Unmanned cluster action strategy recommendation method based on case reasoning
By calculating unmanned swarm action strategies using a case-based reasoning method, and utilizing the similarity and weight values between target and source cases, the problem of insufficient effectiveness in generating unmanned swarm action strategies is solved, enabling rapid and accurate recommendation of unmanned swarm action strategies.
Patent Information
- Application Number
- CN202410827578.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-06-25
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2044-06-25
AI Technical Summary
Existing technologies are not effective enough in generating unmanned swarm action strategies, especially in terms of autonomy, adversarial nature, and uncertainty.
A case-based reasoning approach is adopted. By acquiring the target case and multiple source cases of the unmanned swarm, the similarity and weight values are calculated, and the source case with the highest global similarity is selected to guide the action strategy. Case-based reasoning technology is used to recommend action strategies for the unmanned swarm.
It enables rapid and reasonable recommendation of unmanned swarm action strategies, improves the accuracy and effectiveness of action strategies, and can guide the actions of unmanned swarms in both online and offline states.
Smart Images

Figure CN118917410A_ABST
Abstract
Description
Technical Field
[0001] The embodiments of the present application relate to the technical field of unmanned swarm action, and in particular to a method for recommending an unmanned swarm action strategy based on case reasoning. Background Art
[0002] Unmanned swarm is a disruptive technology that changes the rules of confrontation. Through the close collaboration of individual unmanned equipment (such as drones), it can achieve a leap in the overall capabilities of the swarm, and has outstanding advantages that are difficult to match by a single unmanned equipment.
[0003] After receiving a task, unmanned swarms usually need to determine their action strategy first and formulate an action plan based on the strategy. At present, the research on action strategies for unmanned swarms to perform tasks has become a research hotspot, including interval decision-making with adversarial game models, multi-agent reinforcement learning decision-making, and interference avoidance priority adversarial strategies. However, due to the characteristics of autonomy, adversarialness, and uncertainty in the actions of unmanned swarms, the above research schemes still have some defects in the effectiveness of generating action strategies for unmanned swarms. Summary of the invention
[0004] The first aspect of the embodiment of the present application proposes an unmanned swarm action strategy recommendation method based on case reasoning, and the unmanned swarm action strategy recommendation method based on case reasoning includes:
[0005] Obtaining a target case of an unmanned cluster and a plurality of corresponding source cases, and obtaining a plurality of problem attributes of the case of the unmanned cluster, wherein the problem attributes are used to characterize characteristic attributes of the problem of the case;
[0006] Calculating the similarity between the target case and the multiple source cases in terms of the target problem attribute, wherein the target problem attribute is any one of the multiple problem attributes;
[0007] Calculating normalized values of the multiple source cases on the target problem attribute, calculating the information entropy of the target problem attribute according to the normalized values, and calculating the weight value of the target problem attribute according to the information entropy;
[0008] Calculating the global similarity between the target case and the plurality of source cases according to the similarity and the weight value;
[0009] A target source case is selected from the plurality of source cases according to the global similarity, so as to guide the action strategy of the target case through the target source case.
[0010] In some implementations, calculating the normalized values of the plurality of source cases on the target problem attribute includes:
[0011]
[0012] in, For source case S i The normalized value on the target problem attribute k, For source case S i The value of the target problem attribute k, min i,k is the minimum value of the multiple source cases on the target problem attribute k, max i,k is the maximum value of the multiple source cases on the target problem attribute k, i≤I,l∈N + , I is the total number of the multiple source cases, k≤K,k∈N + , K is the total number of the multiple problem attributes, N + is the set of all positive integers;
[0013] The calculating the information entropy of the target problem attribute according to the normalized value includes:
[0014]
[0015] Among them, E k is the information entropy of the target problem attribute k;
[0016] The calculating the weight value of the target problem attribute according to the information entropy includes:
[0017]
[0018] Among them, ω k is the weight value of the target problem attribute k.
[0019] In some implementations, calculating the global similarity between the target case and the plurality of source cases according to the similarity and the weight value includes:
[0020]
[0021] Among them, Sim(S i ,T j ) is the target case T j With source case S i The global similarity between i,k ,T j,k ) is the target case T j Source Case S i The similarity between the target problem attributes k, S i,k For characterizing source case S i The target problem attributes k, T j,k To characterize the target case T j The target problem attribute k.
[0022] In some implementations, when the target question attribute is a deterministic question attribute, calculating the similarity between the target case and the plurality of source cases in the target question attribute includes:
[0023]
[0024] Among them, Sim(S i,k ,T j,k ) is the target case T j Source Case S i The similarity between the attributes of the deterministic problem, S i,k For characterizing source case S i The deterministic problem attribute k, T j,k To characterize the target case T j The deterministic problem attribute k, For source case S i The value of attribute k in the deterministic problem, For the target case T j The value of attribute k in the deterministic problem, max k is the maximum value of the problem attribute k, min k is the minimum value of the problem attribute k.
[0025] In some implementations, when the target question attribute is a fuzzy question attribute, calculating the similarity between the target case and the plurality of source cases in the target question attribute includes:
[0026] Sim(S i,k ,T j,k )=1-d
[0027] Among them, Sim(S i,k ,T j,k ) is the target case T j Source Case S i The similarity between the attributes of fuzzy questions, S i,k For characterizing source case S i The fuzzy problem attribute k, T j,k To characterize the target case T j The fuzzy problem attributes k and d are based on the calculation of the bulldozer distance S i,k With T j,k The similarity of attributes between .
[0028] In some implementations, when the target question attribute is an interval-type question attribute, calculating the similarity between the target case and the plurality of source cases in the target question attribute includes:
[0029]
[0030] Among them, Sim(S i,k ,T j,k ) is the target case T j Source Case S i The similarity between the interval problem attributes, S i,k For characterizing source case S i The interval problem attribute k, T j,k To characterize the target case T j The interval problem attribute k, For source case S i The value of attribute k in interval type problem, For the target case T j The value of attribute k in interval type problem, For measurement Length, For measurement Length, For measurement Length.
[0031] In some implementations, when the target question attribute is a categorical question attribute, calculating the similarity between the target case and the plurality of source cases in the target question attribute includes:
[0032]
[0033] Among them, Sim(S i,k ,T j,k ) is the target case T j With source case S i The similarity between the attributes k of the classification problem, S i,k For characterizing source case S i The classification problem attributes k, T j,k To characterize the target case T j The problem attribute k, For source case S i The value of attribute k in the classification problem, For the target case T j The value of attribute k in the classification problem, B p and B q Any two subspaces of the attribute value space of the classification problem.
[0034] A second aspect of an embodiment of the present application proposes an unmanned swarm action strategy recommendation device based on case reasoning, and the unmanned swarm action strategy recommendation device based on case reasoning includes:
[0035] A data acquisition unit, used to acquire a target case of an unmanned cluster and a plurality of corresponding source cases, and to acquire a plurality of problem attributes of the case of the unmanned cluster, wherein the problem attributes are used to characterize characteristic attributes of the problem of the case;
[0036] a similarity calculation unit, configured to calculate the similarity between the target case and the plurality of source cases in terms of the target problem attribute, wherein the target problem attribute is any one of the plurality of problem attributes;
[0037] A weight value calculation unit, used to calculate the normalized values of the multiple source cases on the target problem attribute, calculate the information entropy of the target problem attribute according to the normalized value, and calculate the weight value of the target problem attribute according to the information entropy;
[0038] A global similarity calculation unit, configured to calculate the global similarity between the target case and the plurality of source cases according to the similarity and the weight value;
[0039] A source case recommendation unit is used to select a target source case from the multiple source cases according to the global similarity, so as to guide the action strategy of the target case through the target source case.
[0040] A third aspect of an embodiment of the present application proposes an electronic device, characterized in that it includes at least one controller and a memory for communicating with the controller; the memory stores instructions that can be executed by the at least one controller, and the instructions are executed by at least one of the controllers to enable at least one of the controllers to execute the unmanned cluster action strategy recommendation method based on case reasoning as described above.
[0041] A fourth aspect of an embodiment of the present application proposes a computer-readable storage medium, characterized in that a computer program is stored on the computer-readable storage medium, and when the computer program is executed, the unmanned cluster action strategy recommendation method based on case reasoning as described above is implemented.
[0042] This application proposes a method for recommending action strategies for unmanned clusters based on case reasoning. This method applies case reasoning technology to the process of recommending action strategies for unmanned clusters, and provides a new method for recommending action strategies. By comparing the similarity with historical cases, it can quickly and reasonably find the most similar source case as a recommendation guide for the action strategy of the target case. It can recommend action strategies for unmanned cluster target cases in both online and offline states, and with the increase of historical source cases, it can continuously improve the accuracy of action strategy recommendations for target cases. This method also provides a method for calculating problem attribute similarity and global case similarity in the process of searching for target source cases, which is used to match the target source case with the highest similarity from multiple source cases, which can improve the accuracy of the searched target source cases and the effectiveness of the target case action strategy recommendation.
[0043] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS
[0044] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.
[0045] Figure 1 It is a flowchart of case reasoning provided by an embodiment of the present application;
[0046] Figure 2 It is a flowchart of a method for recommending an unmanned swarm action strategy based on case reasoning provided in an embodiment of the present application;
[0047] Figure 3 It is a structural diagram of an unmanned swarm action strategy recommendation device based on case reasoning provided in an embodiment of the present application;
[0048] Figure 4 It is an architectural diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0049] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.
[0050] The following introduces the technical concepts used in this application:
[0051] (1) Case-Based Reasoning (CBR) technology is a knowledge reasoning method that solves problems by imitating human thinking and reasoning patterns. CBR mainly uses historical experience to solve similar current problems. It has unique advantages such as weak dependence on domain knowledge, intuitive case expression, efficient problem solving, and continuous knowledge accumulation.
[0052] Reference Figure 1 At present, the most commonly used case reasoning framework is the 4R framework, which consists of four parts: case retrieval (Retrieve), case reuse (Reuse), case revision (Revise) and case storage (Retain). When the framework is running, the problem to be solved in the real world is first represented as a target case through a case. Then, the case retrieval is used to match the source case with the highest similarity from the historical case library. If the source case can be directly used to solve the target case, the case reuse is adopted to complete the case reasoning process. Finally, according to the solution and effect of the target case, the target case can be stored in the historical case library as a new case for solving subsequent real-world problems.
[0053] From the perspective of knowledge reasoning, case representation is the knowledge modeling process of real-world problems and their solutions in the form of cases, which has a weak dependence on domain knowledge. Case retrieval, case reuse, and case correction can be understood as the knowledge application process of historical case data, which is both intuitive and efficient. Case preservation is the knowledge accumulation process of case data, thereby continuously enhancing case reasoning capabilities.
[0054] (2) Action strategy: The action strategy of unmanned swarm is an important prerequisite for the formulation of action plans. The action plans under the guidance of different action strategies vary greatly, and multiple different action plans can be formulated under the guidance of the same strategy.
[0055] This application uses CBR technology to recommend unmanned cluster action strategies, and when selecting source cases for target cases, it uses problem attribute similarity and global case similarity calculation methods to improve the accuracy of source case selection.
[0056] Reference Figure 2 The present application provides a method for recommending action strategies of unmanned swarm based on case-based reasoning. The method comprises the following steps S110 to S150:
[0057] Step S110 : obtaining a target case of an unmanned cluster and a plurality of corresponding source cases, and obtaining a plurality of problem attributes of the cases of the unmanned cluster.
[0058] An unmanned swarm may be a swarm composed of drones, or a swarm composed of robots, robot dogs, or fixed unmanned devices, or a swarm composed of unmanned devices such as drones, robots, and robot dogs. There is no specific limitation here.
[0059] The target case refers to the case representation of the problem to be solved by the unmanned swarm. For CBR technology, the case usually includes three parts: problem description, solution and solution effect. When CBR technology is applied to the problem of recommending the action strategy of the unmanned swarm, the problem description is a standardized representation of the mission, force, scene and other information of the unmanned swarm action. The solution is mainly a variety of action strategies. The effect of the solution includes the achievement of the goal and the combat damage of the party. It should be noted that the action strategy can be set according to the actual operation of the unmanned swarm, and no limitation is made here.
[0060] Source cases refer to cases in the historical case library that are relevant to the target case. The historical case library is a database of historical source cases of unmanned clusters collected by some means and can be preset in advance. The multiple source cases here can be all cases in the entire historical case library, or some source cases extracted from the historical case library.
[0061] Because the target case is a problem to be solved, the source case includes: problem description, solution and solution effect. Problem attributes are the characteristic attributes of the problem used to characterize the case. The problem description part can be standardized by extracting the characteristic attributes of the problem.
[0062] It is worth noting that each source case has a corresponding action strategy, and the correspondence between the source case and the action strategy can be set in advance. Since the setting of the action strategy and the relationship with the source case are not the focus of this application, and the action strategy is common knowledge in the field, it will not be described in detail here.
[0063] In step S110, the target case and multiple source cases of the unmanned cluster are first obtained, the purpose of which is to find a target source case that is most similar to the target case from multiple source cases, and then the action strategy of the target case can be implemented according to the action strategy corresponding to the case.
[0064] Step S120: Calculate the similarity between the target case and the multiple source cases in terms of the target problem attributes.
[0065] The target problem attribute in step S120 is any one of the multiple problem attributes. Since there are multiple problem attributes, there are different calculation methods for calculating the similarity between the target case and the multiple source cases on each problem attribute. On the same problem attribute, the calculation method of the similarity between the target case and the multiple source cases is the same, and the calculation method of the similarity can be such as Euclidean distance, bulldozer distance, etc. See the introduction of the subsequent embodiments for details.
[0066] Step S130: Calculate the normalized values of the multiple source cases on the target problem attribute, calculate the information entropy of the target problem attribute according to the normalized values, and calculate the weight value of the target problem attribute according to the information entropy.
[0067] Because the calculation of the similarity between cases depends on the similarity of the problem attributes and the weight of the attributes. Therefore, in order to solve the problem of weight allocation of problem attributes, this implementation method starts from the data and allocates attribute weights according to the discreteness of the problem attribute data, thereby improving the discrimination. This embodiment takes into account the discreteness of the problem attribute data and uses the entropy weight method to calculate the weight of the problem attribute. This is because the entropy weight method uses the concept of indicator variation in information entropy to characterize the discreteness of the attribute data. Therefore, if the data variation of a problem attribute indicator is large, the role of this attribute in distinguishing different action cases is more prominent, so a larger weight should be assigned. Vice versa, the discrimination of small-weight problem attributes is poor. The calculation process includes data standardization, calculation of indicator variability, calculation of information entropy, and determination of weights.
[0068] In some implementations of the present application, step S130 may be calculated as follows:
[0069] Step S131, calculating the normalized values of multiple source cases on the target problem attributes, including:
[0070]
[0071] in, For source case S i The normalized value on the target problem attribute k, For source case S i The value of the target problem attribute k, min i,k is the minimum value of multiple source cases on the target problem attribute k, max i,k is the maximum value of multiple source cases on the target problem attribute k, i≤I,i∈N + , I is the total number of source cases, k≤K,k∈N + , K is the total number of problem attributes, N + is the set of all positive integers.
[0072] Step S132, calculating the information entropy of the target problem attribute according to the normalized value, including:
[0073]
[0074] Among them, E k is the information entropy of the target problem attribute k.
[0075] Step S133, calculating the weight value of the target problem attribute according to the information entropy, including:
[0076]
[0077] Among them, ω k is the weight value of the target problem attribute k.
[0078] Step S140: Calculate the global similarity between the target case and multiple source cases based on the similarity and weight value.
[0079] Here, after obtaining the similarity between the target case and multiple source cases in the target problem attributes, as well as the weight values corresponding to different problem attributes, the global similarity between the target case and multiple source cases can be calculated using a weighted summation method.
[0080] In some implementations of the present application, step S140 calculates the global similarity between the target case and the multiple source cases by the following formula:
[0081]
[0082] Among them, Sim(S i ,T j ) is the target case T j With source case S i The global similarity between i,k ,T j,k ) is the target case T j Source Case S i The similarity between the target problem attributes k, S i,k For characterizing source case S i The target problem attributes k, T j,k To characterize the target case T j The target problem attribute k.
[0083] Step S150: selecting a target source case from multiple source cases according to the global similarity, so as to guide the action strategy of the target case through the target source case.
[0084] This step obtains the global similarity between the target case and the source case, selects the source case with the highest global similarity (here referred to as the target source case), and Figure 1The provided CBR technical solution guides the action strategy of the target case through the target source case. It should also be noted that the CBR calculation also includes the correction of the case. The embodiment of the present application does not involve the improvement of the correction step, so it will not be described in detail here.
[0085] The unmanned swarm action strategy recommendation method based on case reasoning provided in this embodiment has at least the following effects:
[0086] This method applies case reasoning technology to the process of recommending action strategies for unmanned clusters, and provides a new action strategy recommendation method. By comparing the similarity with historical cases, it can quickly and reasonably find the most similar source case as the recommended guidance for the action strategy of the target case. It can recommend the action strategy of the target case of unmanned clusters in both online and offline states, and with the increase of historical source cases, it can continuously improve the accuracy of the action strategy recommendation of the target case. In the process of finding the target source case, this method also provides the problem attribute similarity and global case similarity calculation methods, which are used to match the target source case with the highest similarity from multiple source cases, which can improve the accuracy of the target source case found and the effectiveness of the action strategy recommendation of the target case.
[0087] In some embodiments of the present application, unmanned swarm action has typical characteristics such as autonomy, confrontation and uncertainty. For example, the effect of the action is closely related to the degree of protection and air defense force of the target, the equipment scale and load quantity of the cluster, the external electromagnetic and physical environment, etc. Except for some attributes such as cluster equipment, which can obtain precise values, most of the information related to the target and the external environment is fuzzy and uncertain. Therefore, the problem attributes of the unmanned swarm action case are divided into four categories, namely deterministic (such as the number of attacking drones), fuzzy (such as the solidity of target facilities), interval (such as target reaction radius) and sub-type (such as physical weather), and the corresponding attribute similarity calculation model is designed for each type.
[0088] (1) When the target problem attribute is a deterministic problem attribute, the similarity between the target case and multiple source cases in the target problem attribute is calculated, including:
[0089]
[0090] Among them, Sim(S i,k ,T j,k ) is the target case T j Source Case S i The similarity between the attributes of the deterministic problem, S i,k For characterizing source case S i The deterministic problem attribute k, T j,k To characterize the target case T jThe deterministic problem attribute k, For source case S i The value of attribute k in the deterministic problem, For the target case T j The value of attribute k in the deterministic problem, max k is the maximum value of the problem attribute k, min k is the minimum value of the problem attribute k. Sim(S i,k ,T j,k ) has a similarity value range of [0,1].
[0091] (2) When the target problem attribute is a fuzzy problem attribute, the similarity between the target case and multiple source cases in the target problem attribute is calculated, including:
[0092] The bulldozer distance represents the optimal distance between two distributions. It is used here to calculate the optimal distance between distributions represented by triangular fuzzy numbers and as the similarity between problem attributes.
[0093] First, remember the problem attribute a k The value space of The problem attribute value represented by triangular fuzzy numbers is given as follows:
[0094]
[0095] Secondly, define the calculation formula of attribute similarity based on bulldozer distance:
[0096] Sim(S i,k ,T j,k )=1-d (6)
[0097] Among them, Sim(S i,k ,T j,k ) is the target case T j Source Case S i The similarity between the attributes of the fuzzy questions, S i,k For characterizing source case S i The fuzzy problem attribute k, T j,k To characterize the target case T j The fuzzy problem attributes k and d are based on the calculation of the bulldozer distance S i,k With T j,k The attribute similarity between To express, The calculation formula based on triangular fuzzy number distribution is as follows:
[0098]
[0099] in,
[0100] (3) When the target problem attribute is an interval-type problem attribute, the similarity between the target case and multiple source cases in the target problem attribute is calculated, including:
[0101] The interval-type numerical similarity is usually calculated using the Jaccard coefficient. Although it has a good degree of leniency, it is slightly insufficient in terms of discrimination. Therefore, the improved Jaccard coefficient method is used here to calculate the interval-type problem attribute similarity:
[0102]
[0103] Among them, Sim(S i,k ,T j,k ) is the target case T j Source Case S i The similarity between the interval problem attributes, S i,k For characterizing source case S i The interval problem attribute k, T j,k To characterize the target case T j The interval problem attribute k, For source case S i The value of attribute k in interval type problem, For the target case T j The value of the interval problem attribute k. and is the number of intervals, that is, the interval consisting of the minimum and maximum values [x - ,x + ], L(·) is the function of measuring the length of the interval number, For measurement Length, For measurement Length, For measurement Length.
[0104] (4) When the target problem attribute is a categorical problem attribute, the similarity between the target case and multiple source cases in the target problem attribute is calculated, including:
[0105] The similarity calculation of the attributes of the categorical question is relatively subjective. To facilitate processing, firstly, the value space of the categorical question attributes {b1, b2, …, b n} is divided into several subspaces {B1,B2,…,B n′}. Each subspace contains multiple similar attribute values, usually n≥n′. Given the classification problem attribute S i,k and T j,k The similarity calculation formula is:
[0106]
[0107] Among them, Sim(S i,k ,T j,k ) is the target case T j With source case S i The similarity between the attributes k of the classification problem, S i,k For characterizing source case S i The classification problem attributes k, T j,k To characterize the target case T j The problem attribute k, For source case S i The value of attribute k in the classification problem, For the target case T j The value of attribute k in the classification problem, B p and B q Any two subspaces of the attribute value space of the classification problem.
[0108] This implementation method divides the problem attributes of unmanned swarm action cases into deterministic, fuzzy, interval and classified types, and designs corresponding attribute similarity calculation models for each type, which enriches the problem description of the target case and can improve the accuracy of finding the corresponding most similar source case for the target case.
[0109] In order to verify the practicality of this application, a set of experimental examples are provided below:
[0110] (1) Experimental background;
[0111] The scenario is that a Red Army unmanned swarm attacks a Blue Army ground facility. The Red Army's unmanned equipment includes several attack drones and reconnaissance drones, which can form groups to carry out swarm operations according to mission requirements. The Blue Army has deployed several anti-drone equipment for ground facilities. Now it is required to recommend the most appropriate Red Army unmanned swarm action strategy for several typical Blue Army ground facilities and guide the formulation of the Red Army unmanned swarm action plan.
[0112] (2) Experimental analysis;
[0113] The case includes: problem description, solution and solution effect. Based on the experimental assumptions, the problem attributes of the unmanned swarm action case are first sorted out, including the red team equipment information, blue team target information and action environment information. Specifically, the red team equipment information represents the main force for implementing the action, and the selected attributes mainly include the number of offensive drones (a1), the number of reconnaissance drones (a2) and the action radius (a3). The blue team target information reflects the basic situation of the action target object, and the selected attributes mainly include the solidity of the facilities (a4), the number of anti-drone equipment (a5), the number of targets (a6) and the reaction distance (a7). The action environment information reflects the environmental factors that have a greater impact on the unmanned swarm action, and the selected attributes mainly include physical weather (a8) and electromagnetic environment (a9).
[0114] Among them, a1, a2, a5, and a6 are deterministic problem attributes, and their values are exact values.
[0115] a3 and a7 are essentially interval distances from 0 to the maximum radius, and should be regarded as interval problem attributes, with values being interval values.
[0116] a4 and a9 are usually described qualitatively and their values have a certain continuity. They should be regarded as fuzzy problem attributes and their values are triangular fuzzy numbers.
[0117] a8 has classification characteristics, such as sunny day (5), cloudy day (4), rainy day (3), snowy day (2) and foggy day (1), and should be processed according to the attributes of the classification problem and take the value of one of the types.
[0118] For ease of processing, a4 and a9 need to be graded first. The possible values of a4 are fragile (1), poor (2), good (3), and strong (4). The possible values of a9 are poor (1), poor (2), average (3), better (4), and good (5).
[0119] Secondly, the solution attribute of the source case is the action strategy taken (r1), which is one of multiple action strategies. Here, it is assumed that there are 6 action strategies (effect priority strategy AT1, time priority strategy AT2, cost priority strategy AT3, effect aversion strategy AT4, time aversion strategy AT5, cost aversion strategy AT6), that is, r1∈{AT1,AT2,AT3,AT4,AT5,AT6}. Finally, the operation effect of the solution includes goal achievement (e1) and cost-effectiveness ratio (e2).
[0120] Among them, the target achievement attribute value may be success or failure, but all cases in this experiment are taken as success, that is, e1 = Yes; the cost-effectiveness attribute value is represented by the number of red party drone losses divided by the number of blue party anti-drone equipment and ground facilities, that is, e2∈[0,1].
[0121] To verify the performance of the method, we searched for public drone action cases on the Internet, sorted out the problem and solution attribute values and execution effects, and manually supplemented some of the unpublished data to construct a dataset with 10 source cases (S1-S 10 ) historical case library. In terms of target cases, four typical blue ground facility scenarios (T1-T4) are designed, as shown in Table 1.
[0122]
[0123] Table 1
[0124] In order to calculate the global similarity between the four target cases and all the source cases, the attribute similarity calculation process is first explained. First, a1, a2, a3, a5, a6 and a7 are deterministic and interval problem attributes, which can be calculated according to the above formulas (5) and (6). Secondly, a8 is a classification problem attribute. According to the impact of weather on drone operations, the subspace is divided into B1 = {5,4}, B2 = {3,2}, B3 = {1}. Finally, a4 and a9 belong to fuzzy problem attributes, and the calculation process is relatively complicated. Here, a4 is taken as an example, and its value space is a4∈{1,2,3,4}, then t=3. The value of the s4 attribute is represented by a triangular fuzzy number, and the source case S is given. i With target case T j The similarity calculation results for all possible values of the a4 attribute are shown in Table 2.
[0125]
[0126] Table 2
[0127] Based on this, the global similarity calculation results between all target cases and source cases are given in Table 3. Taking T1 as an example, the similarity results between the target case and the source case are given in Table 4. Among them, the problem attribute weight calculated based on the entropy weight method is ω k =(0.1741,0.015,0.174,0.094,0.102,0.164,0.201,0.035,0.04).
[0128]
[0129]
[0130] Table 3
[0131] <![CDATA[Sim(S i,k ,T 1,k )]]> <![CDATA[a1]]> <![CDATA[a2]]> <![CDATA[a3]]> <![CDATA[a4]]> <![CDATA[a5]]> <![CDATA[a6]]> <![CDATA[a7]]> <![CDATA[a8]]> <![CDATA[a9]]> <![CDATA[S1]]> 0.5 0.333 0.775 0.599 0.333 0.8 0.707 1 0.805 <![CDATA[S2]]> 0.5 0.667 1 1 0.667 0.8 0.866 0.7 0.441 <![CDATA[S3]]> 0.667 0.667 0.775 0.649 1 1 0.612 0.3 1 <![CDATA[S4]]> 0.333 1 0.816 0.599 0.667 0.6 0.707 0.3 0.473 <![CDATA[S5]]> 0.667 0.667 0.926 0.739 1 0.6 0.935 0.3 0.736 <![CDATA[S6]]> 0.5 0.333 0.775 1 1 0.6 0.707 0.7 0.805 <![CDATA[S7]]> 0.667 0.333 0.739 0.599 0.667 0.8 0.866 0.3 0.805 <![CDATA[S8]]> 0.333 0.667 0.866 0.649 0.667 0.8 0.671 0.3 0.441 <![CDATA[S9]]> 1 0.333 1 0.649 0.667 1 0.707 0.3 0.736 <![CDATA[S 10 ]]> 0.667 0.667 0.775 1 1 0.4 0.758 0.3 0.473
[0132] Table 4
[0133] (3) Experimental conclusions;
[0134] Finally, the recommendation results of the four target cases are summarized in Table 5.
[0135] Target Case Recommended Cases Recommended strategies Policy Type Name <![CDATA[T1]]> <![CDATA[S9]]> AT3 Cost-first strategy <![CDATA[T2]]> <![CDATA[S1]]> AT3 Cost-first strategy <![CDATA[T3]]> <![CDATA[S2]]> AT1 Effect-first strategy <![CDATA[T4]]> <![CDATA[S8]]> AT2 Time priority strategy
[0136] Table 5
[0137] Reference Figure 3 In one embodiment of the present application, a device for recommending an action strategy of an unmanned swarm based on case reasoning is provided. The device for recommending an action strategy of an unmanned swarm based on case reasoning comprises:
[0138] The data acquisition unit is used to acquire a target case of an unmanned cluster and a plurality of corresponding source cases, and to acquire a plurality of problem attributes of the case of the unmanned cluster, wherein the problem attribute is used to characterize a characteristic attribute of the problem of the case.
[0139] The similarity calculation unit is used to calculate the similarity between the target case and the multiple source cases in the target problem attribute, wherein the target problem attribute is any one of the multiple problem attributes.
[0140] The weight value calculation unit is used to calculate the normalized values of multiple source cases on the target problem attribute, calculate the information entropy of the target problem attribute according to the normalized value, and calculate the weight value of the target problem attribute according to the information entropy.
[0141] The global similarity calculation unit is used to calculate the global similarity between the target case and multiple source cases according to the similarity and weight value.
[0142] The source case recommendation unit is used to select a target source case from multiple source cases according to global similarity, so as to guide the action strategy of the target case through the target source case.
[0143] It should be noted that this embodiment and the above method embodiment are based on the same inventive concept, so the relevant contents of the above method embodiment are also applicable to this embodiment and will not be described in detail here.
[0144] Reference Figure 4 The embodiment of the present application further provides an electronic device, the electronic device comprising:
[0145] at least one memory;
[0146] at least one processor;
[0147] at least one program;
[0148] The program is stored in the memory, and the processor executes at least one program to implement the unmanned swarm action strategy recommendation method based on case reasoning described above.
[0149] The electronic device may be any intelligent terminal including a mobile phone, a tablet computer, a personal digital assistant (PDA), a vehicle-mounted computer, etc.
[0150] The electronic device according to the embodiment of the present application is described in detail below.
[0151] The processor 1600 may be implemented by a general-purpose central processing unit (CPU), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application;
[0152] The memory 1700 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 1700 can store an operating system and other application programs. When the technical solution provided in the embodiment of this specification is implemented by software or firmware, the program code is stored in the memory 1700, and the processor 1600 calls and executes the unmanned cluster action strategy recommendation method based on case reasoning in the embodiment of this application.
[0153] Input / output interface 1800, used to implement information input and output;
[0154] The communication interface 1900 is used to realize the communication interaction between the device and other devices. The communication can be realized through a wired manner (such as USB, network cable, etc.) or a wireless manner (such as mobile network, WIFI, Bluetooth, etc.);
[0155] Bus 2000 , which transmits information between various components of the device (e.g., processor 1600 , memory 1700 , input / output interface 1800 , and communication interface 1900 );
[0156] The processor 1600 , the memory 1700 , the input / output interface 1800 , and the communication interface 1900 are connected to each other in communication within the device via the bus 2000 .
[0157] An embodiment of the present application also provides a storage medium, which is a computer-readable storage medium. The computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are used to enable a computer to execute the above-mentioned case-based reasoning unmanned cluster action strategy recommendation method.
[0158] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.
[0159] The embodiments described in this application are intended to more clearly illustrate the technical solutions of the embodiments of this application and do not constitute a limitation on the technical solutions provided by the embodiments of this application. Those skilled in the art will appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.
[0160] Those skilled in the art will appreciate that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than shown in the figures, or a combination of certain steps, or different steps.
[0161] The device embodiments described above are merely illustrative, and the units described as separate components may or may not be physically separated, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
[0162] Those skilled in the art will appreciate that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices may be implemented as software, firmware, hardware, or a suitable combination thereof.
[0163] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0164] It should be understood that in the present application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.
[0165] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.
[0166] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
[0167] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.
[0168] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including multiple instructions to enable an electronic device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of each embodiment method of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, ROM), random access memory (Random Access Memory, RAM), disk or optical disk and other media that can store programs.
[0169] The above is a specific description of the preferred implementation of the embodiments of the present application, but the embodiments of the present application are not limited to the above-mentioned implementation methods. Technical personnel familiar with the field can also make various equivalent modifications or substitutions without violating the spirit of the embodiments of the present application. These equivalent modifications or substitutions are all included in the scope defined by the claims of the embodiments of the present application.
Claims
1. A method for recommending action strategies for unmanned swarms based on case-based reasoning, characterized in that: The unmanned swarm action strategy recommendation method based on case reasoning includes: Obtaining a target case of an unmanned cluster and a plurality of corresponding source cases, and obtaining a plurality of problem attributes of the case of the unmanned cluster, wherein the problem attributes are used to characterize characteristic attributes of the problem of the case; Calculating the similarity between the target case and the multiple source cases in terms of the target problem attribute, wherein the target problem attribute is any one of the multiple problem attributes; Calculating normalized values of the multiple source cases on the target problem attribute, calculating the information entropy of the target problem attribute according to the normalized values, and calculating the weight value of the target problem attribute according to the information entropy; Calculating the global similarity between the target case and the plurality of source cases according to the similarity and the weight value; A target source case is selected from the multiple source cases according to the global similarity, so as to guide the action strategy of the target case through the target source case.
2. The method for recommending action strategies for unmanned swarms based on case-based reasoning according to claim 1, characterized in that: The calculating the normalized values of the plurality of source cases on the target problem attribute comprises: in, For source case S i The normalized value on the target problem attribute k, For source case S i The value of the target problem attribute k, min i,k is the minimum value of the multiple source cases on the target problem attribute k, max i,k is the maximum value of the multiple source cases on the target problem attribute k, i≤I,i∈N + , I is the total number of the multiple source cases, k≤K,k∈N + , K is the total number of the multiple problem attributes, N + is the set of all positive integers; The calculating the information entropy of the target problem attribute according to the normalized value includes: Among them, E k is the information entropy of the target problem attribute k; The calculating the weight value of the target problem attribute according to the information entropy includes: Among them, ω k is the weight value of the target problem attribute k.
3. The method for recommending action strategies for unmanned swarms based on case-based reasoning according to claim 2, characterized in that: The calculating the global similarity between the target case and the plurality of source cases according to the similarity and the weight value includes: Among them, Sim(S i ,T j ) is the target case T j With source case S i The global similarity between i,k ,T j,k ) is the target case T j Source Case S i The similarity between the target problem attributes k, S i,k For characterizing source case S i The target problem attributes k, T j,k To characterize the target case T j The target problem attribute k.
4. The method for recommending action strategies for unmanned swarms based on case-based reasoning according to claim 1, characterized in that: In the case where the target question attribute is a deterministic question attribute, the calculating the similarity between the target case and the plurality of source cases in the target question attribute includes: Among them, Sim(S i,k ,T j,k ) is the target case T j Source Case S i The similarity between the attributes of the deterministic problem, S i,k For characterizing source case S i The deterministic problem attribute k, T j,k To characterize the target case T j The deterministic problem attribute k, For source case S i The value of attribute k in the deterministic problem, For the target case T j The value of attribute k in the deterministic problem, max k is the maximum value of the problem attribute k, min k is the minimum value of the problem attribute k.
5. The method for recommending action strategies for unmanned swarms based on case-based reasoning according to claim 1, characterized in that: In the case where the target question attribute is a fuzzy question attribute, calculating the similarity between the target case and the plurality of source cases in the target question attribute includes: Sim(S i,k ,T j,k )=1-d Among them, Sim(S i,k ,T j,k ) is the target case T j Source Case S i The similarity between the attributes of fuzzy questions, S i,k For characterizing source case S i The fuzzy problem attribute k, T j,j To characterize the target case T j The fuzzy problem attributes k and d are based on the calculation of the bulldozer distance S i,k With T j,k The similarity of attributes between .
6. The method for recommending action strategies for unmanned swarms based on case-based reasoning according to claim 1, characterized in that: In the case where the target question attribute is an interval-type question attribute, calculating the similarity between the target case and the plurality of source cases in the target question attribute includes: Among them, Sim(S i,k ,T j,k ) is the target case T j Source Case S i The similarity between the interval problem attributes, S i,k For characterizing source case S i The interval problem attribute k, T j,k To characterize the target case T j The interval problem attribute k, For source case S i The value of attribute k in interval type problem, For the target case T j The value of attribute k in interval type problem, For measurement Length, For measurement Length, For measurement Length.
7. The method for recommending action strategies for unmanned swarms based on case-based reasoning according to claim 1, characterized in that: In the case where the target question attribute is a categorical question attribute, the calculating the similarity between the target case and the plurality of source cases in the target question attribute includes: Among them, Sim(S i,k ,T j,k ) is the target case T j With source case S i The similarity between the attributes k of the classification problem, S i,k For characterizing source case S i The classification problem attribute k, T j,k To characterize the target case T j The problem attribute k, For source case S i The value of attribute k in the classification problem, For the target case T j The value of attribute k in the classification problem, B p and B q Any two subspaces of the attribute value space of the classification problem.
8. An unmanned swarm action strategy recommendation device based on case reasoning, characterized in that: The unmanned swarm action strategy recommendation device based on case reasoning includes: A data acquisition unit, used to acquire a target case of an unmanned cluster and a plurality of corresponding source cases, and to acquire a plurality of problem attributes of the case of the unmanned cluster, wherein the problem attributes are used to characterize characteristic attributes of the problem of the case; a similarity calculation unit, configured to calculate the similarity between the target case and the plurality of source cases in terms of the target problem attribute, wherein the target problem attribute is any one of the plurality of problem attributes; A weight value calculation unit, used to calculate the normalized values of the multiple source cases on the target problem attribute, calculate the information entropy of the target problem attribute according to the normalized values, and calculate the weight value of the target problem attribute according to the information entropy; A global similarity calculation unit, configured to calculate the global similarity between the target case and the plurality of source cases according to the similarity and the weight value; A source case recommendation unit is used to select a target source case from the multiple source cases according to the global similarity, so as to guide the action strategy of the target case through the target source case.
9. An electronic device, characterized in that: It includes at least one controller and a memory for communicating with the controller; the memory stores instructions that can be executed by the at least one controller, and the instructions are executed by at least one of the controllers to enable at least one of the controllers to execute the case-based reasoning unmanned cluster action strategy recommendation method as described in any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, which, when executed, implements the unmanned swarm action strategy recommendation method based on case reasoning as described in any one of claims 1 to 7.
Citation Information
Patent Citations
Underwater unmanned cluster formation key node identification method based on swarm intelligence
CN115047892A
Consilence of data-mining
WO2007147166A2
