Experimental Space Assisted Generation Method for Accelerated Discovery

Abstract

The present application relates to an experimental space auxiliary generation method for accelerating discovery. The method includes: obtaining original case data from an experimental case database; processing the original case data to determine the scenario of the current combat experiment; performing semantic matching on the current combat experiment and the original case data to obtain at least one recommended experimental index for the current combat experiment; determining at least one recommended experimental factor for the current combat experiment according to the scenario of the experiment; and determining the recommended value-taking method for each experimental index in the current combat experiment according to the recommended experimental index, the recommended experimental factor, and a preset mapping relationship. This method can assist in the generation of the experimental space by automatically generating various indicators of the experimental space, simplify the complexity of relevant work links, reduce the workload of experimental personnel, and thus improve the efficiency of combat experiments.

Description

Technical Field

[0001] This application relates to the field of simulation technology, and particularly to an experimental space assisted generation method for accelerating discovery. Background Art

[0002] Combat experiments are scientific experimental activities for studying combat problems. That is, by applying the principles, methods, and technologies of scientific experiments, in a controllable and measurable virtual confrontation environment, empirically study the characteristics and laws of war and military combat operations, and provide a scientific basis for military decision-making and war practice.

[0003] With the development of the war form towards intelligence, multi-domain, and systemization, the experimental plan design in modern war experiments has become increasingly complex, and there are more and more factors affecting the generation of the experimental space. In order to explore the laws in war through combat experiments, it is usually necessary to carry out a series of compilations, settings, and selections for the experimental scenario, experimental indicators, experimental factors, and factor levels, so as to generate a large number of combat experiment sample points and generate an initial combat experiment space. The traditional experimental space generation process based on manual design completely relies on the experience of experimental personnel, has high requirements for experimental personnel in terms of experimental scenarios, military knowledge, mathematical statistics, etc., can only construct an experimental space with a limited scale, and the entire process takes a long time, making it difficult to meet the requirements of modern joint combat experiments. Summary of the Invention

[0004] Based on this, in view of the above technical problems, it is necessary to provide an experimental space assisted generation method for accelerating discovery, which can simplify the complexity of relevant work links, reduce the workload of experimental personnel, and improve the efficiency of combat experiments.

[0005] This application provides an experimental space assisted generation method for accelerating discovery, and the method includes:

[0006] Obtain original case data from an experimental case database;

[0007] Process the original case data to determine the experimental scenario of the current combat experiment;

[0008] Perform semantic matching on the current combat experiment and the original case data to obtain at least one recommended experimental indicator for the current combat experiment;

[0009] Determine at least one recommended experimental factor for the current combat experiment according to the experimental scenario;

[0010] Determine the recommended value-taking method for each experimental indicator in the current combat experiment according to the recommended experimental indicators, the recommended experimental factors, and a preset mapping relationship.

[0011] Among them, the described experiment scenario includes multiple elements; the processing of the original case data to determine the experiment scenario of the current combat experiment further includes:

[0012] Obtain an experiment scenario template;

[0013] Perform templating processing on the original case data according to the experiment scenario template to obtain templated case data;

[0014] Match the experiment questions of the current combat experiment with the templated case data, respectively determine the recommended data corresponding to each element, and obtain the experiment scenario of the current combat experiment.

[0015] Among them, the multiple elements of the experiment scenario include at least one combination of military background, force deployment, force composition, geographical environment, action plan, etc.;

[0016] The experiment scenario template is obtained by analyzing the scenario characteristics of the original case data, and the experiment scenario template includes multiple elements.

[0017] Among them, the semantic matching of the current combat experiment and the original case data to obtain at least one recommended experiment index for the current combat experiment further includes:

[0018] Extract semantic features from the original case data to obtain the first semantic feature vector corresponding to the original case data;

[0019] Extract semantic features from the current combat experiment to obtain the second semantic feature vector corresponding to the current combat experiment;

[0020] Calculate the semantic matching degree according to the first semantic feature vector and the second semantic feature vector;

[0021] Select target case data from the original case data according to the semantic matching degree;

[0022] Determine at least one recommended experiment index for the current combat experiment according to the target case data.

[0023] Among them, the selecting of target case data from the original case data according to the semantic matching degree further includes:

[0024] If the semantic matching degree between the original case data and the current combat experiment is greater than or equal to a preset threshold, then determine the original case data as the target case data.

[0025] Among them, the method further includes:

[0026] Obtain at least one experiment index corresponding to the target case data;

[0027] According to the case matching degree and usage frequency, at least one experimental objective corresponding to the target case data is sequentially used as the recommended experimental index of the current combat experiment.

[0028] Among them, determining at least one recommended experimental factor of the current combat experiment according to the experimental scenario further includes:

[0029] Determine a set of candidate factors according to the experimental scenario;

[0030] Input the semantic features of the current combat experiment into a semantic network for operation to obtain at least one recommended experimental factor of the current combat experiment; wherein, the semantic network is used to match and sort the semantic features of the current combat experiment with the set of candidate factors.

[0031] Among them, the set of candidate factors includes at least one candidate factor; determining the set of candidate factors according to the experimental scenario further includes:

[0032] Use the node parameters and / or model parameters of the experimental scenario as candidate factors to determine the set of candidate factors;

[0033] Among them, the experimental scenario is a structured experimental scenario;

[0034] The node parameters refer to the variable parameters in the task nodes and action nodes that support the operation of the experimental scenario;

[0035] The model parameters refer to the variable parameters in the entity model, rule model, and environment model that support the operation of the experimental scenario.

[0036] Among them, the method further includes:

[0037] Extract the experimental features of the original case data to label the experimental case data;

[0038] Train a neural network according to the labeled experimental case data to construct the semantic network.

[0039] Among them, the preset mapping relationship includes the corresponding relationship between experimental indicators, experimental factors, and the value-taking methods of experimental factors; the method further includes:

[0040] Analyze the original case data in the case database to obtain the corresponding relationship between experimental factors and experimental indicators in each original case data;

[0041] According to the corresponding relationship between experimental factors and experimental indicators in each original case data, statistically obtain the value-taking methods under different corresponding relationships, and generate and store a value-taking method library.

[0042] The above-mentioned experimental space assisted generation method for accelerated discovery can automatically generate the experimental scenario of combat experiments, and automatically select experimental indicators, experimental factors and their value-taking methods, so as to generate the experimental space of combat. By automatically generating various indicators of the experimental space, this method can assist in the generation of the experimental space, simplify the complexity of relevant work links, reduce the workload of experimental personnel, and thus improve the efficiency of combat experiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 FIG. is a schematic flow chart of the experimental space assisted generation method for accelerated discovery in one embodiment;

[0044] Figure 2 FIG. is a schematic flow chart of the determination process of the experimental scenario in one embodiment;

[0045] Figure 3 FIG. is a schematic flow chart of the experimental indicator recommendation process in one embodiment;

[0046] Figure 4 FIG. is a schematic flow chart of the experimental factor recommendation process in one embodiment;

[0047] Figure 5 FIG. is a schematic flow chart of the value-taking method recommendation process of experimental indicators in one embodiment. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0048] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to 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.

[0049] Combat experiment is a scientific experimental activity for studying combat problems, that is, using the principles, methods and technologies of scientific experiments to empirically study the characteristics and laws of war and military combat operations in a controllable and measurable virtual confrontation environment, so as to provide a scientific basis for military decision-making and war practice.

[0050] With the development of the war form towards the directions of intelligence, multi-domain and system, the experimental plan design in modern war experiments is becoming increasingly complex, and there are more and more factors affecting the generation of the experimental space. In order to explore the laws in war through combat experiments, it is usually necessary to carry out a series of compilations, settings and selections for the experimental scenario, experimental indicators, experimental factors and factor levels, so as to generate a large number of combat experiment sample points and generate an initial combat experimental space. The traditional experimental space generation process based on manual design completely relies on the experience of experimental personnel, has high requirements for experimental personnel in terms of experimental scenarios, military knowledge, mathematical statistics, etc., can only construct an experimental space with a limited scale, and the entire process takes a long time, making it difficult to meet the requirements of modern joint combat experiments.

[0051] Based on this, the present application provides an experimental space assisted generation method for accelerating discovery. This method can automatically generate the scenario of a combat experiment, and automatically select experimental indicators, experimental factors, and their value-taking methods, so as to generate the experimental space of combat. By automatically generating various indicators of the experimental space, this method can assist in the generation of the experimental space, simplify the complexity of relevant work links, reduce the workload of experimental personnel, and thus improve the efficiency of combat experiments.

[0052] As Figure 1 shown, the present application provides an experimental space assisted generation method for accelerating discovery, and the method may include:

[0053] S110. Obtain the original case data from the experimental case database; wherein, the original case data may be the combat experiment cases corresponding to historical combat experiments, and at least one combat experiment case corresponding to multiple historical combat experiments is stored in an experimental case database.

[0054] S120. Process the original case data to determine the scenario of the current combat experiment;

[0055] Among them, the scenario refers to the assumption of the basic situation, combat intention, and combat development of both sides in combat. Each combat experiment may correspond to a scenario, which is used to represent various assumptions of the current combat experiment. Optionally, the scenario may include a combination of one or more elements such as military background, force deployment, force composition, geographical environment, and action plan.

[0056] Embodiments of the present application can intelligently determine the scenario of the current combat experiment based on the original case data corresponding to historical combat experiments and in combination with the experimental requirements of the current combat experiment. Among them, the scenario of the current combat experiment may be a combination of multiple elements such as military background, force deployment, force composition, geographical environment, and action plan.

[0057] S130. Perform semantic matching on the current combat experiment and the original case data to obtain at least one recommended experimental indicator for the current combat experiment.

[0058] Embodiments of the present application can obtain the semantic features of the current combat experiment and the semantic features of at least one original case data in the experimental case database, and perform semantic matching on the semantic features of the current combat experiment and the semantic features of at least one original case data, so as to determine at least one recommended experimental indicator for the current combat experiment according to the experimental indicators of the original case data, thereby realizing the intelligent assisted screening of the experimental indicators of the current combat experiment. Among them, at least one recommended experimental indicator corresponding to the current combat experiment may be displayed in the form of a list, and each recommended experimental indicator in the list is arranged in a certain order.

[0059] S140. Determine at least one recommended experimental factor for the current combat experiment according to the described experimental scenario;

[0060] In the embodiments of the present application, the intelligent assisted screening of experimental factors is obtained through in-depth analysis of the experimental scenario, so as to recommend at least one recommended experimental factor for the experimenter to choose. Among them, at least one recommended experimental factor corresponding to the current combat experiment can be displayed in the form of a list, and each recommended experimental factor in the list is arranged in a certain order.

[0061] S150. Determine the recommended value-taking method for each experimental index in the current combat experiment according to the recommended experimental index, the recommended experimental factor and the preset mapping relationship.

[0062] Among them, the preset mapping relationship can include the corresponding relationship between the experimental index, the experimental factor and the value-taking method of the experimental factor. In the embodiments of the present application, the recommendation of the value-taking method of the experimental factor is based on the experimental factor to be set currently and the determined experimental index, and the relationship between the current independent variable and the dependent variable is deduced from the preset mapping relationship statistically obtained from the experimental cases, and then the best value-taking method is indexed from the factor value-taking method library, and this best value-taking method is used as the recommended value-taking method. Thus, this method can intelligently recommend the value-taking method of the experimental factor for the experimenter and assist in setting the experimental factor level.

[0063] It can be seen that the method provided by the present application can intelligently generate the experimental scenario of the combat experiment, and intelligently select the experimental index, the experimental factor and its value-taking method, so as to assist in the generation of the experimental space, simplify the complexity of the relevant work links, reduce the work burden of the experimenter, and thus improve the efficiency of the combat experiment.

[0064] Optionally, the intelligent assisted generation of the experimental scenario in the embodiments of the present application can be realized based on the experimental scenario template and case data recommendation. As Figure 2 shown, the above step S120 may further include the following steps:

[0065] S210. Obtain the experimental scenario template; wherein, the experimental scenario template is used to uniformly format the experimental scenarios in the experimental case data, and then recommend the case data as needed for the experimenter when formulating the experimental scenario, so as to obtain the experimental scenario of the current combat experiment.

[0066] In the embodiments of the present application, the experimental scenario template is obtained by analyzing the scenario characteristics of the original case data, and the experimental scenario template includes multiple elements. Specifically, the experimental scenario template can be obtained by extracting the scenario characteristics of the original case data and abstracting the corresponding framework from multiple elements of the experimental scenario (such as military background, force deployment, force composition, geographical environment and action plan).

[0067] S220. Perform templating processing on the original case data according to the experimental scenario template to obtain templated case data. As Figure 2 shown, in the embodiment of the present application, at least one piece of original case data in the experimental case database can be processed according to the determined experimental scenario template, so that all the original case data has a unified form. After the original case data is processed by templating, the corresponding templated case data can be obtained. The templated case data refers to the case data in which the experimental scenario of the case data is represented according to the experimental scenario template. Optionally, the templated case data can be stored in a database.

[0068] S230. Perform case matching between the experimental problem of the current combat experiment and the templated case data, respectively determine the recommended data corresponding to each element, and obtain the experimental scenario of the current combat experiment.

[0069] Case data recommendation is to, after processing the experimental scenario of the original case data according to the experimental scenario template, based on the case-based reasoning theory, recommend the information of the experimental scenario in similar experimental cases as recommended data to the experimenters, so as to quickly specify the experimental scenario. Among them, the case-based reasoning theory is a technology for reasoning about solving similar current problems based on past experience knowledge. It emphasizes that when solving a new problem, find the most relevant case to the current problem from memory or the case library, and solve the current problem based on this.

[0070] The method of the present application can, based on the case-based reasoning theory, perform case matching between the experimental problem of the current combat experiment and at least one piece of templated case data, so as to select the case data most relevant to the current combat experiment from at least one piece of templated case data, and determine the recommended data for each element of the experimental scenario in the current combat experiment according to the experimental scenario corresponding to the selected case data, so as to obtain the experimental scenario of the current combat experiment. Optionally, the experimental scenario of the current combat experiment can be represented according to the above experimental scenario template.

[0071] Optionally, the intelligent auxiliary selection method for experimental indicators can include steps such as experimental feature extraction, semantic matching, and experimental indicator recommendation. As Figure 3 shown, the above step S130 can further include:

[0072] S310. Extract semantic features from the original case data to obtain the first semantic feature vector corresponding to the original case data.

[0073] Specifically, the method of the present application can extract experimental features from at least one piece of original case data in the experimental case database according to the combat experiment theory and combat experiment case data, and obtain the first semantic feature corresponding to each piece of original case data according to the above experimental features. The first semantic feature can be represented in the form of a vector, that is, the first semantic feature vector. Among them, the experimental features can include experimental name, experimental question, experimental object, experimental background, etc. Further, the method of the present application can also perform data annotation on each piece of original case data in the experimental case database based on the extraction of the above experimental features to obtain the annotated case data, that is, each piece of annotated case data has a corresponding first semantic feature vector. The annotated case data can be stored in a database for subsequent semantic matching.

[0074] S320. Extract semantic features of the current combat experiment to obtain a second semantic feature vector corresponding to the current combat experiment.

[0075] Specifically, the method of the present application can also extract semantic features of the current combat experiment according to the above experimental features to obtain a second semantic feature corresponding to the current combat experiment. The second semantic feature can be represented in the form of a vector, that is, the second semantic feature vector.

[0076] S330. Calculate the semantic matching degree according to the first semantic feature vector and the second semantic feature vector, and select target case data from the original case data according to the semantic matching degree.

[0077] Specifically, the method of the present application can perform semantic matching between the second semantic feature vector corresponding to the current combat experiment and the first semantic feature vectors of each case data in the database, and obtain the corresponding semantic matching degree. Optionally, the method of the present application can use the cosine similarity algorithm or the neural network algorithm for semantic matching to calculate the semantic matching degree.

[0078] Further, based on the calculated semantic matching degree, the method of the present application can select the original case data from the experimental case database as the target case data to determine at least one recommended experimental index corresponding to the current combat experiment according to the target case data.

[0079] Further optionally, if the semantic matching degree between the original case data and the current combat experiment is greater than or equal to a preset threshold, the original case data is determined as the target case data. That is, if the semantic matching degree calculated according to the first semantic feature vector and the second semantic feature vector corresponding to the original case data is greater than or equal to the preset threshold, the original case data is used as the target case data. If the semantic matching degree calculated according to the first semantic feature vector and the second semantic feature vector corresponding to the original case data is less than the preset threshold, the target case data is selected from other original case data in the experimental case database.

[0080] S340. Determine at least one recommended experimental index for the current combat experiment according to the target case data.

[0081] The method of the present application can further recommend the experimental indexes corresponding to the target case data to the experimenter to assist in generating at least one recommended experimental index for the current combat experiment. Specifically, the method of the present application can obtain at least one experimental index corresponding to the target case data, and sequentially use at least one experimental target corresponding to the target case data as the recommended experimental index of the current combat experiment according to the case matching degree and the usage frequency of the experimental index. As Figure 3 shown, the recommended experimental indexes can be presented to the experimenter in the form of a list, and the experimental indexes in the list are arranged in order according to factors such as the case matching degree and the usage frequency of the experimental index. The experimenter can select the experimental indexes of the current combat experiment from the recommended list, so as to achieve the purpose of assisting in generating experimental indexes.

[0082] Optionally, the intelligent auxiliary screening method for experimental factors is based on the experimental scenario, supported by semantic analysis technology and case-based reasoning theory, and recommends optional experimental factors to the experimenter in order. The intelligent auxiliary screening method for experimental factors can include a step of obtaining a set of candidate factors, constructing a semantic network for experimental factor recommendation, and experimental factor recommendation based on the semantic network. As Figure 4 shown, the above step S140 can further include:

[0083] S410. Determine a set of candidate factors according to the experimental scenario;

[0084] Wherein, the experimental scenario is a structured experimental scenario, and the structured experimental scenario can be obtained based on the experimental scenario determined in step S120. As Figure 4 shown, the structured experimental scenario can be in a tree structure, and the structured experimental scenario includes multiple nodes.

[0085] The node parameters refer to the variable parameters in the task nodes and action nodes that support the operation of the experimental scenario; the model parameters refer to the variable parameters in the entity model, rule model, and environment model that support the operation of the experimental scenario.

[0086] The method of this application can traverse the above-mentioned structured experimental scenario, obtain the node parameters of the structured experimental scenario, and the node parameters of this experimental scenario can be the candidate factors that have not been used in the experimental scenario. The method of this application can also obtain the entity model, rule model, and environment model that support the operation of this experimental scenario through entity extraction, model matching, etc., and select the variable parameters (i.e., model parameters) of these models as candidate factors. Further, the method of this application can determine the set of candidate factors for realizing the scenario based on the above-mentioned node parameters and / or model parameters.

[0087] S420. Input the semantic features of the current combat experiment into a semantic network for operation to obtain at least one recommended experimental factor for the current combat experiment; wherein, the semantic network is used to match and sort the semantic features of the current combat experiment with the set of candidate factors, and this semantic network can be obtained through neural network training based on the case data in the experimental case database.

[0088] Specifically, the method of this application can be based on at least one original case data in the experimental case database, supported by natural language processing technology and knowledge graph theory, extract the experimental features of each case data, and label the case data based on the extracted experimental features to obtain the labeled experimental case data. The process of labeling the cases can refer to the above-mentioned process of determining the first feature vector. Further, the method of this application can use the labeled experimental case data to train a neural network to obtain this semantic network, so that this semantic network can be used to recommend experimental factors.

[0089] After constructing the semantic network, the method of this application can input the semantic features of the current combat experiment (such as the above-mentioned second feature vector) into a semantic network for operation. The semantic network can match and sort the candidate factors in the set of candidate factors according to the semantic features of the current combat experiment, so as to obtain at least one recommended experimental factor for the current combat experiment, and recommend and display this at least one recommended experimental factor to the experimenter. As Figure 4 shown, the recommended experimental factors can be displayed to the experimenter in the form of a list, and the experimental factors in the list are arranged in order according to factors such as case matching degree and usage frequency of experimental indicators. The experimenter can select the experimental factors of the current combat experiment from this recommended list, so as to achieve the purpose of assisting in generating experimental factors.

[0090] Optionally, the intelligent auxiliary setting of experimental factor levels is based on experimental cases, supported by the value-taking methods of various experimental factors, and recommends the value-taking methods of experimental factors for experimenters based on the correlation between experimental indicators and experimental factors in the experimental cases, so as to assist in setting the levels of experimental factors. To support the recommendation process of experimental factor value-taking methods and achieve the intelligent auxiliary setting of experimental factor levels, the method of this application may further include a statistical extraction method for the relationship between experimental factors and experimental indicators and the construction of an experimental factor value-taking method library. As Figure 5 shown, the above method may further include:

[0091] S510. Analyze the original case data in the case database to obtain the corresponding relationship between experimental factors and experimental indicators in each original case data.

[0092] Among them, the method of this application can analyze at least one original case data in the experimental case database respectively to obtain the corresponding relationship between experimental factors and experimental indicators in each original case data, and lay a foundation for the recommendation of value-taking methods by statistically analyzing the value-taking methods of experimental indicators under different relationships.

[0093] As Figure 5 shown, an experimental factor a can correspond to multiple experimental indicators a1, a2... The value-taking method of experimental indicator a1 can be value-taking method a1, and the value-taking method of experimental indicator a2 can be value-taking method a2... Similarly, an experimental factor b can correspond to multiple experimental indicators b1, b2... The value-taking method of experimental indicator b1 can be value-taking method b1, and the value-taking method of experimental indicator b2 can be value-taking method b2...

[0094] S520. According to the corresponding relationship between experimental factors and experimental indicators in each original case data, statistically obtain the value-taking methods under different corresponding relationships, and generate and store a value-taking method library.

[0095] The method of this application can also statistically obtain the value-taking methods of experimental factors under different corresponding relationships between experimental factors and experimental indicators in the above at least one original case data, and establish an index with the variable relationship applicable to the experimental factor value-taking method (i.e., the corresponding relationship between experimental factors and experimental indicators), so as to generate a value-taking method library. Further, the value-taking method library can be stored in a database.

[0096] As Figure 5As shown, based on the statistical analysis of each original case data in the experimental case database, the method of this application can first obtain the corresponding relationships between experimental factors, experimental indicators, and their value-taking methods in each original case data. Then, based on the corresponding relationships of different experimental case data, statistical analysis can be performed to obtain a value-taking method library for different experimental factors, and this value-taking method library can be stored in the form of a database. For example, experimental factor X can correspond to multiple experimental indicators (experimental indicator 1, experimental indicator 2... experimental indicator n). In the intelligent experimental factor level recommendation step, the method of this application can be indexed through the experimental factors and experimental coordinates of the current combat experiment to retrieve the current best value-taking method from the value-taking method library according to the preset mapping relationship, and display this value-taking method as the recommended value-taking method to the experimenters, so as to assist the experimenters in setting the experimental factor levels.

[0097] The experimental space auxiliary generation technology for accelerating discovery proposed in this application relies on technologies such as natural language processing technology, case matching technology, and deep neural network algorithms to construct an experimental space auxiliary generation system composed of four sub-modules: intelligent auxiliary generation based on experimental scenarios, intelligent auxiliary selection of experimental indicators, intelligent auxiliary screening of experimental factors, and intelligent auxiliary setting of factor levels. It can simplify the complexity of the experimental space generation link, reduce the workload of experimenters, and thus improve the efficiency of combat experiments.

[0098] It should be understood that although Figures 1 - 5 the steps in the flowchart of Figures 1 - 5 are shown in sequence according to the arrows, these steps do not necessarily need to be executed in the order indicated by the arrows. Unless there is a clear indication in this article, the execution of these steps has no strict order limit, and these steps can be executed in other orders. Moreover,

[0099] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combinations of these technical features do not conflict, they should all be considered as the scope described in this specification.

[0100] The above-described embodiments merely represent several implementation manners of the present application. The description thereof is relatively specific and detailed, but it should not be construed as a limitation on the scope of the invention patent. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all fall within the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the appended claims.

Claims

1. A method for assisted generation of experimental space for accelerated discovery, characterized in that: The method comprises: Obtain original case data from the experimental case database; Processing the original case data to determine the experimental scenario of the current combat experiment; Performing semantic matching on the semantic features of the current operational experiment and the original case data to obtain at least one recommended experimental indicator of the current operational experiment; Determine at least one recommended experimental factor for the current operational experiment based on the experimental scenario; wherein, determining at least one recommended experimental factor for the current operational experiment based on the experimental scenario further includes: determining a set of candidate factors based on the experimental scenario; inputting the semantic features of the current operational experiment into a semantic network for operation to obtain at least one recommended experimental factor for the current operational experiment; wherein, the semantic network can be obtained by neural network training based on case data in an experimental case database; the semantic network is used to match and sort the semantic features of the current operational experiment with the set of candidate factors; Determining a recommended value-taking method for each experimental indicator in the current operational experiment based on the recommended experimental indicator, the recommended experimental factor, and a preset mapping relationship; wherein the preset mapping relationship includes a correspondence between the experimental indicator, the experimental factor, and the value-taking method for the experimental factor; the method further includes: Analyzing the original case data in the case database to obtain the corresponding relationship between the experimental factors and the experimental indicators in each original case data; According to the correspondence between experimental factors and experimental indicators in each original case data, the value-taking methods under different correspondences are statistically obtained, and a value-taking method library is generated and stored.

2. The method according to claim 1, characterized in that The experimental scenario includes multiple elements; processing the original case data to determine the experimental scenario of the current combat experiment also includes: Get the experimental scenario template; Performing template processing on the original case data according to the experimental scenario template to obtain templated case data; The experimental questions of the current operational experiment are matched with the templated case data, and the recommended data corresponding to each element is determined respectively to obtain the experimental assumptions of the current operational experiment.

3. The method according to claim 2, characterized in that The multiple elements of the experimental scenario include a combination of at least one of military background, force deployment, force composition, geographical environment, and action plan; The experimental scenario template is obtained based on the scenario feature analysis of the original case data, and the experimental scenario template includes multiple elements.

4. The method according to any one of claims 1 to 3, characterized in that The semantic matching of the current operational experiment and the original case data to obtain at least one recommended experimental indicator for the current operational experiment further includes: Performing semantic feature extraction on the original case data to obtain a first semantic feature vector corresponding to the original case data; Performing semantic feature extraction on the current combat experiment to obtain a second semantic feature vector corresponding to the current combat experiment; Calculating a semantic matching degree based on the first semantic feature vector and the second semantic feature vector; selecting target case data from the original case data according to the semantic matching degree; At least one recommended experimental indicator of the current operational experiment is determined based on the target case data.

5. The method according to claim 4, characterized in that The step of selecting target case data from the original case data according to the semantic matching degree further includes: If the semantic matching degree between the original case data and the current operational experiment is greater than or equal to a preset threshold, the original case data is determined as the target case data.

6. The method according to claim 4, characterized in that The method further comprises: Obtaining at least one experimental indicator corresponding to the target case data; According to the case matching degree and the usage frequency, at least one experimental target corresponding to the target case data is sequentially used as a recommended experimental indicator for the current operational experiment.

7. The method according to claim 1, characterized in that The set of candidate factors includes at least one candidate factor; and determining the set of candidate factors according to the experimental scenario further includes: Taking the node parameters and / or model parameters of the experimental scenario as candidate factors to determine a candidate factor set; Wherein, the experimental scenario is a structured experimental scenario; The node parameters refer to the variable parameters in the task nodes and action nodes that support the operation of the experimental scenario; The model parameters refer to the variable parameters in the entity model, rule model and environment model that support the operation of the experimental scenario.

8. The method according to claim 1, characterized in that The method further comprises: extracting experimental features of the original case data to label the experimental case data; A neural network is trained according to the labeled experimental case data to construct the semantic network.

Images