Method for describing military simulation scenario based on knowledge graph

Abstract

The application relates to the technical field of computer simulation, and discloses a military simulation scenario description method based on a knowledge graph, which comprises the following steps: acquiring command message data and electromagnetic environment link state data, and dividing a command session set according to message header identification; constructing a multi-layer associated knowledge graph comprising a session layer, a sentence layer and a traceability layer; according to the electromagnetic environment link state data, the cumulative link connection duration is counted, and when a preset threshold is met, a scenario grounding operation is triggered, and an execution element layer is generated in the graph; the graph element scale change amplitude before and after each grounding operation is counted, and a session traceability inflation index is generated through weighted processing; finally, a complexity identification node carrying the index is established and written into the graph, and the final knowledge graph file, the replayable command message sequence and the simulation executable scene package are output. The application can quantify the structural inflation risk caused by the environment and the protocol mechanism in the description stage, and is suitable for large-scale simulation interoperation.

Description

Technical Field

[0001] This invention relates to the field of computer simulation technology, and more specifically, to a method for describing military simulation scenarios based on knowledge graphs. Background Technology

[0002] In the field of distributed federated simulation and command and control system interoperability (C2SIM), the description and generation of military simulation scenarios are crucial links connecting operational intentions with simulation execution. Typically, operational scenarios need to be transformed into machine-readable, executable scenarios for initialization, task distribution, and situational awareness simulations within heterogeneous simulation systems. With technological advancements, knowledge graphs, due to their powerful semantic representation capabilities, are increasingly being used to describe complex battlefield entities, command relationships, and mission flows. Standards such as C2SIM utilize XML or ontology models to standardize message formats, supporting command issuance, status monitoring, and message stream recording and playback.

[0003] However, in actual tactical communication simulations and adversarial exercises, existing scenario description methods have significant limitations.

[0004] First, existing description systems often focus on static snapshots of the situation, lacking a detailed depiction of the entire lifecycle of conversational message flows. Under the C2SIM standard, a command is not isolated but includes a sender, receiver, session identifier, acknowledgment, and potential timeout retransmission. To fully represent these interactions in a knowledge graph, a large amount of statement-level metadata and a source structure must be introduced. Existing conventional graph construction methods often lead to a sharp increase in the size of the graph's nodes and edges when processing such high-frequency interaction data, lacking effective structured measurement methods.

[0005] Secondly, existing scenario generation methods fail to fully consider the coupling effect between complex electromagnetic environments and communication protocol mechanisms. In adversarial electromagnetic environments, communication links exhibit unstable on / off states. When a link is in a rejected state, reliable transmission mechanisms such as C2SIM trigger repeated retransmissions and status queries. Current description methods typically assume the network state is ideal or only perform simple packet loss simulations, neglecting the fact that when the simulation element instantiation (i.e., scenario grounding) occurs, the accumulation of historical messages and confirmation of receipts can introduce a massive amount of tracing information instantaneously.

[0006] This deficiency makes it difficult to predict the extent of expansion of the knowledge graph structure when constructing knowledge graphs for large-scale simulation scenarios. As the size of the receiver increases or electromagnetic interference intensifies, the storage overhead and query cost of the graph may increase exponentially, leading to resource exhaustion or response delays in the simulation system during loading or playback analysis. Summary of the Invention

[0007] This invention provides a method for describing military simulation scenarios based on knowledge graphs, which solves the technical problems mentioned in the background art.

[0008] This invention provides a method for describing military simulation scenarios based on knowledge graphs, including:

[0009] Acquire command and control message data and electromagnetic environment link status data, and divide the command and control message data into command session sets according to the message header identifier;

[0010] Construct a multi-layered association knowledge graph that maps the command session set. The multi-layered association knowledge graph includes a session layer that describes the session affiliation relationship, a statement layer that encapsulates instruction semantics and metadata, and a tracing layer that records message derivation and interaction history.

[0011] Based on the electromagnetic environment link status data, the cumulative link connection time is counted. When the cumulative link connection time meets the preset simulation element instantiation threshold, the scenario grounding operation is triggered, and an execution element layer containing simulation entity objects and task parameter objects is generated in the multi-layer association knowledge graph.

[0012] The change in the size of the graph primitives before and after each of the proposed grounding operations is statistically analyzed. All the change amplitudes in the command session set are aggregated and weighted according to the number of receiving nodes to generate a session tracing expansion index that characterizes the increase in graph structure complexity.

[0013] Establish a complexity identifier node that carries the session tracing expansion index and associate it with the multi-layer association knowledge graph. Output a knowledge graph file containing the complexity identifier node, a replayable command and control message sequence, and a simulation executable scenario package.

[0014] The beneficial effects of this invention are as follows: By introducing a grounding triggering mechanism based on the cumulative connection duration of links, the problem of uncontrollable expansion of the knowledge graph structure caused by the instantiation of simulation elements in adversarial electromagnetic environments and complex session interaction scenarios is effectively solved; This invention utilizes the session source expansion index to achieve quantitative prediction and risk identification of the growth rate of the graph primitive size in the scenario description stage, avoiding the interruption of the simulation system due to resource exhaustion during loading or playback, and significantly improving the stability, replayability, and robustness to complex electromagnetic environments of the command and control message flow processing in large-scale federated simulation. Attached Figure Description

[0015] Figure 1 This is a diagram illustrating the grounding step-induced session tracing expansion mechanism of the present invention;

[0016] Figure 2 This is a scenario architecture diagram of the present invention for command and control, simulation interoperability and playback in an electromagnetic environment. Detailed Implementation

[0017] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed only to enable those skilled in the art to better understand and implement the subject matter described herein, and changes may be made to the function and arrangement of the elements discussed without departing from the scope of this specification. Various processes or components may be omitted, substituted, or added as needed in the examples. Furthermore, features described in some examples may be combined in other examples.

[0018] like Figure 1 As shown, the knowledge graph-based method for describing military simulation scenarios includes:

[0019] Acquire command and control message data and electromagnetic environment link status data, and divide the command and control message data into command session sets according to the message header identifier;

[0020] Construct a multi-layered association knowledge graph that maps the command session set. The multi-layered association knowledge graph includes a session layer that describes the session affiliation relationship, a statement layer that encapsulates instruction semantics and metadata, and a tracing layer that records message derivation and interaction history.

[0021] Based on the electromagnetic environment link status data, the cumulative link connection time is counted. When the cumulative link connection time meets the preset simulation element instantiation threshold, the scenario grounding operation is triggered, and an execution element layer containing simulation entity objects and task parameter objects is generated in the multi-layer association knowledge graph.

[0022] The change in the size of the graph primitives before and after each of the proposed grounding operations is statistically analyzed. All the change amplitudes in the command session set are aggregated and weighted according to the number of receiving nodes to generate a session tracing expansion index that characterizes the increase in graph structure complexity.

[0023] Establish a complexity identifier node that carries the session tracing expansion index and associate it with the multi-layer association knowledge graph. Output a knowledge graph file containing the complexity identifier node, a replayable command and control message sequence, and a simulation executable scenario package.

[0024] In a preferred embodiment, command and control message data and electromagnetic environment link status data are acquired, and the command and control message data is divided into command session sets based on message header identifiers, including:

[0025] For any object to be identified in the charge message data Generate hash identifier The calculation formula is:

[0026]

[0027] SHA256 stands for Secure Hash Algorithm. Indicates the object to be identified Perform normalized data concatenation; hex indicates hexadecimal output.

[0028] Parse the message header fields of the charge message data to extract the session identifier. With the ReceiverSet list;

[0029] The charge message data is aggregated into a command session set C using the session identifier, and the calculation formula is as follows:

[0030]

[0031] And obtain the set of receiving nodes. The calculation formula is:

[0032]

[0033] Where X represents the charge message data, m represents a single charge message, cid represents the identifier of a specific session, and c represents a single command session.

[0034] Employing a secure hash algorithm to calculate identifiers from the concatenated normalized data is crucial for ensuring absolute consistency in naming the same object across all heterogeneous systems in a distributed simulation environment without a central node, thus preventing association failures caused by naming offsets. Grouping data based on session identifiers in the message header is grounded in the command and control system's business logic. This means that the causal chain of instructions, retransmission behavior, and the resulting structural bloat all occur within independent session threads; therefore, isolation processing at the session level is essential for accurate localized measurement.

[0035] In a preferred embodiment, a multi-layered relational knowledge graph mapping the command session set is constructed. This multi-layered relational knowledge graph includes a session layer describing session affiliation, a statement layer encapsulating instruction semantics and metadata, and a tracing layer recording message derivation and interaction history.

[0036] Based on a preset statement template library, each command message in the command session set is processed. Mapped to a fixed number statement object and mount statement-level metadata , is represented as:

[0037]

[0038] in, This represents the structure for metadata description of statements (i.e., the reference statement structure of RDF-star). They are subject, predicate, and object, respectively. For the sending time, A unique identifier for a single accusation message. The session identifier to which the accusation message belongs;

[0039] Create a message entity node for each accusation message. With traceability activity nodes And establish derived relationship edges based on message generation, sending and receiving relationships to form the source tracing layer;

[0040] Define the multi-layered relational knowledge graph at any time. Spectral structure volume The calculation formula is:

[0041]

[0042] in, The number of message entity nodes. This is the sum of the number of activity nodes and the number of derived edges. To determine the number of objects in the feature layer, These are the corresponding resource weights.

[0043] Constructing a multi-layered, interconnected knowledge graph aims to overcome the limitations of traditional graph models in simultaneously handling statement-level metadata and dynamic interaction history. By mapping messages to statement objects and attaching metadata using pre-defined templates, a refined description of instruction content is achieved. The tracing layer solidifies the dynamic behaviors of message generation, sending, and receiving into nodes and edges within the graph. The graph structure volume is defined as the weighted sum of nodes and edges at each layer. By mapping the abstract graph topology to computational resource consumption, weighting coefficients reflect the differentiated impact of different types of graph elements on storage space and query efficiency.

[0044] In a preferred embodiment, calculating the cumulative link connectivity time based on the electromagnetic environment link status data includes:

[0045] The electromagnetic environment link state data is constructed as a time-varying data. Changing binary on / off state function Defined as:

[0046]

[0047] in, This indicates that the link is reachable. This indicates that the link is refused.

[0048] For any charge message m in the command session set, based on the sending time of the charge message... Starting from the current moment, calculate... The cumulative connection time of the link is calculated using the following formula:

[0049]

[0050] Where u is the integration variable.

[0051] The cumulative connectivity duration of a statistical link is based on continuous modeling of the impact of the electromagnetic environment. In a highly contested environment, completing a full command interaction loop requires a continuous or cumulative time window to ensure reliable transmission of information and acknowledgments. Therefore, by performing cumulative calculations of the binary on / off state function in the time dimension, the effective availability of the communication link in the time domain can be captured.

[0052] In a preferred embodiment, when the cumulative connection time of the link meets a preset simulation element instantiation threshold, a scenario grounding operation is triggered, generating an execution element layer containing simulation entity objects and task parameter objects in the multi-layered association knowledge graph, including:

[0053] Set a threshold for instantiation of simulation elements Determine the unique grounding trigger moment of the charge message m. The calculation formula is:

[0054]

[0055] Where inf represents the time when the minimum lower bound of the condition is met;

[0056] At the unique ground trigger time Generate based on the preset instantiation template A simulation entity object and a task parameter object and create a grounding active node. Establish the following relationship:

[0057]

[0058]

[0059] in, Indicates usage relationship, Indicates a generating relation. Represents a statement object.

[0060] Setting an instantiation threshold for simulation elements and triggering grounding operations accordingly is to strictly couple the progression of simulation logic with the objective constraints of the communication environment. Specifically, only when the cumulative connection time of the link reaches a level sufficient to support the exchange of critical information can abstract instructions be transformed into concrete simulation entities and task parameters. Generating a fixed number of objects and establishing traceability relationships at this point ensures that each instantiation action leaves a definite structured record in the graph, thereby allowing delays or blockages caused by environmental constraints to be observed through changes in the graph structure.

[0061] In a preferred embodiment, the change in the size of the spectral primitives before and after each planned grounding operation is statistically analyzed, including:

[0062] Determine the trigger time of the desired grounding operation. Left limit moment With right limit moment ;

[0063] Calculate the change in the size of the spectrogram primitives corresponding to a single hypothetical grounding operation. The calculation formula is:

[0064]

[0065] in, This represents the spectral structure volume at the left-hand limit. Let ln denote the spectral structure volume at the right limit time, and ln denote the natural logarithm operation.

[0066] Using the logarithm of the ratio to calculate the magnitude of changes in graph primitive size is intended to eliminate the interference of differences in base size on the sensitivity of the index. Specifically, the structural impact caused by grounding operations is essentially a multiplicative rather than additive growth; that is, the jump in structural complexity is often proportional to the current graph size. By calculating the natural logarithm of the size ratio before and after the change, this nonlinear structural explosion can be transformed into a linear step value, thereby measuring the structural pressure experienced by sessions of different sizes at the moment of instantiation under a unified scale.

[0067] In a preferred embodiment, all the variation magnitudes within the command session set are aggregated and weighted according to the number of receiving nodes to generate a session tracing expansion index that characterizes the increase in the complexity of the graph structure, including:

[0068] Generate the session origination expansion index The calculation formula is:

[0069]

[0070] in, This represents the total number of planned grounding operations that occurred within the aforementioned command session set. This represents the magnitude of change in the size of the spectral primitives corresponding to the k-th hypothetical grounding operation. The number of receiving nodes is represented by ln, which represents the natural logarithm operation.

[0071] The session origination inflation metric is a normalized measure of structural impact throughout the entire lifecycle. By summing the changes in all grounding operations, it reflects the total structural cost throughout the session; weighting based on the number of receiving nodes is to remove the background complexity caused by the natural growth in the size of receiving nodes. This separates the inflation component induced by environmental adversarial factors, protocol retransmissions, and the graph weaving mechanism itself, ensuring that the metric only reflects the trend of increasing complexity at the structural level.

[0072] In a preferred embodiment, a complexity identifier node carrying the session tracing inflation index is established and associated with the multi-layered association knowledge graph. The output includes a knowledge graph file containing the complexity identifier node, a replayable command and control message sequence, and a simulation executable scenario package, including:

[0073] Create a complexity identifier node for each command session c in the command session set. and write it into the attribute vector. , is represented as:

[0074]

[0075] in, This refers to the session origination expansion index. To determine the total number of grounding operations, The number of receiving nodes. The graph structure volume after the k-th hypothetical grounding operation;

[0076] Establish With all grounded active nodes within this command session The member relationships are determined, and the knowledge graph file is generated.

[0077] Generate the replayable charge message sequence, ensuring the hash identifier of any message m. It is consistent with the identifier of the corresponding message entity node in the knowledge graph file;

[0078] The simulation executable scenario package is generated, and the simulation executable scenario package is composed of simulation entity objects in the execution element layer. Obtained by conversion with task parameter object;

[0079] Establishing complexity marker nodes and writing them back to the graph is to internalize the calculated structural risks as attributes of the data itself, enabling the graph to have self-describing capabilities. Requiring consistency in markers across the three types of output files ensures seamless interoperability between static graph descriptions, dynamic message replay sequences, and executable simulation scenarios via a unified hash key. Furthermore, during simulation execution or post-mortem analysis, the system can directly read and utilize pre-calculated bloat metrics for resource scheduling or performance optimization.

[0080] Object to be identified These are the various entities that need to be uniquely identified in the command and control message data, covering elements such as messages, statements, activities, and grounding artifacts. Their function is to provide a unified reference object for the interaction of heterogeneous systems in a distributed simulation environment, avoiding naming confusion. They are obtained from the input command and control message data. Extract all independent units that need to be tracked, associated, or stored, such as a single command and control message, a fragment of a message, or activity records during message transmission. For example, a command and control message containing operational orders, or a task description statement derived from that message, can both serve as objects to be identified. .

[0081] Hash identifier The object to be identified A unique digital identifier is used to ensure consistency in naming the same object across heterogeneous systems in a distributed environment, preventing association failures due to naming offsets. The calculation formula is... Among them, SHA256 is the preferred hash algorithm, which has the characteristics of high security and extremely low collision probability; Yes The normalized concatenation result, hex indicates that the output is a hexadecimal string. It is obtained by following a fixed order: |Type|Source System|Timestamp|Payload Hash| After normalization and concatenation, the hash is calculated using the SHA256 algorithm. For example, for an accusation message originating from Command Center A with a timestamp of 2024-05-20|14:30:00, after concatenation according to the rules and calculation of the SHA256 hash, the resulting 64-bit hexadecimal string is its hash. .

[0082] Standardized data concatenation For identifiable objects The attribute information is concatenated in an ordered manner according to a unified rule to generate a unique hash identifier. The foundation of this framework is to eliminate inconsistencies in hash values ​​caused by differences in the order of object attributes, ensuring that the same object generates the same identifier across different systems. The concatenation order is fixed as |Type|Source System|Timestamp|Payload Hash|, all attributes are UTF-8 encoded, and missing attributes are filled with fixed null values. The acquisition method is extraction. The data type (e.g., message, statement, activity), source system (e.g., command node, simulation node), timestamp (e.g., generation time, sending time), and payload hash (e.g., the hash value of the message content) are concatenated into a string in sequence. For example, a statement object... If the type is a task statement, the source system is tactical terminal B, the timestamp is 2024-05-2014:35:00, and the payload hash is abc123..., then... For mission statement | Tactical terminal B | 2024-05-20 14:35:00 | abc123....

[0083] SHA256 is used to generate objects to be identified. A cryptographic hash algorithm that uniquely identifies a hash identifier is used to concatenate normalized data of arbitrary length. Mapped to a fixed-length (256-bit) hash value, it possesses collision resistance and one-way hashing properties, ensuring the uniqueness and security of the identifier. SHA256 is the preferred hash value, which can be directly implemented using publicly available algorithms. The output is presented in hexadecimal (hex) format for easy storage and transmission. For example, using the standard SHA256 hash function library in a programming language, input... The corresponding string will output the corresponding hash value.

[0084] Hexadecimal output It is a hash identifier The data presentation format is used to convert the 256-bit binary hash value generated by the SHA256 algorithm into a 64-bit hexadecimal string. Its purpose is to simplify the storage, transmission, and visualization of hash identifiers, making it easier for humans and systems to recognize and process compared to the binary format. The acquisition method involves converting the binary result output by the SHA256 algorithm into hexadecimal encoding, which is a fixed format conversion step in the hash identifier generation process.

[0085] Accusation message data This includes all command instructions, status reports, and confirmation messages used to describe military simulation scenarios. The data format must conform to the C2SIM standard and include both a message header and a message body. Acquisition methods include real-time collection via a legitimate hardware interface or extraction from the simulation system's log files. The data must include key fields such as sender, receiver, MessageID, and ConversationID. For example, a command message from the command center to a combat unit ordering the capture of a high ground, or a report message from the combat unit confirming arrival at a designated area, both fall under this category. Components of.

[0086] Single accusation message It is the data of the accusation. The basic building block is an independent data block that carries specific command intentions or status information. Its function is to serve as the smallest data granularity for session segmentation, statement splitting, and source tracing. Each message has a unique MessageID and its corresponding ConversationID. It is obtained from command and control message data. The message is split and extracted based on MessageID. Each message must contain a message header (sender, receiver, MessageID, ConversationID, etc.) and a message body (the specific instruction or report content). For example, a message with MessageID MSG001, ConversationID CONV005, and content "Immediately activate air defense early warning" is a single command and control message. .

[0087] Session identifier It is a unique identifier used to identify the session to which a group of related command and control messages belong. It serves as the basis for dividing command and control sessions, and its function is to aggregate messages with causal, retransmission, or interactive relationships into independent sessions, achieving a localized measurement of structural expansion. It is obtained by extracting the message header fields from the command and control message data, ensuring that all messages within the same session have the same identifier. The identifiers for different sessions are unique and conform to the naming conventions of the C2SIM standard. For example, messages such as command transmissions, confirmations, and status feedback related to a specific raid mission between the command center and combat units all carry the same identifier. This forms an independent session.

[0088] Receiver list It contains a single accusation message. This collection of all recipient information records the target receiving nodes of the message. Its purpose is to determine the scale of the receiving nodes in the session, providing fundamental parameters for calculating the session scalability index. It is obtained by parsing the receiver field of the command and control message header, extracting the unique identifiers of all recipients (such as node names and device numbers), and forming a set of data. For example, if the recipients of a message are Combat Unit 1, Combat Unit 2, and Radar Station A, then its... .

[0089] Command Session Collection It is a session identifier The aggregated set of message groups, i.e. A single command session Its function is to isolate accusation messages by session dimension, enabling the measurement of structural expansion to accurately pinpoint the specific command intent chain. The acquisition method involves traversing the accusation message data. They will have the same All messages are aggregated into a single session, and multiple sessions together constitute a single session. For example, all The messages for CONV005 are aggregated into a session. ,all Messages for CONV008 are aggregated into a session. ,but .

[0090] Specific session identifier It is a single command session A unique identifier, similar to a session identifier. One-to-one correspondence is a key parameter for distinguishing different sessions. Its function is to assign a unique identifier to each aggregated session, facilitating individual session management, expansion metric calculation, and result traceability. It is obtained from the session identifier. Extracting from this, each session corresponds to a unique [database / database]. ,and The values ​​must remain consistent. For example, the session identifier. If it is CONV005, then it corresponds to a specific session identifier. .

[0091] Single command session It is a set of command sessions The basic building blocks are composed of the same The message group, comprising all associated accusation messages, serves as the basic unit for measuring structural expansion. All statement generation, source tracing, and grounding operations related to this session occur within this unit. It is obtained based on a specific session identifier. From the data of the allegations Related messages are filtered and aggregated to form independent session data units. For example, the CONV005 group, which contains multiple related messages such as command sending, acknowledgment confirmation, status feedback, and timeout retransmission, constitutes a single command session. .

[0092] Receiver node set It is a single command session The deduplicated set of all receivers in the middle, i.e. Its scale This is a crucial parameter in calculating the conversation origination expansion index. Its function is to reflect the coverage of conversations and provide a basis for the normalization of the expansion index. It is obtained by analyzing individual command conversations. List of all message receivers The union of the sets eliminates duplicate receivers, resulting in a unique set of receivers. For example, in a session... The list of recipients for the three messages are as follows: , , ,but .

[0093] The preset statement template library is a collection of fixed-number, fixed-format statement templates used to translate single command and control messages. The mapping is done to standardized statement objects, which convert messages into statements, ensuring a fixed number of statements and a uniform structure for each message. The acquisition method is based on the C2SIM message body schema design, covering four core predicate categories: object, behavior, time and space, and constraint. The number of templates can be determined through offline calibration, with an initial size of 12 being preferred. For example, the template library contains fixed-format templates such as "[Subject] executes [Task] at [Time]" and "[Subject]'s position is [Coordinates]".

[0094] Number of statement objects It is every accusation message The fixed number of statement objects generated by mapping reflects the size of the preset statement template library. Its purpose is to ensure that the number of statements after each message is transformed is consistent, avoiding process branches due to differences in message content and ensuring the determinism of the graph structure size calculation. The preferred value is 12 (which can be adjusted through offline calibration). This value is obtained by counting the number of valid templates in the preset statement template library and applying a coverage threshold. (Optimal 0.98) After trimming, the template is determined to ensure that it covers the core semantics of the vast majority of messages. For example, if the preset statement template library retains 12 valid templates after trimming, then 12 statement objects are generated for each message, and missing fields are filled with null value nodes.

[0095] Statement object It is a message of accusation. The standardized semantic units generated by mapping through the preset statement template library are in the form of ( ),in , , These are the subject, predicate, and object, respectively, providing the basic objects for mounting and grounding statement-level metadata. The method of acquisition is to send the message... The content is filled according to a preset statement template, and missing fields are filled with fixed null nodes. (IRI is ex:null) Replacement, generating a fixed number of RDF-star format statement objects. For example, after template mapping, the message "Message Combat Unit 1 occupies High Ground A at 15:00" generates statement objects.

[0096] Statement-level metadata It is mounted on the statement object The additional information set on the above, which is fixed to include Its function is to record key information such as the source, session, and generation time of the statement object. It is retrieved from the statement object. Corresponding accusations Extract relevant fields and attach them to the statement object in a fixed format to ensure a one-to-one correspondence between metadata and the statement object. For example, the statement object... The corresponding message ID is MSG001, ConversationID is CONV005, and the sending time is... If SenderID is Command Center A, then its .

[0097] Message Identifier It is a single accusation message The unique identifier is used to distinguish different message entities and serves as the basis for creating and associating message entity nodes. Its function is to achieve accurate message location and traceability, ensuring that each message corresponds to a unique entity node in the graph. It is obtained by parsing the message header fields of the command and control message, conforming to the C2SIM standard's UUID generation specification, ensuring that the identifiers of different messages are not duplicated.

[0098] Sending time It is a single accusation message The transmission time is a time parameter used for calculating the cumulative link connectivity duration and determining the grounding trigger time. Its function is to provide a benchmark for time-related calculations and process triggering, ensuring the accuracy of timing logic. It is obtained by parsing the timestamp field in the message header of the command and control message. If the timestamp is not included in the message, it is determined by mapping the message sequence number within the session to a uniform time grid, with the format uniformly set as year-month-day|hour:minute:second.

[0099] Message entity node In a knowledge graph, a single accusation message The created node object is identified as This is used to represent message entities in the graph. Its function is to serve as a concrete carrier of the message within the graph, providing associated targets for tracing activity nodes and establishing derived relationship edges. The acquisition method involves processing each accusation message... Generate a unique hash identifier And create corresponding nodes in the graph, with node attributes including , , Information such as messages. of In the graph, a node is created that uses this identifier as its unique identifier; this node is the message entity node. .

[0100] Source tracing activity nodes It is used to record a single accusation message. The graph nodes representing various activities throughout the message lifecycle (such as sending, receiving, retransmitting, and grounding) are part of the tracing layer. Their function is to solidify the dynamic interaction history of messages, provide node support for establishing derived relationship edges, and enable traceability of the message flow process. The acquisition method is based on the message itself. For each type of activity (such as sending activity, grounding activity), create a corresponding node. The node type is based on the preset PROV activity type set. Defined, identified as For example, for messages Sending activity creates node Create nodes for grounding activities .

[0101] Derivation edges are graph edges that connect message entity nodes, source activity nodes, statement objects, and other elements. They are used to represent the relationships between nodes, such as generation, usage, and association, as shown in the PROV standard. , The role of these elements is to construct the topological structure of the graph, reflecting the logical relationships between elements, and forming the basis of the tracing layer. They are obtained by creating relationships based on the actual relationships between nodes, according to the relationship types defined in the PROV-O standard, ensuring that the establishment of each relationship is supported by clear business logic (e.g., message entities are generated by sending activities, and grounding activities use statement objects). For example, message entity nodes... With sending activity nodes Establish between Derivative relationship edge.

[0102] The tracing layer is one of the core layers of a multi-layered relational knowledge graph. It consists of message entity nodes, tracing activity nodes, and derived relationship edges. Built based on the PROV-O standard's entity-activity-proxy model, its function is to record the entire lifecycle of a message, including generation, sending, receiving, retransmission, and grounding, enabling complete traceability of the message flow process. This is achieved by creating corresponding entity nodes and activity nodes for each message and establishing derived relationship edges based on the message flow logic, forming a tracing network covering the entire message lifecycle. For example, the entire process of a message being sent from the command center, received by the combat unit, generating a receipt, and triggering grounding is fully presented in the tracing layer through entity nodes, activity nodes, and derived relationship edges.

[0103] at any time It is a time variable used to calculate parameters such as the size of the graph structure and the cumulative connectivity duration of links, and can cover a predetermined time axis. All moments within this timeframe serve to provide a time reference for time-related calculations, supporting dynamic characterization of the graph structure and link states. This is obtained from a predetermined time axis. The selection can be made from discrete time points or continuous time intervals depending on the calculation requirements, and must be related to the message sending time. Grounding trigger time Time parameters should maintain a consistent timestamp format. For example, consider a defined timeline. If the time frame is 2024-05-2014:00:00-2024-05-2016:00:00, then at any given time... You can select 14:35:00, 15:10:00, etc. within this interval.

[0104] Atlas structure volume It represents any time interval Single command session The quantitative index corresponding to the size of the map is calculated using the following formula: Its function is to unify the scale of elements such as message entities, tracing activities, and grounding products in the measurement graph, providing a basis for calculating the magnitude of structural changes before and after grounding. As the benchmark, , , For offline-calibrated resource weights, , , These represent the number of message entities, the number of tracing activities and derived edges, and the number of grounded products at the corresponding time.

[0105] Number of message entity nodes It is any time Single command session The total number of message entity nodes already generated in the process, i.e. Its function is to form a core component of the graph structure, reflecting the cumulative scale of messages in a session. It is obtained by statistically analyzing individual command sessions. Sending time The number of message entity nodes corresponding to all messages is counted by deduplicating MessageID. For example, at time... Before, conversation If 5 messages have been sent, then .

[0106] Origin tracing activities and derived edge count It is any time Single command session The sum of the number of originating activity nodes and derived edges already generated in the graph is a crucial component of the graph structure, reflecting the cumulative scale of message interaction history within a session. It is obtained by statistically analyzing individual command sessions. Mid-time The total number of all source activity nodes and the number of derived edges are counted separately and then summed. For example, at time... Before, conversation There are 8 tracing activity nodes and 12 derived relationship edges in the middle, then .

[0107] Number of grounding products It is any time Single command session The total number of simulation entity objects, task parameter objects, and their derived statements already generated in the process, i.e. Its function is to form a component of the map structure, reflecting the scale of structural expansion caused by grounding operations. It is obtained by statistically analyzing individual command sessions. Grounding time All grounding artifacts, including those generated by each message. A simulated entity object and The derived statements are summed to obtain the result.

[0108] Resource weight These are the resource consumption weights corresponding to message entity nodes, tracing activities and derived edges, and grounded products. They are used to map structure counts to actual resource costs (such as memory and time consumption), and their function is to increase the graph structure volume. This approach reflects actual resource consumption and improves the engineering practicality of the inflation index. The optimal value is obtained through offline calibration, which involves collecting structure counts and measured resource metrics (such as memory and index size) from multiple batches of session samples. Least squares fitting is then used for calculation. .

[0109] The subject of the statement object ,predicate ,object It is a statement object The semantic components correspond to the subject, action, and object of the statement, respectively. Their function is to clarify the semantic connotation of the statement and realize the structured expression of the instruction content. The acquisition method is from the command and control message. Extract the corresponding components from the content, fill them according to the preset statement template, and use null nodes for missing components. Alternative. For example, if drone A conducts reconnaissance of area B, after mapping according to a template, , , .

[0110] Electromagnetic environment link state data is the raw data describing the reachability of tactical communication links. It contains information such as the link's on / off state and interference intensity at different times, and its function is to construct a binary on / off state function. This provides the basic data for calculating the cumulative link connectivity duration. The data can be obtained either by generating it using an electromagnetic environment simulation tool (considering factors such as terrain, weather, and interference) or by deducing it from the message logs using the send-receive timestamps. For example, the link status data for a certain time period might be: 14:00-14:05 reachable, 14:05-14:10 rejected, 14:10-14:15 reachable.

[0111] Binary on / off state function It is a time-series function characterizing link reachability, defined as ,in This indicates that the link is reachable (ON state). This indicates a link failure (OFF state). Its function is to simplify the complex electromagnetic environment link state into a computable model, supporting the integral calculation of the cumulative link connectivity duration. The data is obtained by binarizing the electromagnetic environment link state data, dividing it into time intervals, and determining the on / off state at each moment. A two-state update process can be used for modeling, and parameters are calibrated using historical data.

[0112] Integral variable This is a placeholder variable used to calculate the cumulative link connectivity time. In integration, it represents a continuous variable in the time dimension. Its purpose is to support the continuous-time integration calculation of the cumulative link connectivity time, ensuring that the integration result accurately reflects the cumulative duration of effective link connectivity. It is obtained through the standard variable definition for integration operations and is used only as an integration variable in the formula for calculating the cumulative link connectivity time.

[0113] The cumulative connection time of the link refers to the time from the accusation message Sending time up to the current moment The cumulative time the link remains reachable during this period is calculated using the following formula: Its function is to determine whether a message meets the grounding trigger condition, and it reflects the cumulative effective communication time of the link. It is obtained by using the message sending time. Starting from the point of integration, at the current moment As the endpoint of integration, for the binary on / off state function It is obtained by continuous-time integration.

[0114] Simulation element instantiation threshold This represents the minimum effective communication time required for the command and control messages to complete a closed-loop interaction. It is the critical value for triggering the hypothetical grounding operation. Its function is to ensure that the grounding operation is triggered only when the link has sufficient effective communication time to support the exchange of critical information, thus avoiding structural expansion distortion caused by link instability. The preferred value is obtained through offline calibration by statistically analyzing the cumulative reachability time samples of historical message sending and critical feedback. Take the preset signal level (Preferred quantile 0.95) .

[0115] Grounding trigger time It is a single accusation message Meets the instantiation threshold of simulation elements The first moment is calculated using the following formula: Its function is to determine the execution time of a planned grounding operation, ensuring the uniqueness and determinism of the grounding operation. It is obtained by real-time monitoring of the cumulative link connectivity time; when this time first reaches [a certain threshold]... When, record the corresponding moment. ,in This indicates the time when the minimum lower bound that satisfies the condition is reached.

[0116] Grounding products At the time of ground triggering The generated simulation entity objects and task parameter objects are the result of the transformation from abstract statements to executable simulation elements. Their function is to provide machine-readable entities and parameters for simulation execution, serving as the key carrier connecting the scenario description and simulation execution. The acquisition method involves generating a fixed number of [objects / parameters] for each trigger grounding message based on a preset instantiation template. Each grounded artifact contains the attributes and parameters required for simulation execution. For example, the generated Tank Company A simulation entity object and the 15:30 attack launch mission parameter object are both grounded artifacts. .

[0117] Number of grounding products It is each message that triggers grounding. The fixed number of generated simulation entity objects and task parameter objects serves to ensure that the number of products generated in each grounding operation is consistent, thus guaranteeing the quantity of grounding products. Determinism in computation. Optimal values ​​are obtained through offline calibration, specifically by statistically analyzing the average number of primitives corresponding to each command in historical executable scenario packages. Take the median and round up, that is The preferred value is 8.

[0118] Grounding active node It records a single message in the graph. The traceability node for grounding operations is a crucial component of the traceability layer. Its function is to solidify the execution record of grounding operations, providing node support for the association between grounding artifacts and statement objects. It is obtained through messages... Grounding trigger time Create the corresponding traceability activity node and identify it as... The node attributes include information such as grounding time and associated message identifier.

[0119] PROV relationship and Based on the W3CPROV-O standard, these are traceability relationships that represent the logical associations between entities used by activities and entities generated by activities. Their function is to establish traceability relationships between grounded activity nodes, statement objects, and grounded artifacts, thus improving the topology of the traceability layer. They are obtained by creating relationships based on the actual logical relationships between entities and activities, such as grounded activity nodes. With statement object Establish Relationship, grounding products and Establish relation.

[0120] Grounding trigger time It is a single command session The Middle The execution time of a planned grounding operation is determined by the grounding trigger times of all messages involved in that grounding operation. After deduplication, the determination is made, and its function is to identify the session. The timing of each grounding operation provides a time reference for calculating the structural volume of the graph before and after grounding. This is obtained through a session. Grounding trigger time for all messages After removing duplicates and sorting in ascending order, we get ,in For the session The total number of groundings.

[0121] Left limit moment It is the ground trigger moment The previous instantaneous moment, the right limit moment It is the ground trigger moment The subsequent instantaneous moment serves to define the time points before and after the grounding operation, ensuring that the impact of the grounding operation on the spectral structure can be quantified individually. This is obtained based on the grounding trigger moment. definition, infinitely close and less than , infinitely close And greater than In engineering implementation, this can be replaced by instantaneous counting before grounding and instantaneous counting after grounding.

[0122] Structural change range It is a single command session No. The logarithmic change in the structural volume of the graph before and after the grounding operation is calculated using the following formula: Its function is to quantify the structural expansion intensity caused by a single grounding operation, converting multiplicative growth into a linear step value for easier and standardized measurement. It is obtained by calculating the spectral structural volume at the left-hand limit time. and the right limit moment Calculate the natural logarithm of the ratio of the two according to the formula.

[0123] natural logarithm operation This is the mathematical operation used for data transformation in this method, mainly used to calculate the magnitude of structural changes. and conversation origination expansion indicators Its function is to transform the nonlinear multiplicative structural growth brought about by grounding operations into a linear additive step, eliminate the interference of the difference in base size on the sensitivity of the index, and make the expansion intensity of sessions of different sizes measurable under a unified scale.

[0124] Total number of grounding incidents It is a single command session The total number of times a planned grounding operation occurs during a session, i.e., the number of sessions. Set of grounding trigger times The number of elements is used to count the total number of grounding operations throughout the entire session's lifecycle, providing a basic parameter for calculating the session origination expansion index. This is obtained by analyzing the session... The total number of deduplicated ground trigger times is calculated by counting all the deduplicated ground trigger times. For example, a conversation If there are 3 different ground trigger times, then .

[0125] Number of receiving nodes It is a single command session Receiver node set The scale, that is Its function is to serve as a normalization factor for the conversation origination expansion index, stripping away the background complexity caused by the natural growth of the receiving node scale, so that the index only reflects the expansion trend at the structural level. It is obtained by analyzing the set of receiving nodes. Count the elements in the array, and the total number obtained is . .For example, ,but .

[0126] Session origination expansion index It represents a single command session The core indicator of the intensity of the spectral structure expansion induced by the grounding step is calculated using the following formula: Its function is to quantitatively assess the structural expansion risk of a session under grounding operations, providing a basis for resource planning and simulation optimization. The method of acquisition is to first accumulate the structural change magnitudes of all grounding operations within the session. Then divide by the normalization factor of the number of receiving nodes. get.

[0127] Complexity identifier node It carries a single command session The graph nodes for structural expansion information serve as carriers that internalize expansion indices into graph attributes. Their function is to endow the graph with self-describing capabilities, facilitating direct reading of structural expansion risk information by subsequent simulation systems and supporting resource scheduling and performance optimization. Acquisition is performed for each command session. Create an independent node, identified as and mount the attribute vector .

[0128] Attribute vector It is stored in the complexity identifier node. The structured attribute set in the data is in the form of Its function is to centrally store session expansion information, including expansion indicators, number of groundings, receiver size, and volume after each grounding, facilitating rapid querying and analysis. The acquisition method involves integrating session data in a fixed order. , , The graph structure volume after each grounding is formed into a vector and written into the complexity identifier node.

[0129] Membership relationships are indicators of connection complexity among nodes. With corresponding conversation All grounded active nodes The graph relationship is used to establish the correlation between expansion indicators and specific grounding operations, enabling traceability from expansion indicators to grounding activities and grounding products. The data is obtained according to a preset relationship type (e.g., ...). (or custom relationship), in With all in this session Establish connecting edges between them.

[0130] The knowledge graph file contains a conversation layer, a statement layer, a source tracing layer, an execution element layer, and complexity identifier nodes. The knowledge graph is stored in RDF-star+PROV-O format. Its purpose is to fully present the structured description of the military simulation scenario, including semantic information, source relationships, and structural expansion indicators, supporting cross-system interoperability and subsequent analysis. The acquisition method involves serializing the constructed multi-layered relational knowledge graph according to the RDF-star and PROV-O standard formats to generate a file (such as TTL format).

[0131] The replayable command and control message sequence is a set of messages that retains the original message structure and identifiers, corresponding one-to-one with the message entity node identifiers in the knowledge graph file. Its function is to support message flow replay during the simulation process, providing dynamic data support for simulation review and analysis, and ensuring consistency between the replay process and the graph description. It is obtained by extracting command and control message data. The core fields (including MessageID, ConversationID, sender, receiver, content, etc.) are used to maintain consistency between the message identifier and the message entity node identifier in the knowledge graph, and to generate a replayable sequence file.

[0132] The simulation executable scenario package consists of simulation entity objects in the execution element layer. The machine-readable file package obtained by converting task parameter objects conforms to NATOGSD's definition of ExecutableScenario. Its function is to directly interface with the simulation execution environment, providing executable scenario elements for simulation initialization, task distribution, and situational simulation. The acquisition method involves converting grounding products into a format supported by the simulation platform (such as XML or JSON) and integrating them into a standardized scenario package file.

[0133] like Figure 2 As shown, Figure 2The demonstration showcases the C2SIM server acting as a hub between the command and control system and the simulation system, considering electromagnetic interference or denial. The server receives command messages from the command and control system and link status data from electromagnetic environment / link status sources, then engages in bidirectional interaction with the simulation system cluster for initialization, control, command, and reporting. Furthermore, the C2SIM server is responsible for recording message streams to or replaying them from a message playback log library, and ultimately generating knowledge graph files and executable scenario packages, which are stored in the corresponding repositories.

[0134] It is important to note that all input data described in this solution is acquired in real-time through legal and compliant hardware interfaces with the user's full knowledge, explicit consent, and active cooperation. The preset parameters, prior constants, and statistical means are all derived from publicly available scientific literature data, de-identified general research datasets, or calibration data from laboratory environments, and do not contain any unauthorized sensitive third-party information. The system's data processing is limited to local or volatile memory computation transmitted via encrypted channels. There is no illegal collection, theft, or retention of user biometric data or infringement of user privacy without the user's knowledge. All parameter calls and generation comply with the principles of data minimization, legality, legitimacy, and necessity.

[0135] The embodiments of this example have been described above. However, this example is not limited to the specific implementation methods described above. The specific implementation methods described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms based on the guidance of this example, and all of them are within the protection scope of this example.

Claims

1. A method for describing military simulation scenarios based on knowledge graphs, characterized in that, include: Acquire command and control message data and electromagnetic environment link status data, and divide the command and control message data into command session sets according to the message header identifier; Construct a multi-layered association knowledge graph that maps the command session set. The multi-layered association knowledge graph includes a session layer that describes the session affiliation relationship, a statement layer that encapsulates instruction semantics and metadata, and a tracing layer that records message derivation and interaction history. Based on the electromagnetic environment link status data, the cumulative link connection time is counted. When the cumulative link connection time meets the preset simulation element instantiation threshold, the scenario grounding operation is triggered, and an execution element layer containing simulation entity objects and task parameter objects is generated in the multi-layer association knowledge graph. The changes in the size of the graph primitives before and after each planned grounding operation are statistically analyzed. All changes within the command session set are aggregated and weighted according to the number of receiving nodes to generate a session tracing expansion index characterizing the increase in graph structural complexity, including: Determine the left and right limit times of the proposed grounding operation trigger time, respectively; Calculate the spectral structure volume at the left limit time and the spectral structure volume at the right limit time, respectively. Calculate the ratio of the spectral structure volume at the right limit time to the spectral structure volume at the left limit time, and perform a natural logarithmic operation on the ratio to obtain the change range of the spectral primitive size corresponding to a single scenario grounding operation. The variation ranges of the graph element sizes corresponding to all hypothetical grounding operations within the command session set are summed. The number of receiving nodes is incremented by one, and then the natural logarithm is calculated. The result is used as a normalization factor. Divide the summation result by the normalization factor to obtain the session origination expansion index; The graph structure volume of the multi-layered knowledge graph at any time is defined as the weighted sum of the number of message entity nodes, the number of traceability activity nodes, the number of derived relationship edges, and the number of objects in the execution element layer. Establish a complexity identifier node that carries the session tracing expansion index and associate it with the multi-layer association knowledge graph. Output a knowledge graph file containing the complexity identifier node, a replayable command and control message sequence, and a simulation executable scenario package.

2. The military simulation scenario description method based on knowledge graphs according to claim 1, characterized in that, Acquire command and control message data and electromagnetic environment link status data, and divide the command and control message data into command session sets based on message header identifiers, including: For any object to be identified in the charge message data, normalized data concatenation is performed according to the order of type, source system, timestamp, and payload hash, and a hash algorithm is used to calculate the concatenation result to generate a hash identifier; Parse the message header fields of the charge message data to extract the session identifier and the list of receivers; The charge message data is aggregated into a command session set using the session identifier, and the union of all receiver lists in the command session set is taken to obtain a receiver node set.

3. The military simulation scenario description method based on knowledge graphs according to claim 1, characterized in that, A multi-layered relational knowledge graph mapping the command session set is constructed. This multi-layered relational knowledge graph includes a session layer describing session affiliation, a statement layer encapsulating instruction semantics and metadata, and a tracing layer recording message derivation and interaction history. Based on a preset statement template library, each command and control message in the command session set is mapped to a fixed number of statement objects, and each statement object is attached with statement-level metadata containing message identifier, session identifier and sending time; For each accusation message, a message entity node and a tracing activity node are established, and derived relationship edges are established based on the message generation, sending, and receiving relationships to form the tracing layer.

4. The military simulation scenario description method based on knowledge graphs according to claim 1, characterized in that, Based on the electromagnetic environment link status data, the cumulative link connectivity time is calculated, including: The electromagnetic environment link status data is constructed as a time-varying binary on / off state function, where on state indicates link reachability and off state indicates link rejection. For any command and control message in the command and control session set, the binary on / off state function is continuously integrated with the sending time of the command and control message as the starting point to obtain the cumulative link connection time of the command and control message at the current time.

5. The military simulation scenario description method based on knowledge graphs according to claim 4, characterized in that, When the cumulative connection time of the link meets the preset simulation element instantiation threshold, a scenario grounding operation is triggered, generating an execution element layer containing simulation entity objects and task parameter objects in the multi-layered association knowledge graph, including: Set a threshold for instantiation of simulation elements. The threshold for instantiation of simulation elements represents the minimum effective communication time required for the command and control messages to complete the closed-loop interaction. Monitor the cumulative connection time of the link, and determine the moment when the instantiation threshold of the simulation element is first reached as the grounding trigger moment of the accusation message; At the grounding trigger moment, a fixed number of simulation entity objects and task parameter objects are generated according to a preset instantiation template, and grounding activity nodes are created in the multi-layered association knowledge graph. The grounding activity nodes are then linked to statement objects in the statement layer and simulation entity objects in the execution element layer for reference and generation relationships.

6. The military simulation scenario description method based on knowledge graphs according to claim 1, characterized in that, Establish a complexity identifier node carrying the session tracing inflation index and associate it with the multi-layered association knowledge graph. Output a knowledge graph file containing the complexity identifier node, a replayable command and control message sequence, and a simulation executable scenario package, including: Create a complexity identifier node for each command session in the command session set, and write the session tracing expansion index, the total number of planned grounding operations that occur in the command session, the number of receiving nodes, and the graph structure volume sequence after each planned grounding operation as fixed attributes into the complexity identifier node. Establish the member association relationship between the complexity identifier node and all grounded active nodes in the command session, and embed the complexity identifier node into the multi-layer association knowledge graph; Generate the knowledge graph file, which includes the conversation layer, the statement layer, the source tracing layer, the execution element layer, and the complexity identifier node; Generate the replayable accusation message sequence, maintaining the consistency between the message identifier of each accusation message in the replayable accusation message sequence and the message entity node identifier in the knowledge graph file; The simulation executable scenario package is generated, which is obtained by converting the simulation entity object and the task parameter object in the execution element layer.

Images