleets), agents may join or exit at any time due to task requirements or environmental interference. To solve the coordination problem of such open cluster systems, studies have proposed dynamic adjustment strategies based on consensus algorithms, such as proportional protocols and pulse observer protocols, to support the dynamic changes in the number of agents. However, formation optimization and tracking not only need to solve the consistency coordination problem, but also need to meet the expected formation configuration and state constraints under the time-varying optimization objectives, which puts higher requirements on the robustness and computational efficiency of distributed algorithms. Although some progress has been made in the current research on time-varying constraint optimization of open cluster systems, such as introducing projection operators to process set constraints and using dual average methods to process equality constraints, these methods all rely on accurate neighbor information interaction, resulting in excessive consumption of communication resources and difficulty in adapting to real-time requirements or strong interference environments.

[0004] Although communication compression technology has been applied in the consistency and static optimization problems of fixed-scale clusters, its expansion faces severe challenges in time-varying optimization scenarios, especially in open cluster systems. The main reason is that the dynamically changing cluster scale will introduce time-varying communication errors, and the existing compression methods lack effective modeling and control mechanisms for the relationship between scale changes and error accumulation. Therefore, how to achieve time-varying formation optimization tracking in open cluster systems under the condition of limited communication resources has become a technical problem that needs to be overcome urgently. Summary of the Invention

[0005] The objective of this application is to provide an open cluster formation optimization tracking method and related devices based on communication compression, which can achieve time-varying formation optimization tracking of an open cluster system.

[0006] To achieve the above objective, this application provides the following solutions:

[0007] In a first aspect, this application provides an open cluster formation optimization tracking method based on communication compression, including:

[0008] For each participating agent, perform formation optimization tracking operations to obtain the control input of each participating agent at the current moment;

[0009] Based on the control input of each participating agent at the current moment, obtain the state of each participating agent at the next moment;

[0010] Among them, the formation optimization tracking operation specifically includes:

[0011] Obtain the state of the target agent at the current moment and the communication compression information of all neighboring agents corresponding to the target agent; the target agent is any participating agent; the neighboring agent is a participating agent that has a communication relationship with the target agent;

[0012] Determine the weighting factors of all neighboring agents corresponding to the target agent at the current moment;

[0013] Based on the communication compression information and weighting factors of all neighboring agents corresponding to the target agent at the current moment, calculate the weighted average value of the target agent at the current moment;

[0014] Based on the state and weighted average value of the target agent at the current moment, calculate the control input of the target agent at the current moment.

[0015] In a second aspect, this application provides a computer device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor executes the computer program to implement the open cluster formation optimization tracking method based on communication compression described above.

[0016] In a third aspect, this application provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the open cluster formation optimization tracking method based on communication compression described above.

[0017] In a fourth aspect, this application provides a computer program product, including a computer program, and when the computer program is executed by a processor, it implements the open cluster formation optimization tracking method based on communication compression described above.

[0018] According to the specific embodiments provided by the present application, the following technical effects are disclosed in the present application:

[0019] The present application provides an open cluster formation optimization tracking method based on communication compression and related devices. The method includes performing formation optimization tracking operations for each participating agent to obtain the control input of each participating agent at the current moment; obtaining the state of each participating agent at the next moment according to the control input of each participating agent at the current moment. Among them, the formation optimization tracking operation specifically includes: calculating the weighted average value of the target agent at the current moment according to the communication compression information and weighting factors of all neighboring agents corresponding to the target agent at the current moment; calculating the control input of the target agent at the current moment according to the state and weighted average value of the target agent at the current moment.

[0020] The present application utilizes the communication compression information of the neighboring agents corresponding to the participating agents, and performs information fusion through the weighted average value, enabling each participating agent to fully understand the global information without relying on the communication information of all agents, effectively reducing the complexity and redundancy of information transmission among all participating agents; moreover, since the target tracking operation and optimization process of each participating agent are based on the communication compression information of the neighboring agents, the system can achieve stable and efficient cluster control without relying on global information, and can adapt to a large-scale dynamically changing agent group, thereby realizing the time-varying formation optimization tracking of the open cluster system. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0022] Figure 1 It is an application environment diagram of an open cluster formation optimization tracking method based on communication compression in an embodiment of the present application;

[0023] Figure 2 It is a schematic flowchart of an open cluster formation optimization tracking method based on communication compression provided in an embodiment of the present application;

[0024] Figure 3 For Figure 2 It is a detailed flowchart of step 201 in

[0025] Figure 4 It is a schematic diagram of the change curve of the number of participating agents provided in another embodiment of the present application;

[0026] Figure 5a Schematic diagram of the communication topology of the participating agents at the initial moment provided by another embodiment of the present application;

[0027] Figure 5b Schematic diagram of the communication topology of the participating agents at the arrival moment provided by another embodiment of the present application;

[0028] Figure 5c Schematic diagram of the communication topology of the participating agents at the departure moment provided by another embodiment of the present application;

[0029] Figure 6a Schematic diagram of the state of the participating agents at the current moment k = 0 provided by another embodiment of the present application;

[0030] Figure 6b Schematic diagram of the state of the participating agents at the current moment k = 114 provided by another embodiment of the present application;

[0031] Figure 6c Schematic diagram of the state of the participating agents at the current moment k = 180 provided by another embodiment of the present application;

[0032] Figure 6d Schematic diagram of the state of the participating agents at the current moment k = 300 provided by another embodiment of the present application;

[0033] Figure 6e Schematic diagram of the state of the participating agents at the current moment k = 450 provided by another embodiment of the present application;

[0034] Figure 6f Schematic diagram of the state of the participating agents at the current moment k = 495 provided by another embodiment of the present application;

[0035] Figure 7 Schematic diagram of the dynamic regret value curve for time average provided by another embodiment of the present application;

[0036] Figure 8 Schematic diagram of the formation constraint violation regret value curve for time average provided by another embodiment of the present application;

[0037] Figure 9 Schematic diagram of the structure of a computer device provided by an embodiment of the present application. Detailed implementation manner

[0038] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0039] To make the above objects, features, and advantages of the present application more obvious and understandable, the present application will be further described in detail below in conjunction with the drawings and specific embodiments.

[0040] The open cluster formation optimization tracking method based on communication compression provided by the embodiments of the present application can be applied to, for example, Figure 1 the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be set separately, integrated on the server 104, placed in the cloud or on other servers. The terminal 102 can send the state of the target agent at the current moment and the communication compression information of all neighbor agents corresponding to the target agent to the server 104. The server 104 calculates the weighted average value of the target agent at the current moment according to the communication compression information of all neighbor agents corresponding to the target agent at the current moment and the weighting factor; according to the state of the target agent at the current moment and the weighted average value, the control input of the target agent at the current moment is calculated. The server 104 can feedback the obtained control input of the target agent at the current moment to the terminal 102. Among them, the server 104 can be implemented by an independent server or a server cluster composed of multiple servers, and can also be a cloud server.

[0041] In an exemplary embodiment, as Figure 2 shown, an open cluster formation optimization tracking method based on communication compression is provided. This method is executed by a computer device and can be specifically executed by a terminal and a server together. In the embodiments of the present application, taking this method applied to Figure 1 the server 104 in as an example for illustration, it includes the following steps 201 to step 202. Among them:

[0042] Step 201, for each participating agent, perform a formation optimization tracking operation to obtain the control input of each participating agent at the current moment.

[0043] Among them, the formation optimization tracking operation specifically includes:

[0044] Step 2011: Obtain the state of the target agent at the current moment and the communication compression information of all neighbor agents corresponding to the target agent; the target agent is any participating agent; the neighbor agent is a participating agent that has a communication relationship with the target agent.

[0045] Step 2012: Determine the weighting factors of all neighbor agents corresponding to the target agent at the current moment.

[0046] Step 2013: Calculate the weighted average value of the target agent at the current moment according to the communication compression information and the weighting factors of all neighbor agents corresponding to the target agent at the current moment.

[0047] Step 2014: Calculate the control input of the target agent at the current moment according to the state and the weighted average value of the target agent at the current moment.

[0048] Step 202: Obtain the state of each participating agent at the next moment according to the control input of each participating agent at the current moment.

[0049] By implementing the above Steps 201 to 202, the present application utilizes the communication compression information of neighbor agents and performs information fusion through the weighted average value, enabling each participating agent to fully understand the global information without relying on the communication information of all agents, effectively reducing the complexity and redundancy of information transmission among all participating agents; moreover, since the target tracking operation and optimization process of each participating agent are based on the communication compression information of neighbor agents, the system can achieve stable and efficient cluster control without relying on global information, can adapt to a large-scale dynamically changing agent group, and thus realizes the time-varying formation optimization tracking of an open cluster system.

[0050] Before performing the formation optimization tracking operation for each participating agent, it is necessary to perform a systematic mathematical modeling on the open cluster system and the formation optimization tracking problem. Specifically: First, describe the communication topology structure of the open cluster system and define the openness characteristics (including the dynamic joining, leaving, and retention mechanisms of agents); Second, establish the dynamic models of each participating agent in the open cluster system to characterize its state evolution law; Finally, based on the above modeling content, clearly define the formation optimization tracking problem to be solved in the present application, thereby providing a mathematical basis and constraint framework for the design of subsequent distributed formation optimization tracking methods.

[0051] Here, first give the definitions of all symbols in the present application: Denote the set of m-dimensional real-valued vectors; Denote the set of real-valued scalars; Denotes the empty set; for a set, |·| represents the total number of elements in the set; for a vector or matrix, ||·||1 represents the L1 norm and ||·||2 represents the L2 norm; A\B represents the difference set of sets A and B; A∩B represents the intersection of sets A and B; A∪B represents the union of sets A and B; for a vector [x] j represents the j-th element, sgn(x) = [sign([x]1),..., sign([x] n )] T is the sign function with respect to the vector, where sign(·) is the standard sign function, (·) T or [·] T denotes the transpose; Π X [·] represents the projection onto X; for a vector x and a function f, ▽f(x) represents the subgradient of f with respect to x; if means if, otherwise means otherwise; min means minimize; s.t. means subject to the constraint.

[0052] Step (1) Modeling the communication network of the open cluster system.

[0053] The communication network of the open cluster system at any moment is represented by an undirected communication topology . Among them, represents the time domain set; T represents the number of time domain steps and is a positive integer; and ε k represent the set of agents (also called participating agents hereinafter) at the current moment k and the set of communication edges respectively. The edge set of the target agent is defined as ε i,k = {e ∈ ε k | (i, j) ∈ ε k}, where, (i, j) ∈ ε k means that the neighbor agent j can transmit information to the target agent i. The neighbor set of the target agent is defined as that is, all neighbor agents that can transmit information to agent i. The target agent is any participating agent.

[0054] Further define the set of participating agents. For the current moment the set of participating agents can be divided into two subsets: the arriving agent set represents that it did not participate at the previous moment k - 1 but participates at the current moment k; the retained agent set represents that it participates at both the current moment k and the previous moment k - 1. In addition, the departing agent set is defined as Indicates that it participated at the previous moment k - 1 but did not participate at the current moment k. At the current moment k = 0, the arriving agent set is initialized to The departing agent set is initialized to

[0055] Step (2) Dynamics modeling of the open cluster system.

[0056] At the current moment The target agent The dynamics model is:

[0057]

[0058] Where is the state of the target agent i at the next moment k + 1; is the control input of the target agent i at the current moment k.

[0059] Step (3) Formation optimization tracking problem modeling.

[0060] At the current moment The target agent can obtain the following local information (i.e., the information at the current moment) from the local data storage system: the state set constraint The objective function of the target agent i at the current moment k and the desired formation configuration of the target agent i at the current moment k Both the objective function and the desired formation configuration are time-varying. Each participating agent aims to design the protocol form of the control input based on the above-given local information and the communication compression information of the neighboring agents obtained through the communication network of the open cluster system, so as to solve the following time-varying formation optimization tracking problem:

[0061]

[0062] Where is the state of the target agent i at the current moment k; h i,k is the desired formation configuration of the target agent i at the current moment k; x j,k is the state of the neighboring agent j at the current moment k; h j,k is the desired formation configuration of the neighboring agent j at the current moment k.

[0063] Definition (formation optimization tracking): Considering the open cluster system described in step (1), for bounded initial states and initial control inputs, if the dynamic regret value Reg d and the formation constraint violation regret value Reg fIf it has a linear upper bound, it indicates that the open cluster system has achieved formation optimization tracking; the dynamic regret value Reg d and the formation constraint violation regret value Reg f are as follows:

[0064]

[0065] wherein, is the optimal state of the target agent i at the current time k.

[0066] In summary, the objective of this application is to design a distributed formation optimization tracking method such that the dynamic regret value Reg d and the formation constraint violation regret value Reg f have a linear upper bound, under the condition that each participating agent in the open cluster system can obtain the state set constraint information, the local information at the current time, and the communication compression information of the neighbor agents at the current time through the communication network.

[0067] Next, a specific design of the distributed formation optimization tracking operation is carried out.

[0068] Furthermore, in step 2011, the participating agents include arriving agents and remaining agents. Then, obtaining the state of the target agent at the current time specifically includes:

[0069] a. When the target agent is an arriving agent, an initialization operation is performed on the target agent to obtain the initial state of the target agent. The specific process is as follows:

[0070] According to the formula the initial state of the target agent is obtained and the initial state is used as the state of the target agent at the current time.

[0071] wherein, if the current time k = 0, the target agent i is initialized using the state set constraint to obtain the initial state x i,0 of the target agent i at the current time k = 0. The state set constraint is used to keep the state of the target agent i in the safe interval; if the current time k > 0, the target agent i is initialized using the desired formation configuration h i,k of the target agent i at the current time k, the state x j,k of the neighbor agent j at the current time k, and the desired formation configuration h j,k of the neighbor agent j at the current time k to obtain the initial state x i,k of the target agent i at the current time k; is the set of neighbor agents corresponding to the target agent i at the current time k; is the set of participating agents at the previous moment k-1; is the set of participating agents at the current moment k.

[0072] b. When the target agent is a retained agent, directly obtain the state of the target agent at the current moment from the local data storage system.

[0073] Furthermore, in step 2012, to determine the weighting factors of all neighbor agents corresponding to the target agent at the current moment, the specific process is as follows:

[0074] According to the formula p ij,k >0, determine the weighting factors of all neighbor agents corresponding to the target agent at the current moment.

[0075] where p ij,k is the weighting factor of the target agent i for the neighbor agent j at the current moment k, and the sum of all weighting factors of the target agent i is 1; is the set of neighbor agents corresponding to the target agent i at the current moment k.

[0076] Furthermore, in step 2013, according to the communication compression information and weighting factors of all neighbor agents corresponding to the target agent at the current moment, calculate the weighted average value of the target agent at the current moment. The specific process is as follows:

[0077] According to the formula calculate the weighted average value of the target agent at the current moment;

[0078] where y i,k is the weighted average value of the target agent i at the current moment k; j is the neighbor agent; is the set of neighbor agents corresponding to the target agent i at the current moment k; p ij,k is the weighting factor of the target agent i for the neighbor agent j at the current moment k; is the communication compression information of the neighbor agent j corresponding to the target agent i at the current moment k, sgn(x) = [sign([x]1),..., sign([x] n )] T is the sign function for vectors, sign(·) is the standard sign function, [x] n represents the nth element; x i,k is the state of the target agent i at the current moment k; h i,k is the expected formation configuration of the target agent i at the current moment k; x j,k is the state of the neighbor agent j at the current moment k; h j,k is the expected formation configuration of the neighbor agent j at the current moment k.

[0079] Further, in step 2014, the control input of the target agent at the current moment is calculated according to the state and weighted average of the target agent at the current moment, specifically including:

[0080] According to the formula Determine the weighted average step size of the target agent at the current moment.

[0081] Where L is the Lipschitz constant of the objective function f of the target agent i at the current moment k, L > 0; p i,k is the minimum value among all the weighting factors of the target agent i at the current moment k; p is the minimum value among all the weighting factors of all the participating agents; i,k is the maximum value of the number of neighbor agent sets of all the participating agents; δ is the weighted average step size of the target agent i at the current moment k. i,k is the weighted average step size of the target agent i at the current moment k.

[0082] According to the formula Determine the learning step size of the target agent at the current moment.

[0083] Where α k is the learning step size of the target agent i at the current moment k; T is the number of time domain steps; n α is the first proportionality coefficient, n α > 0; k0 is a constant greater than or equal to 0; 0 < γ α < min{γ φ , γ a}; γ φ is the upper bound of the optimal state change of the target agent i at the current moment k satisfying γ a is the upper bound of the number of arriving agents at the current moment k, which satisfies n φ is the second proportionality coefficient, n φ > 0; 0 < γ a , γ φ < 1;

[0084] According to the formula Calculate the control input of the target agent at the current moment.

[0085] Where u i,k is the control input of the target agent i at the current moment k; is the state set constraint; represents the projection on ; x i,k is the state of the target agent i at the current moment k; y i,k is the weighted average of the target agent i at the current moment k; ▽fi,k is the subgradient of the objective function, which is used to find the optimization direction of the objective function of the target agent i at the current time k.

[0086] In another exemplary embodiment of the present application, a simulation routine in a two-dimensional plane is considered, that is, the state of the target agent i at the current time k where

[0087] The specific settings of the formation optimization tracking problem in formula (2) are as follows: f i,k (ξ) = a i,k ξ T ξ + [b i,k , c i,k ξ + d i,k . Among them, a i,k , b i,k , c i,k , d i,k are the objective function coefficients of the target agent i at the current time k. At the arrival time of the target agent i they are randomly sampled from a uniform distribution with intervals of [0.1, 0.2], [-0.5, 0.5], [-0.5, 0.5], [-2, 2] respectively; subsequently, before the target agent i leaves, there is The state set constraint is set to a circle centered at the origin with a radius of 3. The expected formation configuration h of the target agent i at the current time k i,k is set to a circle with a radius of 1, and the expected formation configuration is a circle that rotates continuously over time. At the current time k = 0, the state components of each dimension of the arriving agent are randomly sampled from a uniform distribution with an interval of [-2, 2].

[0088] The change curve of the number of participating agents is as Figure 4 shown, where the rhombus indicates that an agent arrives or leaves at the current time, and the circle indicates that an agent arrives and leaves at the current time. Within the time domain steps T = 500, the number of participating agents varies between 5 and 10.

[0089] The communication topology diagrams of the participating agents at the initial time (k = 0), arrival time (k = 62), and departure time (k = 271) are respectively as Figure 5a , Figure 5b and Figure 5c shown, where the circle represents the participating agent node, the gray color represents that the node arrives / leaves at the current time, and the solid line represents the communication edge between two participating agents.

[0090] The weighting factor, weighted average step size, and learning step size of the target agent i for the neighbor agent j at the current time k are set as follows: For any neighbor agent There is δ i,k = 10, α i,k = T -0.6 .

[0091] Based on the above simulation settings, the operation results of the distributed formation optimization tracking method of this application are given. The following Figure 6 respectively gives the state screenshots at the current times k of 0, 114, 180, 350, 400, and 495. The circles represent the remaining agents, the diamonds represent the arriving agents, the rectangles represent the leaving agents, the solid lines represent the expected formation configurations formed by the agents, and the dashed lines represent the expected formation configurations formed by the leaving agents and the remaining agents before leaving.

[0092] As can be seen from the above figure, the distributed formation optimization tracking method proposed in this application allows the arbitrary joining and leaving of agents and can execute tasks with a given expected formation configuration.

[0093] Figure 7 And Figure 8 respectively show the curves of the time-averaged dynamic regret value and the time-averaged formation constraint violation regret value of the participating agents. Among them, the diamond markers represent the arrival times of the agents, the rectangle markers represent the departure times of the agents, the solid lines represent the overall regret value curves with respect to the time average, that is And The dashed lines represent the regret value curves of the target agent i with respect to the time average, that is And The specific expressions are as follows:

[0094]

[0095] When k = T, Therefore, it can be observed from And The changing trends of Reg d And The changing trends; And Are And The components of each participating agent in, so they can be used to observe the influence of each participating agent on Reg d And Reg f The influence of.

[0096] From Figure 7 And Figure 8It can be seen that the newly arrived agent has a greater impact on the formation constraint violation regret value at the arrival moment. However, the time-averaged dynamic regret value and the formation constraint violation regret value will generally converge to a small neighborhood of 0, that is, both the dynamic regret value and the formation constraint violation regret value increase linearly. Therefore, under the distributed formation optimization tracking method proposed in this application, the open cluster system can achieve formation optimization tracking.

[0097] This application studies the time-varying formation optimization tracking problem of an open cluster system. Among them, agents are allowed to join and leave at any time. Each participating agent aims to obtain the optimal trajectory by minimizing the cumulative local time-varying objective function of all participating agents while satisfying the time-varying formation configuration constraints. In addition, this application provides a method for determining parameters (including the weighting factor, weighted average step size, and learning step size), which can make the upper bound of the cumulative error increase linearly. That is, when the time domain is large enough, the actual trajectory of the open cluster system can approximate a sufficiently small neighborhood of the optimal state, so as to achieve time-varying formation optimization tracking.

[0098] This application also provides an application scenario that applies the above-mentioned open cluster formation optimization tracking method based on communication compression. Specifically: The open cluster formation optimization tracking method based on communication compression provided in this embodiment can be applied to open cluster formation optimization tracking. Open cluster formation optimization tracking includes a formation optimization link and a state update link; the formation optimization link is used to perform formation optimization tracking operations for each participating agent to obtain the control input of each participating agent at the current moment; the state update link is used to obtain the state of each participating agent at the next moment according to the control input of each participating agent at the current moment. The open cluster formation optimization tracking method based on communication compression provided in this embodiment belongs to the formation optimization link and the state update link.

[0099] In an exemplary embodiment, a computer device is provided. The computer device can be a server or a terminal, and its internal structure diagram can be as Figure 9As shown in the figure. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O), and a communication interface. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store processed data. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals through a network connection. When the computer program is executed by the processor, it implements an open cluster formation optimization tracking method based on communication compression.

[0100] Those skilled in the art can understand that Figure 9 the structure shown in the figure is only a block diagram of some structures related to the solution of this application, and does not constitute a limitation on the computer device to which the solution of this application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0101] In an exemplary embodiment, a computer device is provided, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the steps in the above method embodiments are implemented.

[0102] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0103] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0104] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use, and processing of relevant data need to comply with relevant regulations.

[0105] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the various embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memories can include read-only memory (ROM), magnetic tapes, floppy disks, flash memories, optical memories, high-density embedded non-volatile memories, resistive random-access memories (ReRAMs), magnetoresistive random-access memories (MRAMs), ferroelectric random-access memories (FRAMs), phase change memories (PCMs), graphene memories, etc. Volatile memories can include random access memories (RAMs) or external caches, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0106] The databases involved in the various embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the various embodiments provided in the present application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logics, data processing logics based on quantum computing, etc., without limitation.

[0107] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the various technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this specification.

[0108] Specific examples are used in this article to elaborate on the principles and implementation manners of the present application. The description of the above embodiments is only used to help understand the method and its core idea of the present application; at the same time, for those of ordinary skill in the art, according to the idea of the present application, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to the present application.

Claims

1. An open cluster formation optimization tracking method based on communication compression, characterized in that Including: For each participating agent, perform formation optimization tracking operation to obtain the control input of each participating agent at the current moment; According to the control input of each participating agent at the current moment, obtain the state of each participating agent at the next moment; Among them, the formation optimization tracking operation specifically includes: Obtain the state of the target agent at the current moment and the communication compression information of all neighbor agents corresponding to the target agent; the target agent is any participating agent; the neighbor agent is a participating agent having a communication relationship with the target agent; Determine the weighting factors of all neighbor agents corresponding to the target agent at the current moment; According to the communication compression information and weighting factors of all neighbor agents corresponding to the target agent at the current moment, calculate the weighted average value of the target agent at the current moment; According to the state and weighted average value of the target agent at the current moment, calculate the control input of the target agent at the current moment.

2. The open cluster formation optimization tracking method based on communication compression according to claim 1, wherein The participating agents include arriving agents and retaining agents; Obtaining the state of the target agent at the current moment specifically includes: When the target agent is an arriving agent, perform an initialization operation on the target agent to obtain the initial state of the target agent, and use the initial state as the state of the target agent at the current moment; When the target agent is a retaining agent, directly obtain the state of the target agent at the current moment.

3. The open cluster formation optimization tracking method based on communication compression according to claim 2, characterized in that When the target agent is an arriving agent, the process of performing an initialization operation on the target agent to obtain the initial state of the target agent is as follows: According to the formula obtain the initial state of the target agent; Among them, if the current time k = 0, the state set constraint is used to perform an initialization operation on the target agent i to obtain the initial state x of the target agent i at the current time k = 0 i,0 , and the state set constraint is used to keep the state of the target agent i within the safe range; if the current time k > 0, the desired formation configuration h of the target agent i at the current time k i,k , the state x of the neighbor agent j at the current time k j,k and the desired formation configuration h of the neighbor agent j at the current time k j,k are used to perform an initialization operation on the target agent i to obtain the initial state x of the target agent i at the current time k i,k ; is the set of neighbor agents corresponding to the target agent i at the current time k; V k-1 is the set of participating agents at the previous time k - 1; V k is the set of participating agents at the current time k.

4. The open cluster formation optimization tracking method based on communication compression according to claim 1, characterized in that The process of determining the weighting factors of all neighbor agents corresponding to the target agent at the current moment is as follows: According to the formula Determine the weighting factors of all neighbor agents corresponding to the target agent at the current moment; Among them, p ij,k is the weighting factor of the target agent i for the neighbor agent j at the current time k; is the set of neighbor agents corresponding to the target agent i at the current time k.

5. The open cluster formation optimization tracking method based on communication compression according to claim 1, characterized in that The process of calculating the weighted average value of the target agent at the current moment according to the communication compression information and weighting factors of all neighbor agents corresponding to the target agent at the current moment is as follows: According to the formula calculate the weighted average of the target agent at the current moment; Among them, y i,k is the weighted average of the target agent i at the current time k; j is the neighbor agent; is the set of neighbor agents corresponding to the target agent i at the current time k; p ij,k is the weighting factor of the target agent i for the neighbor agent j at the current time k; is the communication compression information of the neighbor agent j corresponding to the target agent i at the current time k; x i,k is the state of the target agent i at the current time k; h i,k is the expected formation configuration of the target agent i at the current time k; x j,k is the state of the neighbor agent j at the current time k; h j,k is the expected formation configuration of the neighbor agent j at the current time k.

6. The open cluster formation optimization tracking method based on communication compression according to claim 1, characterized in that Calculating the control input of the target agent at the current moment according to the state and weighted average value of the target agent at the current moment specifically includes: According to the formula to determine the weighted average step size of the target agent at the current moment; According to the formula to determine the learning step size of the target agent at the current moment; According to the formula calculate the control input of the target agent at the current moment; Among them, L is the Lipschitz constant of the objective function f of the target agent i at the current time k i,k ; p i,k is the minimum value among all the weighting factors of the target agent i at the current time k; p is the minimum value among all the weighting factors of all the participating agents; is the maximum value of the number of neighbor agent sets of all the participating agents; δ i,k is the weighted average step size of the target agent i at the current time k; α k is the learning step size of the target agent i at the current time k; T is the number of time domain steps; n α is the proportionality coefficient; k0 is a constant greater than or equal to 0; γ α is the upper bound of the number of arriving agents at any time; u i,k is the control input of the target agent i at the current time k; X is the state set constraint; Π X [·] represents the projection onto X; x i,k is the state of the target agent i at the current time k; y i,k is the weighted average value of the target agent i at the current time k; ▽f i,k is the subgradient of the objective function, which is used to find the optimization direction of the objective function of the target agent i at the current time k.

7. The open cluster formation optimization tracking method based on communication compression according to claim 1, characterized in that The process of obtaining the state of each participating agent at the next moment according to the control input of each participating agent at the current moment is as follows: According to the formula x i,k+1 = u i,k Calculate the state of each participating agent at the next moment; Among them, u i,k is the control input of the target agent i at the current time k; x i,k is the state of the target agent i at the next time k + 1.

8. A computer device, comprising: A memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor executes the computer program to implement the communication compression-based open cluster formation optimization tracking method according to any one of claims 1-7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the communication compression-based open cluster formation optimization tracking method according to any one of claims 1-7.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the communication compression-based open cluster formation optimization tracking method according to any one of claims 1-7.

Images