CN-122021707-A - Intelligent collaboration method, equipment and medium based on CAX

CN122021707ACN 122021707 ACN122021707 ACN 122021707ACN-122021707-A

Abstract

The embodiment of the specification discloses an intelligent collaboration method, equipment and medium based on CAX, and relates to the technical field of intelligent collaboration, wherein the method comprises the steps of receiving natural language design requirements input by a user, carrying out multi-mode retrieval on the natural language design requirements in a preset CAX tool library, determining a candidate tool set, carrying out pre-verification on the candidate tool set in a lightweight sandbox, and screening executable tools; and identifying potential multi-domain constraint conflicts in a preset unified industrial semantic network according to the natural language design requirements, so as to automatically resolve conflicts among a plurality of professional agents through an A2A protocol based on a predefined negotiation rule to generate an executable task sequence, and calling the executable tool to execute corresponding CAX operation according to the executable task sequence through the plurality of professional agents to realize intelligent coordination.

Inventors

XING SHENGNAN
SU XINXIN
ZHANG SHUO
LIU XUYANG
Huo Zhiyan
Hua Chengyi

Assignees

山东华云三维科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260410

Claims (10)

1. A CAX-based intelligent collaboration method, the method comprising: receiving a natural language design requirement input by a user, carrying out multi-mode retrieval on the natural language design requirement in a preset CAX tool library, determining a candidate tool set, carrying out pre-verification on the candidate tool set in a lightweight sandbox, and screening an executable tool; According to the natural language design requirement, potential multi-domain constraint conflicts are identified in a preset unified industrial semantic network, so that automatic conflict resolution is carried out among a plurality of professional intelligent agents through an A2A protocol based on a predefined negotiation rule, and an executable task sequence is generated; And calling the executable tool to execute corresponding CAX operation according to the executable task sequence by the plurality of professional agents to realize intelligent coordination.
2. The intelligent collaboration method based on CAX according to claim 1, wherein in a preset CAX tool library, the natural language design requirements are retrieved in multiple modes, and the candidate tool set is determined, which specifically includes: Carrying out semantic analysis on the natural language design requirement, extracting key design parameters and at least one constraint condition to generate a first query vector according to the key design parameters and text description in the constraint condition, wherein the constraint condition comprises geometric constraint, physical constraint or process constraint; Judging whether the natural language design requirement is associated with a geometric element designated by a user, if so, encoding the type and size information of the geometric element, and determining a second query vector; Reading a structured description file pre-generated by each CAX tool in the preset CAX tool library, wherein the structured description file at least comprises a capability vector obtained by encoding a tool function description text and an enumeration list of geometric feature types supported by the CAX tool; calculating a first similarity score for the first query vector and the capability vector for each CAX tool; if the second query vector exists, calculating a second similarity score of the second query vector and feature vectors corresponding to the geometric feature type enumeration list of each CAX tool; And sorting all CAX tools according to the first similarity score or the weighted sum of the first similarity score and the second similarity score so as to select a preset number of CAX tools ranked in the sorting result and determine the candidate tool set.
3. The CAX-based intelligent collaboration method of claim 1, wherein the candidate tool set is pre-validated in a lightweight sandbox, and executable tools are screened, comprising: creating a container instance for each CAX tool in the set of candidate tools to determine a corresponding lightweight sandboxed environment; obtaining a structured description file of the CAX tool to load in the container instance, simulating a minimized tool call scene; Generating test parameters conforming to the CAX tool input parameter mode according to key design parameters in the natural language design requirements; In the container instance, executing a verification script fragment in the structured description file, transmitting the test parameters as input to the verification script fragment, monitoring the execution process of the container instance, and capturing a corresponding standard output stream, a standard error stream and a final exit code; And analyzing the standard output stream, the standard error stream and the final exit code according to a predefined verification success rule, and if the analysis result accords with the verification success rule, judging that the CAX tool passes the pre-verification to add an executable tool set, wherein the verification success rule comprises that the final exit code is 0, the standard error stream is empty or only contains negligible warning information, and the standard output stream contains an expected successful completion identification character string or can be analyzed to obtain a valid result data fragment.
4. The CAX-based intelligent collaboration method of claim 1, wherein prior to identifying potential multi-domain constraint conflicts in a pre-set unified industrial semantic network, the method further comprises: Acquiring multi-source heterogeneous CAX data from CAD model files, CAE simulation configuration files, CAM process databases and design specification documents; Analyzing the CAD model file, extracting parameterized features, geometric entities and assembly constraint relations among the geometric entities, creating first-class nodes by each parameterized feature and/or geometric entity, and creating relation edges among the nodes based on the assembly constraint relations; analyzing the CAE simulation configuration file, extracting physical material properties, load working conditions, boundary conditions and physical field types, creating a second class node according to each physical material property and/or physical material property, and connecting the first class nodes related to the second class node through relation edges of preset types, wherein the preset types comprise types applied to and constraint to; analyzing the CAM process database, extracting manufacturing resources and machining limits, and creating each manufacturing resource and/or machining limit as a third type of node, wherein the manufacturing resources comprise machine tool capacity, tool parameters, machining precision limits and machining feature types; Creating a plurality of conflict relationship edges between the first class of nodes, the second class of nodes and the third class of nodes based on a predefined conflict rule template, wherein each conflict relationship edge is associated with a conflict type and an initial third-order weight, the conflict type comprising a geometric conflict, a physical conflict and a process conflict, the third-order weight comprising a design layer weight, a process layer weight and a manufacturing layer weight; And storing the nodes and the relation edges in a graph database to form the unified industrial semantic network.
5. The CAX-based intelligent collaboration method of claim 1, wherein identifying potential multi-domain constraint conflicts in a preset unified industrial semantic network according to the natural language design requirements comprises: Mapping key design parameters and constraint conditions in the natural language design requirements into one or more entity nodes in the unified industrial semantic network to form a first entity node set; Taking each entity node in the first entity node set as a starting point, performing traversal searching in the unified industrial semantic network along a relation edge, wherein the depth of the traversal searching is limited within a preset hop count; in the traversal process, at least one conflict relation edge is identified, and two entity nodes connected with each conflict relation edge are obtained; judging whether the two entity nodes belong to different CAX fields, if so, determining a multi-field constraint conflict pair by the two entity nodes to summarize at least one multi-field constraint conflict pair and generate a conflict list.
6. The intelligent collaboration method based on CAX of claim 5, wherein performing a traversal search along a relational edge in the unified industrial semantic network, comprises: Initializing a to-be-accessed node queue and an accessed node set, and adding a current starting node into the to-be-accessed node queue and the accessed node set; Taking any node out of the node queue to be accessed as a current node, traversing each adjacent relation edge of the current node, and judging whether the adjacent relation edge is a conflict relation edge or not, wherein the conflict relation edge comprises geometric conflict, physical conflict and process conflict; if yes, recording the adjacent relation edge into a conflict relation edge set; for each adjacent node connected with the current node through a non-conflict relation edge, if the adjacent node is not recorded in the accessed node set, adding the adjacent node into the to-be-accessed node queue and the accessed node set until the to-be-accessed node queue is empty; and after all the entity nodes in the first entity node set are used as starting nodes to finish breadth-first traversal, merging conflict relation edges recorded in all traversal processes to determine a conflict relation edge set.
7. The CAX-based intelligent collaboration method of claim 1, wherein the automatic conflict resolution among the plurality of specialized agents via the A2A protocol based on predefined negotiation rules generates the executable task sequence, comprising: Determining a current conflict pair to be negotiated from a multi-domain constraint conflict pair of a conflict list, and determining at least two related professional agents related to the current conflict pair to be negotiated; Generating a bid message containing conflict description according to the type of the conflict pair to be negotiated currently by a negotiation initiator agent, and transmitting the bid message to the related professional agent through the A2A protocol; Acquiring historical negotiation data related to the conflict pair to be negotiated between the related professional agent and the current to-be-negotiated from the unified industrial semantic network, generating a bid message according to an internal objective function of the related professional agent and the historical negotiation data, and sending the bid message back to the negotiation initiator agent through the A2A protocol; Sequencing solution proposals in at least one bidding message according to a preset decision rule by the negotiation initiator agent, and selecting the solution proposal meeting the preset requirement to determine negotiation consensus; And according to the negotiation consensus, adjusting or confirming design parameters, simulation settings and/or process planning related to the current to-be-negotiated conflict to perform automatic conflict resolution, and arranging and generating the executable task sequence according to a logic dependency relationship among each CAX operation, wherein the logic dependency relationship is derived from a time sequence or a causal edge in the unified industrial semantic network.
8. The CAX-based intelligent collaboration method of claim 7, wherein generating the bid message based on the internal objective function and history negotiation data of the associated professional agent, comprises: retrieving at least one predefined solution proposal related to the conflict type from the unified industrial semantic network according to the conflict type in the bid message so as to acquire a modification instruction of a CAX parameter set corresponding to each solution proposal; According to the modification instruction, simulating and calculating the variation of the internal objective function of the related professional intelligent agent, and determining a first evaluation factor; inquiring a history negotiation record connected with a current conflict pair node in the unified industrial semantic network, acquiring the success rate of history proposal adopting the same solution, and determining a second evaluation factor; Performing fusion calculation on the first evaluation factor and the second evaluation factor according to preset weights, and determining a comprehensive evaluation score of the solution proposal; And packaging the solution proposal and the corresponding comprehensive evaluation score, and determining the bidding message.
9. A CAX-based intelligent collaborative device, the device comprising: at least one processor, and A memory communicatively coupled to the at least one processor, wherein, The memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-8.
10. A non-transitory computer storage medium storing computer-executable instructions configured to perform the method recited in any one of claims 1-8.

Description

Intelligent collaboration method, equipment and medium based on CAX Technical Field The present disclosure relates to the field of intelligent collaboration technologies, and in particular, to a CAX-based intelligent collaboration method, apparatus, and medium. Background Computer aided technology (CAX), including Computer Aided Design (CAD), computer Aided Engineering (CAE), and Computer Aided Manufacturing (CAM), is a core mainstay of modern product development. Traditionally, these fields are operated in stages, isolated by different software systems and professionals, resulting in design, simulation, and manufacturing data chain breaks, long iteration cycles, and severely dependent on engineers' experience for cross-field coordination. In recent years, with the development of artificial intelligence and Multi-agent system (Multi-AGENT SYSTEM) technology, solutions aimed at automating CAX procedures have emerged. For example, the prior art (e.g., 202510569628.1) discloses a multi-agent computer aided engineering and computer aided design method that can understand the natural language instructions of a user and automatically invoke CAD or CAE software to perform single domain tasks such as modeling, simulation, etc. Such a system significantly reduces the operating threshold of specific software and enables task level automation. However, in the prior art, the intelligent agent collaboration mode is essentially task distribution and result recycling under the drive of sequential execution or preset rules. For example, after the CAD modeling agent completes the design, the model is transferred to the CAE simulation agent for analysis, and if the result does not meet the requirement, the system iterates according to a preset rule (e.g., returning to the previous step to modify the parameters). When a design requirement relates to multiple constraints of geometric design, physical performance, manufacturing process and the like, professional intelligent agents in different fields, such as structural design intelligent agents, thermal simulation intelligent agents and manufacturability analysis intelligent agents, possibly draw contradictory conclusions or schemes based on respective targets and knowledge, so that potential conflicts cannot be actively predicted and identified before or during task execution, a typical processing mode is error reporting and flow stopping, or the process is simply returned to the last step and manually adjusted by a user, in fact, the most complex multi-target balance and decision work is reserved for the user, intelligent coordination in a real sense cannot be realized, and an automatic process is easily interrupted at the key decision point. Therefore, in the intelligent collaboration process of CAX, potential conflicts cannot be actively predicted and identified before or during task execution, and the intelligent agent communication is only used for transmitting task instructions or result data in a fixed format, so that the intelligent collaboration requirement cannot be met. Disclosure of Invention One or more embodiments of the present disclosure provide a method, an apparatus, and a medium for intelligent collaboration based on CAX, which are used to solve the technical problem that in the intelligent collaboration process of CAX, potential conflicts cannot be actively foreseen and identified before or during task execution, and intelligent agent communication is only used to transfer task instructions or result data in a fixed format, so that the intelligent collaboration requirement cannot be satisfied. One or more embodiments of the present disclosure adopt the following technical solutions: One or more embodiments of the present disclosure provide a CAX-based intelligent collaboration method, which includes receiving a natural language design requirement input by a user, performing multimodal retrieval on the natural language design requirement in a preset CAX tool library, determining a candidate tool set, performing pre-verification on the candidate tool set in a lightweight sandbox, screening an executable tool, identifying a potential multi-domain constraint conflict in a preset unified industrial semantic network according to the natural language design requirement, performing automatic conflict resolution between a plurality of professional agents through an A2A protocol based on a predefined negotiation rule, generating an executable task sequence, and calling the executable tool to perform corresponding CAX operation according to the executable task sequence through the plurality of professional agents, so as to implement intelligent collaboration. One or more embodiments of the present specification provide a CAX-based intelligent collaborative device, comprising: at least one processor, and A memory communicatively coupled to the at least one processor, wherein, The memory stores instructions executable by the at least one processor to enable the at leas