WMS software development assisting method and system based on artificial intelligence

Abstract

The invention discloses a WMS software development assisting method and system based on artificial intelligence, and relates to the technical field of logistics management. The method comprises the following steps: constructing a federated rule knowledge graph, generating an initial code snippet based on causal reasoning, converting a natural language demand into a code containing conditional logic by utilizing a causal chain mined from the federated rule knowledge graph, performing rule compliance judgment and processing on the initial code snippet, and obtaining the natural language demand. Determining to enter cross-system collaborative code generation or return to regenerate codes according to a judgment result; and generating a cross-system collaborative code and performing causal consistency verification. According to the method, the aviation logistics WMS development efficiency, the adaptation precision and the system reliability are remarkably improved, code redundancy is reduced, the adaptation capacity to complex scenes is enhanced, and the requirements for high compliance and high reliability of aviation logistics are met.

Description

Technical Field

[0001] This invention relates to the field of logistics management technology, specifically to an artificial intelligence-based WMS software development assistance method and system. Background Technology

[0002] In the modern air logistics sector, Warehouse Management Systems (WMS), as core operational support systems, need to be strictly adapted to the unique regulatory framework of air logistics, including the International Air Transport Association (IATA) Dangerous Goods Transport Rules, customs cross-border supervision policies, and special warehousing requirements for various types of goods. These rules are not only highly dynamic and frequently updated, but also involve multiple stakeholders such as airports, customs, freight forwarders, and e-commerce platforms. Each stakeholder needs to share rules to achieve collaborative operations while also protecting the privacy and security of their own proprietary rules. This places extremely high demands on the rule integration and adaptation capabilities of WMS software development.

[0003] In existing technologies, WMS software development often relies on centralized rule bases or manual rule interpretation, making it difficult to achieve efficient rule collaboration and dynamic updates while protecting the privacy of multiple stakeholders. Centralized rule bases either suffer from incomplete rule coverage due to the inability to access private rules of stakeholders, or pose a privacy risk by collecting original private rules. Consequently, the developed WMS is difficult to adapt to the complex scenario of multi-stakeholder collaboration in air logistics, resulting in delayed rule updates and low collaboration efficiency. Summary of the Invention

[0004] The purpose of this invention is to provide an AI-based WMS software development assistance method and system to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a WMS software development assistance method based on artificial intelligence, comprising the following steps:

[0006] A federated rule knowledge graph is constructed. The federated rule knowledge graph adopts a distributed architecture. Multiple rule nodes store public rules in the field of aviation logistics and private rules of each participating entity. The nodes achieve secure exchange and dynamic updating of rule parameters through a federated learning protocol, providing a rule basis for subsequent code generation.

[0007] Initial code snippets are generated based on causal reasoning. Using causal chains mined from the federated rule knowledge graph, natural language requirements are converted into code containing conditional logic. The causal chains include cause, effect, solution, and moderating variables.

[0008] The initial code snippet is subjected to rule compliance judgment and processing, and the decision is made based on the judgment result to proceed to cross-system collaborative code generation or return to regenerate the code;

[0009] Generate cross-system collaborative code and perform causal consistency verification to achieve interface adaptation between WMS and external systems, and ensure that the code logic is consistent with the causal chain through virtual scenario testing.

[0010] Preferably, the construction of the federated rule knowledge graph adopts a distributed architecture, with multiple rule nodes storing public rules in the aviation logistics field and private rules of each participating entity. Nodes securely exchange and dynamically update rule parameters through a federated learning protocol, providing a rule basis for subsequent code generation, including:

[0011] Deploy distributed rule nodes, which include a central node for storing public rules and distributed nodes for storing private rules of aviation logistics participants. The nodes establish encrypted communication links through a federated learning protocol to achieve secure exchange of rule parameters without transmitting raw data.

[0012] By integrating explicit and implicit rules, structured explicit rules are extracted from authoritative data sources using natural language processing technology, and empirical implicit rules are mined from historical operation logs based on reinforcement learning models. After multi-subject distributed node collaborative verification, these rules are incorporated into the federated rule knowledge graph.

[0013] The system performs rule conflict arbitration and self-evolution. When multiple parties' rules conflict, it calls the aviation scenario weight model to calculate weight coefficients based on cargo type and flight attributes and outputs the optimal coordination solution. At the same time, it automatically optimizes inefficient rules based on rule survival index, which includes compliance contribution and efficiency improvement value.

[0014] Preferably, the federated learning protocol uses homomorphic encryption technology to encrypt the rule parameters. The participants in the air logistics include airports, customs, freight forwarders, and e-commerce platforms. The public rules include the general terms of the International Air Transport Association (IATA), and the private rules include customs inspection standards and the special warehousing requirements of e-commerce platforms.

[0015] Preferably, the initial code snippet generated based on causal reasoning utilizes causal chains mined from the federated rule knowledge graph to convert natural language requirements into code containing conditional logic. The causal chains include causes, effects, solutions, and moderating variables, including:

[0016] Construct a causal rule graph, add a causal relationship layer on the basis of the federated rule knowledge graph, and use causal reasoning algorithms to mine the causal chain behind the rules to clarify the preconditions and exceptions for the application of the rules;

[0017] The natural language requirements are analyzed, the core entities in the requirements are identified through the natural language processing model, and the constraints are analyzed in combination with the causal rule graph to generate a structured list of requirements containing causal logic descriptions.

[0018] The system generates condition-adaptive code by calling an aviation scenario-specific template library based on the structured requirements list. This generates the initial code snippet containing prerequisite judgments and corresponding operational logic. The aviation scenario-specific template library includes sub-templates for dangerous goods management, cross-border supervision, and cold chain transportation.

[0019] Preferably, the step of performing rule compliance judgment and processing on the initial code fragment, and deciding whether to proceed to cross-system collaborative code generation or return to regenerate code based on the judgment result, includes:

[0020] Perform rule compliance verification, extract key operation logic from the initial code snippet, compare it with the latest rules in the federated rule knowledge graph, and calculate a compliance score;

[0021] If the compliance score is ≥95, proceed to the stage of generating cross-system collaborative code and performing causal consistency verification; if the compliance score is <95, return to the stage of generating conditional adaptive code, adjust parameters according to the violation prompt, and regenerate the code.

[0022] Preferably, the step of generating cross-system collaborative code and performing causal consistency verification to achieve interface adaptation between WMS and external systems, and ensuring that the code logic is consistent with the causal chain through virtual scenario testing, includes:

[0023] Construct a system collaboration knowledge graph to record the interface protocols, data formats, and collaboration logic between WMS and external systems, including airport IoT devices, customs supervision systems, and e-commerce platform order systems.

[0024] Generate cross-system interface code. Based on the system's collaborative knowledge graph and the code that passes compliance judgment, automatically generate interface adaptation code to achieve seamless integration between WMS and the external system.

[0025] Perform causal consistency verification by constructing a virtual aviation warehouse scenario based on digital twin technology, simulating extreme scenarios to test code logic, tracing the consistency between the code logic and the causal chain, outputting a causal consistency report, and automatically fixing non-compliant code.

[0026] Preferably, the virtual aviation warehouse scenario includes a tarmac, a hazardous materials warehouse, and a cold chain area; the extreme scenarios include heavy rain and flight cancellations; and the causal consistency verification not only verifies whether the code conforms to the rules, but also verifies whether the code logic follows the inherent logic of the causal chain.

[0027] This invention also provides an artificial intelligence-based WMS software development assistance system, comprising:

[0028] The federated rule management module is used to adopt a distributed architecture, storing public rules in the aviation logistics field and private rules of each participating entity through multiple rule nodes, so as to realize secure exchange of rule parameters between nodes, arbitration of rule conflicts, and self-evolutionary updates.

[0029] The causal code generation module is used to obtain rule graphs from the federated rule management module, mine causal chains, and convert natural language requirements into initial code fragments containing conditional logic. The causal chains include causes, effects, solutions, and moderating variables.

[0030] The compliance assessment module is used to assess and process the initial code snippets for compliance with rules, and decide whether to proceed to cross-system collaborative code generation or return to regenerate the code based on the assessment result.

[0031] The collaborative verification module is used to generate cross-system collaborative code and perform causal consistency verification, realize the interface adaptation between WMS and external systems, and ensure that the code logic is consistent with the causal chain through virtual scenario testing.

[0032] The present invention also provides an electronic device, which is a physical device, comprising:

[0033] The processor and the memory are communicatively connected.

[0034] The memory is used to store at least one executable instruction executed by the processor, which executes the executable instruction to implement the AI-based WMS software development assistance method described above.

[0035] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the artificial intelligence-based WMS software development assistance method described above.

[0036] Compared with the prior art, the beneficial effects of the present invention are:

[0037] This invention achieves secure collaboration and dynamic updates of multi-agent rules through a federated rule knowledge graph, resolving the contradiction between rule lag and privacy protection; it generates code based on causal reasoning to ensure logical traceability and reduce compliance risks; it improves code quality and reliability through compliance judgment and digital twin verification; and it automatically generates cross-system interface code to shorten the integration cycle. Ultimately, it significantly improves the development efficiency, adaptation accuracy, and system reliability of aviation logistics WMS, reduces code redundancy, enhances adaptability to complex scenarios, and meets the high compliance and high reliability requirements of aviation logistics. Attached Figure Description

[0038] Figure 1 The main flowchart of an artificial intelligence-based WMS software development assistance method provided in this embodiment of the invention;

[0039] Figure 2 A schematic diagram of the structure of an AI-based WMS software development assistance system provided in an embodiment of the present invention;

[0040] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0041] 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.

[0042] The method in this embodiment is executed by a terminal, which can be a mobile phone, tablet computer, PDA, laptop or desktop computer, etc. Of course, it can also be other devices with similar functions, and this embodiment does not limit them.

[0043] Please see Figure 1 This invention provides an artificial intelligence-based WMS software development assistance method, which is applied to, including:

[0044] Step 100: Construct a federated rule knowledge graph. The federated rule knowledge graph adopts a distributed architecture, with multiple rule nodes storing public rules in the aviation logistics field and private rules of each participating entity. The nodes achieve secure exchange and dynamic updating of rule parameters through a federated learning protocol, providing a rule basis for subsequent code generation.

[0045] Step 100 aims to build a basic system that can dynamically integrate and update rules in the field of air logistics. Its core purpose is to solve the problems of privacy protection and collaborative use of rules by multiple parties. Through distributed storage and federated learning technology, it realizes secure sharing and real-time updates of rules, providing accurate and comprehensive rule basis for subsequent code generation. The federated rule knowledge graph refers to a structured knowledge set based on a federated learning framework, containing various rules and their relationships in the air logistics field. The distributed architecture refers to a system structure that distributes rule storage and processing functions across multiple nodes rather than concentrating them on a single device. Rule nodes are independent computing units involved in rule storage and processing, responsible for managing and updating local rules. Public rules refer to widely recognized and publicly available rules in the air logistics field, such as the General Carriage Terms and Conditions (GCCT) established by the International Air Transport Association (IATA). Air logistics participants refer to various organizations or institutions involved in air logistics activities, including but not limited to airports, customs, freight forwarders, and e-commerce platforms. Private rules refer to rules owned by each air logistics participant that are confidential, such as customs internal inspection standards or special warehousing requirements of e-commerce platforms. The federated learning protocol refers to a series of agreements used to regulate data exchange and collaborative computation between rule nodes, ensuring that nodes only exchange model parameters, not raw rule data. Rule parameters refer to key feature values ​​extracted from rules for model training or collaborative computation. Through this process, dynamic fusion and updating of rules can be achieved while protecting the privacy of each participant's rules, avoiding problems such as rule lag or difficulties in collaboration.

[0046] In one possible implementation, the rule nodes in the distributed architecture can be divided into a central node and multiple distributed nodes. The central node is deployed on a cloud server and is responsible for storing and managing public rules and the collaborative logic between nodes. The distributed nodes are deployed on the local servers of each participating entity and are responsible for storing private rules and participating in parameter exchange during the federated learning process.

[0047] Specifically, step 100 also includes:

[0048] Step 110: Deploy distributed rule nodes, which include a central node for storing public rules and distributed nodes for storing private rules of aviation logistics participants. The nodes establish encrypted communication links through a federated learning protocol to achieve secure exchange of rule parameters without transmitting the original data.

[0049] Step 120: Integrate explicit and implicit rules. Extract structured explicit rules from authoritative data sources using natural language processing technology. Mine empirical implicit rules from historical operation logs based on reinforcement learning models. After multi-subject distributed node collaborative verification, incorporate them into the federated rule knowledge graph.

[0050] Step 130: Perform rule conflict arbitration and self-evolution. When multiple subject rules conflict, call the aviation scenario weight model to calculate the weight coefficient based on cargo type and flight attributes and output the optimal coordination solution. At the same time, automatically optimize inefficient rules based on rule survival index, which includes compliance contribution and efficiency improvement value.

[0051] Among them, distributed rule nodes refer to rule storage and processing units divided by function, including a central node storing public rules and distributed nodes storing private rules of each entity; federated learning protocols refer to the conventions that regulate data exchange between nodes, ensuring the secure transmission of rule parameters; explicit rules refer to structured rules directly extracted from authoritative data sources (such as IATA announcements and General Administration of Customs documents); implicit rules refer to empirical rules mined from historical operation logs, which need to be verified by multiple entities before being incorporated into the graph; the aviation scenario weight model refers to a model that calculates rule priorities based on factors such as cargo type and flight attributes; and rule survival index refers to parameters that evaluate the effectiveness of rules, including compliance contribution (the degree to which a rule contributes to meeting regulatory requirements) and efficiency improvement value (the extent to which a rule improves warehousing efficiency). Through the above steps, the privacy of rules of each entity can be protected, and the comprehensive integration and dynamic evolution of rules can be achieved, providing high-quality rule input for subsequent code generation.

[0052] In one possible implementation, the central node can be deployed on a cloud server to be responsible for the unified management and updating of public rules; the distributed nodes are deployed on the local servers of each entity and only output encrypted rule parameters when participating in federated learning, without disclosing the original private rules.

[0053] For example, in one feasible implementation, the system first deploys a central node (storing IATA general rules) and distributed nodes (local nodes for airports, customs, and e-commerce platforms, storing airport warehousing specifications, customs inspection standards, and e-commerce cold chain requirements, respectively). The nodes establish encrypted communication through a federated learning protocol. Then, natural language processing technology is used to parse the IATA rules. The 2024 version of the dangerous goods rules obtains explicit rules and extracts implicit rules such as "prioritizing fresh goods for wide-body aircraft" from the airport's nearly one year of flight delay cargo allocation logs. After cross-validation between the airport and e-commerce platform nodes, these rules are incorporated into the knowledge graph. When the airport's rule of "stopping warehousing 2 hours before flight departure" conflicts with the e-commerce rule of "prioritizing warehousing for promotional goods", the aviation scenario weight model calculates the warehousing flexibility weight of 0.6 based on the "promotional daily necessities" type and the "commercial flight" attribute, and outputs a coordination solution of "allowing warehousing with streamlined security check process". At the same time, the "manual verification of paper documents" rule, which has not been called for three consecutive months, is automatically optimized based on the low survival rate index and replaced with the "automatic verification of electronic documents" rule, ultimately forming a dynamically updated federated rule knowledge graph.

[0054] In step 110, the federated learning protocol uses homomorphic encryption technology to encrypt the rule parameters. The participants in the air logistics include airports, customs, freight forwarders, and e-commerce platforms. The public rules include the general terms of the International Air Transport Association (IATA), and the private rules include customs inspection standards and the special warehousing requirements of e-commerce platforms.

[0055] Homomorphic encryption technology refers to an encryption method that allows computation on encrypted data without decryption, ensuring that rule parameters do not leak original information during transmission and computation. Aviation logistics participants refer to organizations involved in the entire aviation logistics chain, including airports (responsible for cargo takeoffs, landings, and warehousing), customs (responsible for cross-border cargo supervision), freight forwarders (responsible for cargo transportation), and e-commerce platforms (responsible for cargo orders and sales). Public rules refer to industry-wide and publicly available rules, such as the dangerous goods classification and transportation standards established by the International Air Transport Association (IATA). Private rules refer to the proprietary rules within each entity, such as customs inspection procedures for specific commodities and special warehousing requirements for high-value goods by e-commerce platforms. These detailed designs ensure the security and specificity of rule sharing under the distributed architecture, avoiding privacy leaks while meeting the collaborative needs of multiple entities.

[0056] In one possible implementation, homomorphic encryption technology can employ the Paillier encryption algorithm to encrypt the rule parameters, enabling nodes to directly perform aggregation calculations on the encrypted parameters and complete collaborative rule updates without decryption.

[0057] For example, in one feasible implementation, the central node stores the IATA public rule that "lithium batteries UN3480 require anti-static storage," the airport distributed nodes store the private rule that "lithium battery storage locations must be ≤5 meters from fire lanes," and the customs distributed nodes store the private rule that "lithium batteries require an additional certificate of origin." When updating the rules, each node exchanges rule parameters encrypted using the Paillier algorithm through a federated learning protocol. The central node aggregates the parameters to generate an updated model, which is then distributed to each distributed node. Each node decrypts and updates the rules using its local private rules. Throughout the process, the original data of private rules such as "fire lane distance" and "certificate of origin requirements" are never leaked. Collaboration is achieved only through encrypted parameters, thus protecting privacy while completing the rule update.

[0058] Step 200: Generate initial code snippets based on causal reasoning. Using the causal chains mined from the federated rule knowledge graph, convert natural language requirements into code containing conditional logic. The causal chains include causes, effects, solutions, and moderating variables.

[0059] The core objective of step S200 is to convert the developers' natural language requirements into initial code snippets with clear logic, ensuring the traceability of the code logic. The key lies in establishing an intrinsic connection between rules and code using causal chains, avoiding logical ambiguity caused by generating code solely based on statistical correlation. Here, causal reasoning refers to the process of mining causal relationships between things through algorithms, i.e., analyzing "why the rules need to be formulated this way"; initial code snippets refer to code portions generated based on requirements that have not yet been fully verified; causal chains refer to the chain describing the causal relationships behind the rules, consisting of cause, effect, solution, and moderating variables; natural language requirements refer to the functional development requirements expressed by developers in everyday language, such as "develop a lithium battery storage module suitable for high-altitude airports"; conditional logic refers to the logical structure in the code that executes different operations based on different preconditions; cause refers to the reason leading to a certain result, such as "low air pressure at high altitudes"; effect refers to the result caused by the cause, such as "increased internal pressure of the lithium battery"; solution refers to the countermeasures taken for the "effect," such as "requires a low-altitude storage area"; and moderating variables refer to factors affecting the strength or applicability of the causal relationship, such as "altitude > 3000 meters". This step generates code that not only meets the rule requirements but also clearly reflects the logic behind the rules, facilitating subsequent verification and maintenance.

[0060] Additionally, it should be noted that the mining of causal chains is the core of this step. By adding a causal relationship layer to the federated rule knowledge graph, the originally isolated rules can be transformed into logically related chains, changing the code generation process from "based on the surface requirements of the rules" to "based on the internal logic of the rules", thereby improving the code's adaptability to complex scenarios.

[0061] In one possible implementation, causal reasoning can employ the Do-Calculus algorithm, which uses intervention analysis to determine the causal relationships between the elements in the rules, identify which factors are the key causes of the results and which are secondary factors, and thus construct an accurate causal chain.

[0062] Specifically, step 200 includes:

[0063] Step 210: Construct a causal rule graph. Add a causal relationship layer to the federated rule knowledge graph. Use causal reasoning algorithms to mine the causal chain behind the rules and clarify the preconditions and exceptions for the application of the rules.

[0064] Step 220: Analyze the natural language requirements, identify the core entities in the requirements through the natural language processing model, and combine the constraints with the causal rule graph to generate a structured requirement list containing causal logic descriptions.

[0065] Step 230: Generate condition-adaptive code. Based on the structured requirements list, call the aviation scenario-specific template library to generate the initial code snippet containing prerequisite judgments and corresponding operation logic. The aviation scenario-specific template library includes sub-templates for dangerous goods management, cross-border supervision, and cold chain transportation.

[0066] Among these, the causal rule graph refers to a knowledge graph containing causal relationships within rules, adding a causal relationship layer to the federated rule knowledge graph; the causal reasoning algorithm refers to algorithms used to mine causal relationships (such as Do-Calculus), which can determine the "cause-effect-solution" association in rules; the natural language processing model refers to AI models used to parse developer requirements (such as BERT), which can identify core entities and constraints in requirements; the structured requirement list refers to converting natural language requirements into machine-understandable structured data, including entities, constraints, and causal logic descriptions; the aviation scenario-specific template library refers to a set of pre-set code templates for different aviation logistics scenarios (dangerous goods management, cross-border supervision, cold chain transportation); and the initial code snippet refers to the code prototype generated based on structured requirements, containing conditional judgments and corresponding operations. Through these steps, the generated code not only meets the rule requirements but also clearly reflects the causal logic behind the rules, facilitating subsequent verification and maintenance.

[0067] In one possible implementation, the aviation scenario-specific template library can be subdivided into sub-templates according to the scenario. Each sub-template contains the code framework for common operations in that scenario (e.g., the dangerous goods management template contains the code structure for storage location allocation and safety monitoring). When generating code, the corresponding sub-template is automatically called and parameters are filled in according to the requirements.

[0068] For example, in one feasible implementation, the system first constructs a causal chain for the "lithium battery storage" rule based on a federated rule knowledge graph using the Do-Calculus algorithm (lithium battery friction-induced static electricity (cause) → short circuit risk (effect) → anti-static storage location (solution), with humidity as the regulating variable), forming a causal rule graph. When developers input the requirement of "developing a lithium battery storage module for high-altitude airports," the BERT model identifies the core entities "lithium battery" and "high-altitude airport," and, combined with the causal rule graph, parses out the constraint condition "low air pressure leads to increased battery pressure, requiring depressurization storage," generating a structured requirement list. Subsequently, the system calls the "high-altitude airport dangerous goods template" to generate an initial code snippet containing the logic of "allocation to depressurization storage location and pressure monitoring when altitude > 3000 meters." Each operation in this code corresponds to a specific link in the causal chain, ensuring logical traceability.

[0069] Step 300: Perform rule compliance judgment and processing on the initial code fragment, and decide whether to enter cross-system collaborative code generation or return to regenerate code based on the judgment result.

[0070] The main purpose of step 300 is to ensure that the initial code snippet conforms to the latest rules in the federated rule knowledge graph. Through strict compliance checks and branching, non-compliant code is prevented from entering subsequent processes, thereby reducing issues in cross-system collaboration and verification phases and improving overall development efficiency. Rule compliance checks refer to verifying whether the logic of the initial code snippet meets the requirements of the rules in the federated rule knowledge graph; processing refers to subsequent actions taken based on the compliance check results, including proceeding to the next stage or regenerating the code; cross-system collaborative code generation refers to the process of generating code for WMS to interface with external systems. This step enables the identification and correction of violations in the early stages of code generation, ensuring that the subsequently generated cross-system collaborative code has basic compliance.

[0071] In one possible implementation, rule compliance judgment can be achieved by building a rule comparison engine. This engine converts the key operational logic (such as library allocation conditions, security threshold parameters, etc.) in the initial code snippet into structured data, compares it item by item with the corresponding rules in the federated rule knowledge graph, calculates the compliance degree of each comparison result, and then calculates the overall compliance score based on the weights of each item.

[0072] Specifically, step 300 includes:

[0073] Step 310: Perform rule compliance verification, extract key operation logic from the initial code snippet, compare it with the latest rules in the federated rule knowledge graph, and calculate the compliance score;

[0074] Step 320: Execute branch processing. If the compliance score is ≥95, proceed to the stage of generating cross-system collaborative code and performing causal consistency verification. If the compliance score is <95, return to the stage of generating conditional adaptive code, adjust parameters according to the violation prompt, and regenerate the code.

[0075] The process includes: rule compliance verification (checking the degree of matching between the initial code snippet and the rules in the federated rule knowledge graph); key operational logic (core logic in the code that affects compliance, such as library allocation conditions and security threshold parameters); compliance scoring (a quantitative indicator measuring the degree of matching between the code and the rules, with a maximum score of 100); branch processing (different operations based on the compliance score: scores ≥ 95 enter cross-system collaborative code generation, scores < 95 return to regenerate the code); and violation alerts (informing the user of specific information about the code's non-compliance with the rules, such as "insufficient spacing threshold"). This process can intercept violation logic early in code generation, reducing the cost of subsequent modifications and improving development efficiency.

[0076] In one possible implementation, the compliance score can be calculated using a weighted average: rules are divided into core rules (such as safety distance requirements) and general rules (such as log recording formats), with a matching weight of 0.7 for core rules and 0.3 for general rules, and the weighted sum is the final score.

[0077] For example, in one feasible implementation, for the initial code snippet of the "lithium battery storage module", the rule compliance verification module extracts key operational logic (such as "the distance between the lithium battery storage location and ordinary goods is 8 meters" and "monitor static electricity once per hour"); it compares it with the core rules ("distance ≥ 10 meters" and "monitor once every 30 minutes") and general rules ("monitoring time needs to be recorded") in the federated rule knowledge graph. The core rule matching degree is 0 (distance does not meet the standard), and the general rule matching degree is 1 (record is complete). The compliance score is calculated as 0.7×0 + 0.3×1 = 30 points (< 95 points); the system triggers branch processing, returns to the code generation stage and prompts "distance is insufficient and monitoring frequency is too low". The code generation module adjusts the distance to 12 meters and changes the monitoring frequency to once every 20 minutes. After regenerating the code, the score reaches 98 points, and the system enters the cross-system collaborative code generation stage.

[0078] Step 400: Generate cross-system collaborative code and perform causal consistency verification to achieve interface adaptation between WMS and external systems, and ensure that the code logic is consistent with the causal chain through virtual scenario testing.

[0079] Step 400 aims to achieve seamless integration between the WMS and external systems, and ensure the reliability and logical consistency of the final code in real-world scenarios. This is achieved by generating cross-system collaborative code to address interface adaptation issues between systems, and by using causal consistency verification to ensure the code logic matches the inherent logic of the rules. Specifically, cross-system collaborative code refers to the code used to achieve data interaction and functional collaboration between the WMS and external systems; causal consistency verification refers to the process of verifying whether the code logic is consistent with the inherent logic of the causal chain; interface adaptation refers to the adaptation process that enables the WMS and external systems to correctly exchange data and perform collaborative operations; external systems refer to other systems that have data interaction or functional collaboration with the WMS, including airport IoT devices (such as RFID scanners, temperature and humidity sensors), customs supervision systems, e-commerce platform order systems, etc.; and virtual scenario testing refers to the process of testing the code functionality in a computer-constructed simulated aviation warehousing environment. This step ensures that the generated code not only achieves cross-system collaboration but also maintains logical correctness under various complex scenarios, meeting the high reliability requirements of air logistics.

[0080] Additionally, it should be noted that the difference between causal consistency verification and rule compliance judgment is that rule compliance judgment mainly checks whether the code meets the surface requirements of the rules, while causal consistency verification delves into whether the code logic follows the causal relationship behind the rules. For example, it not only checks whether "lithium batteries are stored in the anti-static storage location", but also checks whether "the anti-static storage location is skipped when the humidity is ≥80%", to ensure that the code logic is consistent with the causal chain.

[0081] In one possible implementation, virtual scene testing can be based on the Unity engine to build a virtual aviation warehouse scene containing elements such as a tarmac, hazardous materials warehouse, and cold chain area. By simulating extreme scenarios such as rainstorms and flight cancellations, the execution logic of the code in different environments can be tested, and the degree of matching between the code behavior and the causal chain can be recorded.

[0082] Specifically, step 400 includes:

[0083] Step 410: Construct a system collaboration knowledge graph to record the interface protocols, data formats, and collaboration logic between WMS and external systems. The external systems include airport IoT devices, customs supervision systems, and e-commerce platform order systems.

[0084] Step 420: Generate cross-system interface code. Based on the system collaborative knowledge graph and the code that passes the compliance judgment, automatically generate interface adaptation code to achieve seamless integration between WMS and the external system.

[0085] Step 430: Perform causal consistency verification. Construct a virtual scenario of aviation warehouse based on digital twin technology, simulate extreme scenarios to test code logic, trace the consistency between code logic and the causal chain, output a causal consistency report, and automatically repair non-compliant code.

[0086] The system collaboration knowledge graph refers to a graph recording the interface rules between the WMS and external systems, including interface protocols, data formats, and collaboration logic. External systems refer to systems that interact with the WMS, including airport IoT devices (such as RFID scanners and temperature / humidity sensors), customs supervision systems (for cargo clearance data exchange), and e-commerce platform order systems (for order and inventory synchronization). Interface protocols refer to the specifications for data transmission between systems (such as MQTT and HTTP). Data formats refer to the data organization (such as JSON and XML). Cross-system interface code refers to the code that implements data interaction between the WMS and external systems. Digital twin technology refers to the technology of constructing virtual scenarios consistent with the physical warehousing environment. Aviation warehousing virtual scenarios refer to virtual spaces simulating the aviation warehousing environment, including elements such as tarmacs, dangerous goods warehouses, and cold chain areas. Extreme scenarios refer to special situations that may affect warehousing logic (such as heavy rain or flight cancellations). Causal consistency reports refer to reports recording the degree of matching between code logic and causal chains. This ensures that the code not only works collaboratively across systems but also maintains logical correctness in complex scenarios.

[0087] In one possible implementation, the system collaborative knowledge graph can be updated in real time. When the interface rules of external systems change (such as when the customs supervision system upgrades its data format), the graph is automatically updated synchronously to ensure that the generated interface code always adapts to the latest interface requirements.

[0088] For example, in one feasible implementation, the system first constructs a system collaboration knowledge graph to record the interface rules between the WMS and airport temperature and humidity sensors (MQTT protocol, JSON format), customs supervision system (HTTPS protocol, XML format), and the collaboration logic of "adjusting inventory within 15 minutes when flights are delayed"; based on the lithium battery storage code that has passed compliance judgment, interface adaptation code is generated (such as parsing the humidity data of the sensor and uploading the lithium battery inspection record to the customs system); subsequently, a virtual scenario including the high-altitude airport apron and depressurization storage space is constructed using digital twin technology to simulate rainstorm (90% humidity) and flight cancellation scenarios to test the code: verifying whether the code skips the anti-static storage space based on sensor data (causal chain verification) and whether it pushes inventory changes to the e-commerce order system in a timely manner (cross-system collaboration verification); the test found that the code did not handle the storage space release logic after flight cancellation, the system automatically repaired and generated a causal consistency report, and finally output the complete code that meets the requirements.

[0089] In step 430, the virtual aviation warehouse scenario includes a tarmac, a hazardous materials warehouse, and a cold chain area. The extreme scenarios include heavy rain and flight cancellations. The causal consistency verification not only verifies whether the code conforms to the rules, but also verifies whether the code logic follows the inherent logic of the causal chain.

[0090] The virtual aviation warehousing scenario refers to a virtual space constructed using digital twin technology that simulates the actual aviation warehousing environment. It includes key elements such as the apron (cargo loading and unloading area), dangerous goods warehouse (a dedicated area for storing dangerous goods), and cold chain area (an area for storing perishable goods and other temperature-controlled goods). Extreme scenarios refer to special circumstances that may cause changes in the applicable conditions of the rules, such as heavy rain (affecting humidity, thus altering static electricity risk) and flight cancellations (affecting cargo entry time and storage location allocation). The inherent logic of the causal chain refers to the "cause-effect-solution" relationship (e.g., the logic of "low air pressure → increased battery pressure → need for depressurized storage"). Ensuring the logical correctness of the code under various complex scenarios meets the high reliability requirements of aviation logistics.

[0091] In one possible implementation, causal consistency verification can be achieved through log tracing: embedding causal chain identifiers in the code, recording the causal chain links corresponding to each operation, and analyzing the logs during testing to determine whether the code operations match the causal chain links.

[0092] For example, in one feasible implementation, the virtual aviation warehouse scenario includes a high-altitude airport tarmac, dedicated depressurized storage spaces for lithium batteries, a cold chain area, and supporting pressure sensors, temperature and humidity sensors. During testing, extreme scenarios are simulated: an altitude of 4000 meters (low air pressure), humidity of 85% (high humidity), and flight cancellation (requiring delayed storage). The system runs the lithium battery storage code, and log tracing reveals that the code allocates depressurized storage spaces based on low air pressure data (matching the causal chain of "low air pressure → requiring depressurized storage"), skips anti-static operations due to humidity > 80% (matching the adjustment variable of "high humidity reduces static electricity risk"), but fails to handle the logic for extending storage space reservation after flight cancellation (not matching the causal chain of "flight cancellation → delayed storage"). The system then repairs the code, adding logic to "extend storage space reservation by 24 hours when a flight is cancelled." After retesting, the code fully matches the inherent causal chain logic and passes verification.

[0093] In this embodiment, the present invention achieves secure collaboration and dynamic updating of multi-agent rules through a federated rule knowledge graph, resolving the contradiction between rule lag and privacy protection; it generates code based on causal reasoning to ensure logical traceability and reduce compliance risks; it improves code quality and reliability through compliance judgment and digital twin verification; and it automatically generates cross-system interface code to shorten the integration cycle. Ultimately, it significantly improves the development efficiency, adaptation accuracy, and system reliability of aviation logistics WMS, reduces code redundancy, enhances adaptability to complex scenarios, and meets the high compliance and high reliability requirements of aviation logistics.

[0094] Based on the above embodiments, such as Figure 2As shown, the present invention also provides an AI-based WMS software development assistance system to support the AI-based WMS software development assistance method of the above embodiments. The AI-based WMS software development assistance system includes:

[0095] The federated rule management module 11 is used to adopt a distributed architecture, storing public rules in the aviation logistics field and private rules of each participating entity through multiple rule nodes, so as to realize secure exchange of rule parameters between nodes, arbitration of rule conflicts, and self-evolutionary updates.

[0096] The causal code generation module 12 is used to obtain a rule graph from the federated rule management module, mine causal chains, and convert natural language requirements into initial code fragments containing conditional logic. The causal chains include causes, effects, solutions, and moderating variables.

[0097] The compliance judgment module 13 is used to judge and process the rule compliance of the initial code fragment, and decide whether to enter the cross-system collaborative code generation or return to regenerate the code based on the judgment result;

[0098] The collaborative verification module 14 is used to generate cross-system collaborative code and perform causal consistency verification, realize the interface adaptation between WMS and external systems, and ensure that the code logic is consistent with the causal chain through virtual scenario testing.

[0099] In an optional embodiment, the federated rule management module 11 is further configured to deploy distributed rule nodes, which include a central node for storing public rules and distributed nodes for storing private rules of aviation logistics participants. Nodes establish encrypted communication links through a federated learning protocol to achieve secure exchange of rule parameters without transmitting raw data. Explicit and implicit rules are integrated. Structured explicit rules are extracted from authoritative data sources using natural language processing technology, and empirical implicit rules are mined from historical operation logs based on a reinforcement learning model. These implicit rules are then incorporated into the federated rule knowledge graph after collaborative verification by multiple distributed nodes. Rule conflict arbitration and self-evolution are performed. When multiple subject rules conflict, the aviation scenario weight model is invoked to calculate weight coefficients based on cargo type and flight attributes and output the optimal coordination solution. Simultaneously, inefficient rules are automatically optimized based on rule survival indicators, which include compliance contribution and efficiency improvement values.

[0100] The federated learning protocol uses homomorphic encryption to encrypt the rule parameters. The participants in the air logistics include airports, customs, freight forwarders, and e-commerce platforms. The public rules include the general terms of the International Air Transport Association (IATA), and the private rules include customs inspection standards and the special warehousing requirements of e-commerce platforms.

[0101] In an optional embodiment, the causal code generation module 12 is further configured to construct a causal rule graph, adding a causal relationship layer to the federated rule knowledge graph, mining the causal chain behind the rules through causal reasoning algorithms, clarifying the preconditions and exceptions for rule application; parsing natural language requirements, identifying core entities in the requirements through a natural language processing model, and parsing constraints in conjunction with the causal rule graph to generate a structured requirement list containing causal logic descriptions; generating condition-adaptive code, calling an aviation scenario-specific template library based on the structured requirement list to generate the initial code fragment containing premise judgments and corresponding operational logic, wherein the aviation scenario-specific template library includes sub-templates for dangerous goods management, cross-border supervision, and cold chain transportation.

[0102] In an optional embodiment, the compliance judgment module 13 is further configured to perform rule compliance verification, extract key operation logic from the initial code fragment, compare it with the latest rules in the federated rule knowledge graph, calculate a compliance score, and execute branch processing. If the compliance score is ≥95, the module proceeds to the stage of generating cross-system collaborative code and performing causal consistency verification. If the compliance score is <95, the module returns to the stage of generating conditional adaptive code, adjusting parameters according to the violation prompts, and regenerating the code.

[0103] In an optional embodiment, the collaborative verification module 14 is further configured to construct a system collaborative knowledge graph, record the interface protocols, data formats, and collaborative logic between the WMS and external systems, including airport IoT devices, customs supervision systems, and e-commerce platform order systems; generate cross-system interface code, automatically generate interface adaptation code based on the system collaborative knowledge graph and the code that passes compliance judgment, and achieve seamless integration between the WMS and the external systems; perform causal consistency verification, construct an aviation warehouse virtual scenario based on digital twin technology, simulate extreme scenario test code logic, trace the consistency between the code logic and the causal chain, output a causal consistency report, and automatically repair non-compliant code.

[0104] The virtual aviation warehouse scenario includes an apron, a hazardous materials warehouse, and a cold chain area. The extreme scenarios include heavy rain and flight cancellations. The causal consistency verification not only verifies whether the code conforms to the rules, but also verifies whether the code logic follows the inherent logic of the causal chain.

[0105] In this embodiment, the present invention achieves secure collaboration and dynamic updating of multi-agent rules through a federated rule knowledge graph, resolving the contradiction between rule lag and privacy protection; it generates code based on causal reasoning to ensure logical traceability and reduce compliance risks; it improves code quality and reliability through compliance judgment and digital twin verification; and it automatically generates cross-system interface code to shorten the integration cycle. Ultimately, it significantly improves the development efficiency, adaptation accuracy, and system reliability of aviation logistics WMS, reduces code redundancy, enhances adaptability to complex scenarios, and meets the high compliance and high reliability requirements of aviation logistics.

[0106] Furthermore, the AI-based WMS software development assistance system can run the aforementioned AI-based WMS software development assistance method. For specific implementation details, please refer to the method embodiment, which will not be repeated here.

[0107] Based on the above embodiments, such as Figure 3 As shown, the present invention also provides an electronic device, the electronic device comprising:

[0108] The processor 22 includes at least one processor 22, at least one memory 21, a communication interface 23, and a communication bus 24, wherein the processor 22 is communicatively connected to the memory 21.

[0109] In this embodiment, the memory 21 can be implemented in any suitable manner, for example, the memory 21 can be a read-only memory, a hard disk drive, a solid-state drive, or a USB flash drive, etc.; the memory 21 is used to store at least one executable instruction executed by the processor;

[0110] In this embodiment, the processor 22 can be implemented in any suitable manner. For example, the processor 22 can take the form of a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) that can be executed by the (micro)processor, logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers, and embedded microcontrollers, etc.; the processor is used to execute the executable instructions to implement the AI-based WMS software development assistance method as described above.

[0111] Based on the above embodiments, the present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the artificial intelligence-based WMS software development assistance method described above.

[0112] Those skilled in the art will recognize that the modules and method steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0113] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, equipment, and modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0114] In the several embodiments provided in this application, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or units may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interfaces, devices, or equipment, and may be electrical, mechanical, or other forms.

[0115] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0116] In addition, the functional modules in the various embodiments of the present invention can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0117] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program instructions, such as USB flash drives, portable hard drives, read-only storage servers, random access storage servers, magnetic disks, or optical disks.

[0118] Furthermore, it should be noted that the combination of the various technical features in this case is not limited to the combination methods described in the claims of this case or the combination methods described in the specific embodiments. All technical features described in this case can be freely combined or combined in any way, unless they contradict each other.

[0119] It should be noted that the above examples are merely specific embodiments of the present invention, and the present invention is obviously not limited to the above embodiments, with many similar variations. All modifications that can be directly derived or conceived by those skilled in the art from the content disclosed in this invention should fall within the protection scope of this invention.

[0120] The above are merely preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A WMS software development assistance method based on artificial intelligence, characterized in that, Includes the following steps: A federated rule knowledge graph is constructed. The federated rule knowledge graph adopts a distributed architecture. Multiple rule nodes store public rules in the field of aviation logistics and private rules of each participating entity. The nodes achieve secure exchange and dynamic updating of rule parameters through a federated learning protocol, providing a rule basis for subsequent code generation. Initial code snippets are generated based on causal reasoning. Using causal chains mined from the federated rule knowledge graph, natural language requirements are converted into code containing conditional logic. The causal chains include cause, effect, solution, and moderating variables. The initial code snippet is subjected to rule compliance judgment and processing, and the decision is made based on the judgment result to proceed to cross-system collaborative code generation or return to regenerate the code; Generate cross-system collaborative code and perform causal consistency verification to achieve interface adaptation between WMS and external systems, and ensure that the code logic is consistent with the causal chain through virtual scenario testing.

2. The AI-based WMS software development assistance method according to claim 1, characterized in that, The construction of a federated rule knowledge graph employs a distributed architecture, storing public rules in the aviation logistics field and private rules of each participating entity through multiple rule nodes. Nodes securely exchange and dynamically update rule parameters via a federated learning protocol, providing a rule-based basis for subsequent code generation, including: Deploy distributed rule nodes, which include a central node for storing public rules and distributed nodes for storing private rules of aviation logistics participants. The nodes establish encrypted communication links through a federated learning protocol to achieve secure exchange of rule parameters without transmitting raw data. By integrating explicit and implicit rules, structured explicit rules are extracted from authoritative data sources using natural language processing technology, and empirical implicit rules are mined from historical operation logs based on reinforcement learning models. After multi-subject distributed node collaborative verification, these rules are incorporated into the federated rule knowledge graph. The system performs rule conflict arbitration and self-evolution. When multiple parties' rules conflict, it calls the aviation scenario weight model to calculate weight coefficients based on cargo type and flight attributes and outputs the optimal coordination solution. At the same time, it automatically optimizes inefficient rules based on rule survival index, which includes compliance contribution and efficiency improvement value.

3. The AI-based WMS software development assistance method according to claim 2, characterized in that, The federated learning protocol uses homomorphic encryption technology to encrypt the rule parameters. The participants in the air logistics include airports, customs, freight forwarders, and e-commerce platforms. The public rules include the general terms of the International Air Transport Association (IATA), and the private rules include customs inspection standards and the special warehousing requirements of e-commerce platforms.

4. The AI-based WMS software development assistance method according to claim 1, characterized in that, The initial code snippet generated based on causal reasoning utilizes causal chains mined from the federated rule knowledge graph to convert natural language requirements into code containing conditional logic. These causal chains include causes, effects, solutions, and moderating variables, including: Construct a causal rule graph, add a causal relationship layer on the basis of the federated rule knowledge graph, and use causal reasoning algorithms to mine the causal chain behind the rules to clarify the preconditions and exceptions for the application of the rules; The natural language requirements are analyzed, the core entities in the requirements are identified through the natural language processing model, and the constraints are analyzed in combination with the causal rule graph to generate a structured list of requirements containing causal logic descriptions. The system generates condition-adaptive code by calling an aviation scenario-specific template library based on the structured requirements list. This generates the initial code snippet containing prerequisite judgments and corresponding operational logic. The aviation scenario-specific template library includes sub-templates for dangerous goods management, cross-border supervision, and cold chain transportation.

5. The AI-based WMS software development assistance method according to claim 1, characterized in that, The step of performing rule compliance judgment and processing on the initial code fragment, and deciding whether to proceed to cross-system collaborative code generation or return to regenerate the code based on the judgment result, includes: Perform rule compliance verification, extract key operation logic from the initial code snippet, compare it with the latest rules in the federated rule knowledge graph, and calculate a compliance score; If the compliance score is ≥95, proceed to the stage of generating cross-system collaborative code and performing causal consistency verification; if the compliance score is <95, return to the stage of generating conditional adaptive code, adjust parameters according to the violation prompt, and regenerate the code.

6. The AI-based WMS software development assistance method according to claim 1, characterized in that, The process of generating cross-system collaborative code and performing causal consistency verification, achieving interface adaptation between WMS and external systems, and ensuring code logic consistency with the causal chain through virtual scenario testing includes: Construct a system collaboration knowledge graph to record the interface protocols, data formats, and collaboration logic between WMS and external systems, including airport IoT devices, customs supervision systems, and e-commerce platform order systems. Generate cross-system interface code. Based on the system's collaborative knowledge graph and the code that passes compliance judgment, automatically generate interface adaptation code to achieve seamless integration between WMS and the external system. Perform causal consistency verification by constructing a virtual aviation warehouse scenario based on digital twin technology, simulating extreme scenarios to test code logic, tracing the consistency between the code logic and the causal chain, outputting a causal consistency report, and automatically fixing non-compliant code.

7. The AI-based WMS software development assistance method according to claim 6, characterized in that, The virtual aviation warehouse scenario includes a tarmac, a hazardous materials warehouse, and a cold chain area. The extreme scenarios include heavy rain and flight cancellations. The causal consistency verification not only verifies whether the code conforms to the rules, but also verifies whether the code logic follows the inherent logic of the causal chain.

8. A WMS software development assistance system based on artificial intelligence, characterized in that, include: The federated rule management module is used to adopt a distributed architecture, storing public rules in the aviation logistics field and private rules of each participating entity through multiple rule nodes, so as to realize secure exchange of rule parameters between nodes, arbitration of rule conflicts, and self-evolutionary updates. The causal code generation module is used to obtain rule graphs from the federated rule management module, mine causal chains, and convert natural language requirements into initial code fragments containing conditional logic. The causal chains include causes, effects, solutions, and moderating variables. The compliance assessment module is used to assess and process the initial code snippets for compliance with rules, and decide whether to proceed to cross-system collaborative code generation or return to regenerate the code based on the assessment result. The collaborative verification module is used to generate cross-system collaborative code and perform causal consistency verification, realize the interface adaptation between WMS and external systems, and ensure that the code logic is consistent with the causal chain through virtual scenario testing.

9. An electronic device, characterized in that, The electronic device includes: The processor and the memory are communicatively connected. The memory is used to store at least one executable instruction executed by the processor, the processor being used to execute the executable instruction to implement the AI-based WMS software development assistance method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the AI-based WMS software development assistance method as described in any one of claims 1 to 7.

Images