An intelligent ship assembly construction data service call link tracking method

Abstract

The application relates to the technical field of shipbuilding, and discloses an intelligent ship assembly construction data service call link tracking method. The method comprises the following steps: acquiring ship assembly construction full-life cycle data sets, generating a service node topology graph and a load balancing path, performing abnormality detection processing on a task execution result, generating a link tracking feature matrix, further generating a service call link optimization strategy, and outputting a standardized service call link. The system comprises full-life cycle data acquisition, knowledge graph construction, path optimization, abnormality detection, multi-dimensional tracking and link generation modules. The application can effectively integrate data, optimize the service call link, realize load balancing and accurate abnormality detection, improve the efficiency, quality and stability of shipbuilding, reduce the construction cost, and meet the demand of the shipbuilding industry for digital and intelligent management.

Description

Technical Field

[0001] This invention relates to the field of shipbuilding technology, specifically to an intelligent method for tracing the call link of ship assembly and construction data services. Background Technology

[0002] During the ship assembly and construction process, the management and tracking of data service call links face many severe challenges, which seriously restrict the efficiency, quality and cost control of ship construction.

[0003] From a data management perspective, ship assembly and construction involves numerous pieces of equipment and complex processes, generating massive amounts of data—including equipment status parameters, service call logs, and task execution results—that are structurally complex and interconnected. Traditional data management methods cannot efficiently integrate and analyze this data, making it difficult to fully realize its value. For example, different equipment suppliers provide data in varying formats and interfaces, which can easily lead to compatibility issues during data aggregation, making it difficult to guarantee data consistency and accuracy, and posing significant challenges to subsequent analysis and decision-making.

[0004] Regarding service call chains, as shipbuilding becomes increasingly digitalized, the call relationships between various systems and services are becoming more complex. Due to a lack of effective link tracing methods, it is difficult to accurately pinpoint fault points and performance bottlenecks in the service call process. When a service malfunctions, it is impossible to quickly determine whether the problem stems from the service itself or a chain reaction triggered by other links in the upstream and downstream service call chains. This not only prolongs troubleshooting time but may also delay the entire construction schedule and increase unnecessary costs.

[0005] In terms of load balancing, existing load balancing strategies often cannot dynamically adjust based on real-time service call data. When some service nodes are overloaded while others are idle, the load cannot be distributed appropriately to idle nodes in a timely manner, leading to a decline in overall system performance. This load imbalance problem is particularly pronounced during peak shipbuilding seasons, when a large number of tasks are executed concurrently, severely impacting construction efficiency.

[0006] Furthermore, traditional methods for detecting anomalies in task execution results mainly rely on human experience or simple threshold judgments, resulting in poor accuracy and timeliness. They are unable to conduct comprehensive and in-depth analysis of complex anomalies, easily overlooking potential risks and failing to identify and resolve problems early, thus adversely affecting the quality of ship construction.

[0007] With the development of the global shipping industry, higher demands are being placed on the quality, efficiency, and cost control of shipbuilding. Traditional methods for managing data service call chains in ship assembly and construction can no longer meet the needs of the industry. There is an urgent need for an intelligent method and system to optimize data management, track service call chains, achieve efficient load balancing, and accurately detect anomalies, thereby enhancing the overall competitiveness of shipbuilding. Summary of the Invention

[0008] The purpose of this invention is to provide an intelligent method for tracing the call link of ship assembly and construction data services to solve the problems mentioned in the background art.

[0009] To achieve the above objectives, the present invention provides the following technical solution: an intelligent method for tracing the call link of ship assembly and construction data services, the method comprising:

[0010] Obtain a dataset representing the entire lifecycle of ship assembly and construction, including equipment status parameters, service call logs, and task execution results.

[0011] Based on the device status parameters, a service node topology graph is generated using a dynamic knowledge graph construction algorithm. The topology graph includes device relationships and service call path weights.

[0012] Based on the service call logs, a load balancing path is generated using a service path optimization algorithm, and redundant call nodes are removed.

[0013] Anomaly detection processing is performed on the task execution results to generate a service call anomaly marker sequence;

[0014] Input the service node topology map, load balancing path and anomaly marker sequence into the multi-dimensional tracing model to generate a link tracing feature matrix;

[0015] Based on the link tracing feature matrix, a service call link is generated through an adaptive tracing network construction algorithm; the nodes of the adaptive tracing network represent service call modules, and the edges represent call priority and stability weights.

[0016] Preferably, the step of generating a service node topology graph using a dynamic knowledge graph construction algorithm includes:

[0017] Entity relationship modeling is performed on the equipment status parameters to extract collaborative operation features between equipment;

[0018] Based on graph embedding algorithms, device entities are mapped to low-dimensional vectors to generate an initial knowledge graph;

[0019] Hierarchical clustering algorithm is used to group device entities and calculate the connection strength between groups;

[0020] The low-dimensional vector, inter-group connection strength, and service call path weight are encoded into a service node topology graph.

[0021] Preferably, the step of generating a load-balanced path through a service path optimization algorithm includes:

[0022] The service call logs are standardized and cleaned to remove invalid call records;

[0023] The algorithm identifies high-frequency service nodes based on dynamic load balancing and calculates the response time deviation between nodes. Based on the response time deviation and the preset stability threshold, a load balancing path is generated.

[0024] Redundant calling nodes are dynamically removed using a distributed hash table, and the load balancing path is updated.

[0025] Preferably, the anomaly detection process includes:

[0026] The task execution results are segmented in time sequence, and task completion indicators for each stage are extracted.

[0027] An isolated forest algorithm is used to detect abnormal segments and generate preliminary anomaly labels. The preliminary anomaly labels are then corrected based on a dynamic threshold adjustment algorithm to remove false detection labels.

[0028] The corrected anomaly markers are encoded into an anomaly marker sequence in chronological order.

[0029] Preferably, the multi-dimensional tracking model includes a feature alignment module and a matrix generation module, wherein the feature alignment module includes:

[0030] The device association relationships in the service node topology graph are normalized to obtain a first alignment vector; the call priorities in the load balancing path are discretized and mapped to generate a second alignment vector; and the anomaly marker sequence is subjected to window sliding statistics to extract anomaly frequency features to obtain a third alignment vector.

[0031] The first alignment vector, the second alignment vector, and the third alignment vector are fused into a link tracking feature matrix using a tensor decomposition algorithm.

[0032] Preferably, the update method of the dynamic knowledge graph construction algorithm includes:

[0033] Based on the sliding time window, the device state parameters are divided, and the changes in entity relationships within the window are extracted.

[0034] The device relationships and path weights in the knowledge graph are updated through incremental learning algorithms, and the structure of the service node topology is dynamically adjusted based on the updated knowledge graph.

[0035] Preferably, the method for adjusting the parameters of the distributed hash table includes:

[0036] Set the initial number of hash shards based on the number of service call logs;

[0037] By dynamically adjusting the shard distribution using the consistent hashing algorithm, the amount of data migration during node expansion is reduced;

[0038] Optimize hash function parameters based on sharded load monitoring results to improve the efficiency of redundant node removal.

[0039] Preferably, the adaptive tracking network construction algorithm includes:

[0040] Initialize node attributes based on the service call module, and generate edge weight tensors based on stability weights;

[0041] The link tracing feature matrix is ​​used as node input, and the edge weight tensor consists of call priority and stability weights.

[0042] The path selection strategy of each node is iteratively updated through a strategy optimization algorithm to optimize the edge weight tensor.

[0043] Generate the optimal service call chain covering all nodes based on the optimized edge weight tensor.

[0044] Preferably, the present invention also includes an intelligent ship assembly and construction data service call link tracking system, the system comprising:

[0045] Full lifecycle data acquisition module: used to acquire the full lifecycle dataset of ship assembly and construction, which includes equipment status parameters, service call logs and task execution results;

[0046] Knowledge graph construction module: configured to generate a service node topology graph based on the device status parameters using a dynamic knowledge graph construction algorithm;

[0047] Path optimization module: used to generate load-balanced paths based on the service call logs using a service path optimization algorithm;

[0048] Anomaly detection module: Performs anomaly detection processing on the task execution results and generates a service call anomaly marker sequence;

[0049] Multi-dimensional tracking module: Input the service node topology map, load balancing path and anomaly marker sequence into the multi-dimensional tracking model to generate a link tracking feature matrix;

[0050] Link generation module: used to generate service call links based on the link tracing feature matrix and through an adaptive tracing network construction algorithm.

[0051] Compared with the prior art, the beneficial effects of the present invention are:

[0052] In terms of data integration and analysis, by acquiring a dataset covering the entire lifecycle of ship assembly and construction, and utilizing a dynamic knowledge graph construction algorithm to generate a service node topology diagram, complex equipment status parameters can be effectively integrated, clearly presenting equipment relationships and service call path weights. This helps engineers gain a comprehensive understanding of the collaborative relationships between various devices during ship construction, providing strong support for subsequent troubleshooting and performance optimization. For example, when equipment malfunctions, other related equipment can be quickly identified based on the topology diagram, narrowing the scope of investigation and improving troubleshooting efficiency.

[0053] In service call chain optimization, a load-balanced path is generated based on service call logs using a service path optimization algorithm, and redundant call nodes are eliminated, significantly improving system performance and efficiency. The dynamic load balancing algorithm can identify high-frequency service nodes in real time and rationally allocate load based on response time deviations and preset stability thresholds, avoiding situations where some nodes are overloaded while others are idle. Taking the large amount of data processing tasks during shipbuilding as an example, the generation of load-balanced paths ensures that each data processing service node can bear the tasks evenly, greatly improving the overall system response speed, reducing task waiting time, and thus accelerating the shipbuilding progress.

[0054] In terms of anomaly detection and handling, advanced isolated forest and dynamic threshold adjustment algorithms are used to detect anomalies in task execution results, generating accurate anomaly marker sequences. This approach can promptly identify anomalies during task execution, and the dynamic threshold adjustment effectively eliminates false detection markers, improving the accuracy of anomaly detection. In shipbuilding, timely detection and handling of anomalies in critical tasks can prevent problems from escalating and ensure the quality of ship construction. For example, in ship structural welding tasks, if abnormal welding parameters are detected in a timely manner through anomaly detection, the welding process can be adjusted promptly to avoid welding defects and improve the safety of the ship structure.

[0055] The introduction of a multi-dimensional tracing model integrates service node topology, load balancing paths, and anomaly marker sequences to generate a link tracing feature matrix, enabling comprehensive tracking and analysis of service call links from multiple perspectives. Through the collaborative work of the feature alignment module and the matrix generation module, rich feature information can be extracted, providing more accurate data support for subsequent link optimization.

[0056] Finally, based on the link tracing feature matrix, an adaptive tracing network construction algorithm is used to generate service call links. This optimization strategy can dynamically adjust the service call links according to real-time conditions, improving the stability and reliability of the system. In the shipbuilding process, facing constantly changing demands and environments, standardized service call links can ensure smoother collaboration between various stages, reduce the probability of system errors, improve the overall efficiency and quality of shipbuilding, and also help reduce construction costs, enhancing the competitiveness of shipbuilding companies in the market. Attached Figure Description

[0057] Figure 1 This is a schematic diagram illustrating the working principle of the big data-based production scheduling visualization method described in this invention.

[0058] Figure 2 A diagram illustrating the working principle of building a service node topology graph for dynamic knowledge graphs;

[0059] Figure 3 This is a flowchart for handling anomaly detection in task execution results. Detailed Implementation

[0060] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0061] Please see Figures 1-3 This invention provides a technical solution: This invention relates to an intelligent method for tracing the call link of ship assembly and construction data services, and the specific implementation scheme is as follows:

[0062] During the ship assembly and construction process, various sensors, monitoring systems, and related software are used to record and collect equipment status parameters, service call logs, and task execution results, forming a full lifecycle dataset for ship assembly and construction. For example, equipment status parameters may include real-time data such as operating temperature, speed, and pressure; service call logs record information such as the time of each service call, the caller, and the callee; and task execution results reflect the completion status of each construction task, such as whether it was completed, the completion time, and the quality of completion.

[0063] Based on the acquired equipment status parameters, a dynamic knowledge graph construction algorithm is used to generate a service node topology graph. This topology graph includes equipment relationships and service call path weights, demonstrating the collaborative operation relationships between various equipment and the importance of service call paths during the ship assembly and construction process.

[0064] Based on service call logs, a load-balanced path is generated using a service path optimization algorithm. Simultaneously, the algorithm identifies and eliminates redundant call nodes, improving service call efficiency and performance while reducing system resource waste.

[0065] Anomaly detection and handling are performed on task execution results to promptly identify abnormal situations during task execution. The detection results are then used to generate a service call anomaly marker sequence for subsequent analysis and handling of anomalies.

[0066] The service node topology, load balancing path, and anomaly marker sequence are input into the multi-dimensional tracking model. This model processes and fuses these data to generate a link tracking feature matrix, providing data support for the subsequent generation of optimization strategies.

[0067] Based on the link tracing feature matrix, an adaptive tracing network construction algorithm is used to generate service call links.

[0068] The technical solution of the present invention will be further described in detail below through specific embodiments:

[0069] Example 1:

[0070] This embodiment focuses on the specific implementation process of generating a service node topology graph using a dynamic knowledge graph construction algorithm. In the shipbuilding process, there are numerous and complex pieces of equipment. Taking a large shipbuilding project as an example, it involves various types of equipment such as power equipment, welding equipment, and hoisting equipment. After obtaining the state parameters of these devices, entity relationship modeling is first performed. Assuming that equipment A, B, and C are different equipment entities, analysis of the equipment state parameters reveals that temperature changes in equipment A during operation affect the working efficiency of equipment B. This indicates that there are collaborative operation characteristics between equipment A and equipment B, and this relationship is extracted and recorded.

[0071] Device entities are mapped to low-dimensional vectors using a graph embedding algorithm. The graph embedding algorithm used here can be the DeepWalk algorithm. Let the set of device entities be... ,in Indicates the first Each device entity is represented by a DeepWalk algorithm. Mapped to a low-dimensional vector This allows the characteristics of relationships between device entities to be preserved in a low-dimensional space, generating an initial knowledge graph.

[0072] Next, a hierarchical clustering algorithm is used to group the device entities. Let the distance metric function be... It represents the physical device. and The distance between them. Using a hierarchical clustering algorithm, device entities are progressively merged into different groups based on a distance metric function. Calculate the inter-group connection strength, assuming group and The connection strength between them is It can be determined by calculating the number of connections or the sum of connection weights between device entities in the two groups.

[0073] Finally, the low-dimensional vector Intergroup connection strength and service call path weight (in Indicates from device To the equipment The service call path weights are encoded into a service node topology graph. This clearly displays the relationships between devices and the weight information of service call paths, providing an important basis for subsequent analysis and optimization.

[0074] Example 2:

[0075] In shipbuilding and assembly scenarios, service call logs record a large amount of complex data, which is crucial for generating load-balanced paths. Initially, the service call logs need to be standardized and cleaned. During actual shipbuilding, factors such as network fluctuations and transient system failures can generate many invalid call records. For example, during a brief network outage, some service call requests may fail to reach the target server, but still leave call records in the logs without corresponding service response data; these records are invalid. During cleansing, the response field in the log records is checked. If the response field is empty and the call time is less than the minimum normal response time, the record is considered invalid and removed.

[0076] After cleaning, high-frequency service nodes are identified based on a dynamic load balancing algorithm. Assume the set of service nodes is... ,in Representing the Each service node is analyzed statistically, and the number of calls to each service node is recorded. Set a high-frequency threshold. This threshold can be determined based on past construction project data and the actual situation of the current project. For example, in similar past projects, a performance bottleneck may occur when a certain service node is called more than 100 times per unit time. Considering the scale and complexity of the current project, the high-frequency threshold can be adjusted accordingly. Set to 120 times / hour. When a certain service node... Number of calls Greater than At that time, the node is identified as a high-frequency service node.

[0077] Next, the response time deviation between nodes is calculated. Let the nodes be... and The response times between them are respectively and Response time deviation is expressed by the formula To calculate the response time here. This refers to the time interval from when a service request is sent to when a service response is received, which can be precisely obtained from the service call log. A preset stability threshold is set. For example, based on system performance requirements and past experience, Set to 50 milliseconds. When This means that the load between the two nodes is unbalanced, and the service call path needs to be adjusted to achieve load balancing. For example, if node and If the response time deviation is greater than the stability threshold, it is advisable to consider rerouting part of the original shipments to [the relevant destination]. The request was assigned to or to Expand the service resources.

[0078] Finally, redundant calling nodes are dynamically removed using a distributed hash table. A distributed hash table (DHT) is a distributed storage system that stores data across multiple nodes and uses a hash function to determine the data's storage location. In this embodiment, when a calling node has an extremely low call frequency over a period of time, and evaluation reveals that its contribution to the overall service call chain is minimal, this node can be identified as redundant. Utilizing the characteristics of a distributed hash table, based on the node's unique identifier, a hash function is used to dynamically remove redundant calling nodes. The position of the node in the hash table is calculated, and then it is deleted from the hash table, thereby realizing the dynamic removal of redundant calling nodes and timely updating the load balancing path to ensure that the service call chain always maintains an efficient operating state.

[0079] Example 3:

[0080] In the execution of ship assembly and construction tasks, the task execution results are time-varying sequence data, making accurate anomaly detection and handling crucial. First, the task execution results are segmented into time sequences. Assume the sequence of task execution results is as follows: At fixed time intervals This serves as the basis for segmentation. In ship welding tasks, a fixed time interval of 10 minutes is used. The entire welding task execution process is divided into multiple time periods. Extract task completion metrics for each time period.

[0081] After extracting the task completion metrics, the Isolation Forest algorithm is used to detect abnormal segments. The Isolation Forest algorithm isolates data points by constructing a binary tree. Let the set of data points be... , here This represents the task completion rate data points for each time period. For each data point... Calculate its path length in the isolated forest. In an isolated forest, path length The shorter the path length, the more the data point deviates from the normal data distribution, and therefore the more likely it is to be an outlier. A path length threshold should be set. This threshold can be obtained through statistical analysis of historical normal task completion data. For example, by constructing an isolated forest model based on completion data from a large number of normal welding task time periods, the average and standard deviation of the path length of normal data points can be calculated. The average minus twice the standard deviation can then be used as the path length threshold. .when When the time period corresponding to the data point is an anomaly, a preliminary anomaly marker is generated.

[0082] Since initial anomaly labeling may result in false detections, it needs to be corrected using a dynamic threshold adjustment algorithm. Assume the dynamic threshold adjustment factor is... This factor can be dynamically adjusted based on the real-time status of task execution and the changing trends of historical data. For example, if significant fluctuations in recent task completion data are observed, it can be appropriately increased. The value of makes the threshold more stringent; when the data is relatively stable, decrease . The value of . According to the formula Dynamically adjust path length threshold for Then, the initial anomaly markers are reassessed, and false positives are removed.

[0083] Finally, the corrected anomaly markers are encoded into an anomaly marker sequence in chronological order. For example, after correction, within a time period... and If an anomaly is detected, the anomaly marker sequence can be represented as follows: The sequence of anomaly markers, where 1 indicates an anomaly and 0 indicates normal operation, provides a clear and accurate data foundation for in-depth analysis and timely handling of anomalies in the ship assembly and construction process, thus helping to ensure the smooth progress of the entire construction task.

[0084] Example 4:

[0085] This embodiment delves into the specific implementation of the multi-dimensional tracking model. During the ship assembly and construction data processing, the equipment relationships in the service node topology graph are normalized. Let the equipment relationship matrix be... ,in Indicates device and equipment The correlation values ​​between them ( (Through normalization formula) ,Will Normalization to The interval is used to obtain the first alignment vector.

[0086] Discretize and map the call priorities in the load balancing path. Assume the call priority values ​​range as follows: Discretize it into arrive The values ​​between these ranges. For example, priority 1 maps to 0, priority 5 maps to 1, priority 3 maps to 0.5, etc., generating a second alignment vector.

[0087] Window sliding statistics are performed on the anomaly-labeled sequence to extract anomaly frequency features. Let the anomaly-labeled sequence be... Window size is Count the number of anomalies marked as 1 within each window. abnormal frequency This yields the third alignment vector.

[0088] The first alignment vector, second alignment vector, and third alignment vector are fused into a link tracking feature matrix using a tensor decomposition algorithm. Let the first alignment vector be... The second alignment vector is The third alignment vector is Through tensor decomposition algorithm The link tracking feature matrix is ​​obtained. .

[0089] Example 5:

[0090] This embodiment details the update method of the dynamic knowledge graph construction algorithm, the parameter adjustment method of the distributed hash table, and the adaptive tracking network construction algorithm. During the ship assembly and construction process, equipment state parameters change continuously over time. Equipment state parameters are divided based on a sliding time window, with the sliding time window size set as... Within each time window, extract the changes in entity relationships for the device status parameters. For example, at a certain moment, a new device D is installed, establishing a new collaborative working relationship with devices A and B; this is the change in entity relationships.

[0091] The device associations and path weights in the knowledge graph are updated using an incremental learning algorithm. Let the incremental learning algorithm be... Knowledge graph is The device association is The path weight is Based on the changes in entity relationships, through Update the knowledge graph, in which and These represent the changes in device relationships and path weights, respectively. The structure of the service node topology is dynamically adjusted based on the updated knowledge graph to ensure that the topology reflects changes in relationships between devices in real time.

[0092] For parameter adjustments in the distributed hash table, the initial number of hash shards is set based on the number of service call logs. Let the number of service call logs be... The initial number of hash shards is It can be done through formula (This is just an example formula; a more suitable formula can be determined based on experience or testing.) Determine the initial number of hash shards.

[0093] By dynamically adjusting the shard distribution using a consistent hashing algorithm, the amount of data migration during node expansion can be reduced. Let the consistent hashing algorithm be... The hash ring is When a new node joins or an existing node needs to be expanded, through Dynamically adjust the hash ring to achieve a reasonable distribution of shards.

[0094] Optimize hash function parameters based on sharding load monitoring results to improve the efficiency of redundant node removal. Let the sharding load be... ( (representing the shard number), the hash function is Through monitoring According to the optimization algorithm Adjust the hash function parameters to improve the accuracy and efficiency of redundant node removal.

[0095] In the adaptive tracking network construction algorithm, node attributes are initialized according to the service invocation modules, and edge weight tensors are generated based on stability weights. Let the set of service invocation modules be... Each service call module For a given node, the node attributes are: The stability weight is... (Indicates from node) To the node (Stability weights), generating edge weight tensors .

[0096] The link tracing feature matrix is ​​used as node input, and the edge weight tensor consists of call priority and stability weights. The path selection strategy of each node is iteratively updated using a policy optimization algorithm to optimize the edge weight tensor. Let the policy optimization algorithm be... After multiple iterations ,in This represents input data such as the link tracking feature matrix.

[0097] The optimal service call chain covering all nodes is generated based on the optimized edge weight tensor, thereby optimizing the service call chain for ship assembly and construction data and improving the efficiency and reliability of the entire construction process.

[0098] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0099] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An intelligent method for tracing the data service call link in ship assembly and construction, characterized in that, include: Obtain a dataset representing the entire lifecycle of ship assembly and construction, including equipment status parameters, service call logs, and task execution results. Based on the device status parameters, a service node topology graph is generated using a dynamic knowledge graph construction algorithm. The topology graph includes device relationships and service call path weights. Based on the service call logs, a load balancing path is generated using a service path optimization algorithm, and redundant call nodes are removed. Anomaly detection processing is performed on the task execution results to generate a service call anomaly marker sequence; The service node topology, load balancing path, and anomaly marker sequence are input into the multi-dimensional tracing model to generate a link tracing feature matrix, including: The device association relationships in the service node topology graph are normalized to obtain a first alignment vector; the call priorities in the load balancing path are discretized and mapped to generate a second alignment vector; and the anomaly marker sequence is subjected to window sliding statistics to extract anomaly frequency features to obtain a third alignment vector. The first alignment vector, the second alignment vector, and the third alignment vector are fused into a link tracking feature matrix using a tensor decomposition algorithm. Based on the link tracing feature matrix, a service call link is generated through an adaptive tracing network construction algorithm; the nodes of the adaptive tracing network represent service call modules, and the edges represent call priority and stability weights. The adaptive tracking network construction algorithm includes: The node attributes are initialized according to the service call module, and the link tracing feature matrix is ​​used as the node input. The edge weight tensor is composed of call priority and stability weight. The path selection strategy of each node is iteratively updated through a strategy optimization algorithm to optimize the edge weight tensor. Generate the optimal service call chain covering all nodes based on the optimized edge weight tensor.

2. The intelligent ship assembly and construction data service call link tracing method according to claim 1, characterized in that, The generation of the service node topology graph through the dynamic knowledge graph construction algorithm includes: Entity relationship modeling is performed on the equipment status parameters to extract collaborative operation features between equipment; Based on graph embedding algorithms, device entities are mapped to low-dimensional vectors to generate an initial knowledge graph; Hierarchical clustering algorithm is used to group device entities and calculate the connection strength between groups; The low-dimensional vector, inter-group connection strength, and service call path weight are encoded into a service node topology graph.

3. The intelligent ship assembly and construction data service call link tracing method according to claim 1, characterized in that, The process of generating load-balanced paths through a service path optimization algorithm includes: The service call logs are standardized and cleaned to remove invalid call records; The algorithm identifies high-frequency service nodes based on dynamic load balancing and calculates the response time deviation between nodes. Based on the response time deviation and the preset stability threshold, a load balancing path is generated. Redundant calling nodes are dynamically removed using a distributed hash table, and the load balancing path is updated.

4. The intelligent ship assembly and construction data service call link tracing method according to claim 1, characterized in that, The anomaly detection process includes: The task execution results are segmented in time sequence, and task completion indicators for each stage are extracted. An isolated forest algorithm is used to detect abnormal segments and generate preliminary anomaly labels. The preliminary anomaly labels are then corrected based on a dynamic threshold adjustment algorithm to remove false detection labels. The corrected anomaly markers are encoded into an anomaly marker sequence in chronological order.

5. The intelligent ship assembly and construction data service call link tracing method according to claim 2, characterized in that, The update method of the dynamic knowledge graph construction algorithm includes: Based on the sliding time window, the device state parameters are divided, and the changes in entity relationships within the window are extracted. The device relationships and path weights in the knowledge graph are updated through incremental learning algorithms, and the structure of the service node topology is dynamically adjusted based on the updated knowledge graph.

6. The intelligent ship assembly and construction data service call link tracing method according to claim 3, characterized in that, The parameter adjustment method for the distributed hash table includes: Set the initial number of hash shards based on the number of service call logs; By dynamically adjusting the shard distribution using the consistent hashing algorithm, the amount of data migration during node expansion can be reduced. Optimize hash function parameters based on sharded load monitoring results to improve the efficiency of redundant node removal.

7. An intelligent ship assembly and construction data service call link tracing system, used to implement the intelligent ship assembly and construction data service call link tracing method as described in any one of claims 1-6, characterized in that, include: Full lifecycle data acquisition module: used to acquire the full lifecycle dataset of ship assembly and construction, which includes equipment status parameters, service call logs and task execution results; Knowledge graph construction module: configured to generate a service node topology graph based on the device status parameters using a dynamic knowledge graph construction algorithm; Path optimization module: used to generate load-balanced paths based on the service call logs using a service path optimization algorithm; Anomaly detection module: Performs anomaly detection processing on the task execution results and generates a service call anomaly marker sequence; Multi-dimensional tracking module: Input the service node topology map, load balancing path and anomaly marker sequence into the multi-dimensional tracking model to generate a link tracking feature matrix; Link generation module: used to generate service call links based on the link tracing feature matrix and through an adaptive tracing network construction algorithm.

Images