Clustered medical building rapid construction BIM model construction method and system

Abstract

The application relates to a cluster type medical building rapid construction BIM model construction method and system, a thinking tree with multiple attributes is constructed based on design requirements, pruning is carried out in combination with a medical prototype library and compliance constraints, a DAG is generated according to leaf node tasks of the thinking tree after pruning, inter-cluster linkage is used to assign a mutual dependence key level to DAG edges, a weighted DAG is formed, a planning agent determines task priorities and resource allocation according to the weighted DAG, an execution sequence is generated, a BIM code generation agent processes tasks according to the sequence, and a component code is generated through a thinking chain process with phased confidence score evaluation; at the end of each stage, confidence is calculated according to node attributes and medical geometric rules; if the confidence is lower than a threshold value dynamically adjusted based on cluster unit complexity, the thinking chain is adjusted and restarted. When the confidence is continuously too low or a key task cluster is completed, a check and correction agent is intervened to check and correct the generated code; the corrected code is collected to construct a BIM model.

Description

Technical Field

[0001] This invention relates to the field of BIM, specifically to a method and system for rapid construction of BIM models for clustered medical buildings. Background Technology

[0002] Building Information Modeling (BIM) is a crucial technological support for the modern construction industry. Pharmaceutical buildings, especially clusters of pharmaceutical buildings, are far more complex to design and construct than ordinary buildings, not only due to their larger size and numerous functional zones, but also because of their stringent environmental requirements and strict medical and health regulations and industry standards. Traditional BIM model building largely relies on manual operations by BIM engineers and architects. For clusters of pharmaceutical buildings, this manual modeling method is extremely labor-intensive, difficult to meet the demands of rapid construction, and prone to errors and omissions due to structural complexity, requiring frequent changes and rework later on. With the development of artificial intelligence (AI), applying AI agents to BIM model building has become a research hotspot. For example, Jang et al. proposed using natural language models in BIM, and Du et al. proposed Text2BIM, which utilizes large models and intelligent agents to construct BIM models. However, clustered medical buildings have complex functions and require compliance with numerous standards. Existing BIM generation methods based on large models cannot handle such complex thinking and analysis processes, and are prone to generating BIM generation code that seems reasonable but is actually unusable or non-compliant. There is an urgent need for a BIM generation method that can integrate knowledge in the field of medical buildings and has stronger planning, reasoning, generation and correction capabilities. Summary of the Invention

[0003] To address the above problems, in a first aspect of the present invention, a method for rapidly constructing BIM models of clustered medical buildings is provided, the method comprising the following steps:

[0004] A mind tree is constructed based on design requirements. The attributes of the nodes include multiple attributes that represent cross-building function associations and medical standard requirements. The mind tree is pruned using a medical prototype library and compliance constraints. The planning agent generates a Directed Acyclic Graph (DAG) based on the leaf nodes of the mind tree. The interdependence level of the tasks in the DAG is calculated as the edges of the DAG based on the cluster linkage of the tasks in the DAG to obtain a weighted DAG. The planning agent determines the task priority and resource allocation based on the weighted DAG and generates an execution sequence.

[0005] The BIM code generation agent obtains the execution sequence. For each leaf node task, the generation of BIM component code is divided into multiple stages. At the end of each stage, the confidence of the generated code in the current stage is calculated based on the multi-view node attributes and the preset medical geometry rules. When the confidence is lower than the threshold calculated based on the current cluster unit complexity, the reasoning path is adjusted and the thought chain process of the corresponding component is restarted.

[0006] When the confidence score of the BIM code generation agent report continues to be lower than the preset lower limit, or when a specific key task cluster in the weighted DAG is completed, the inspection and correction agent performs inspection and correction on the generated BIM component code set; gather all the corrected BIM component codes, and construct the BIM model of the clustered medical building.

[0007] Preferably, the pruning of the mind tree using a pharmaceutical prototype library and compliance constraints specifically involves:

[0008] Traverse each node of the mind tree and extract the design task information corresponding to the node;

[0009] The matching degree of the design task information is calculated with the functions, spatial topology relationships and essential components of each standard medical functional unit stored in the medical prototype library. When the matching degree of a node is lower than the preset matching threshold, the node and all its child nodes are removed.

[0010] For each remaining node, extract the medical regulatory requirements contained in the node attributes and check their compliance with the current national and local medical building design regulations stored in the compliance constraint database. If the number of violations of any node exceeds the preset number of violations, remove the node and all its child nodes.

[0011] Preferably, the planning agent generates a DAG based on the leaf node tasks of the thought tree, specifically as follows:

[0012] Define a BIM modeling task unit for each leaf node of the mind tree; for any two BIM modeling task units A and B, when the completion of task A is a prerequisite for the start of task B, or the output data of task A is the input data of task B, then create a directed edge from task A to task B in the DAG.

[0013] Repeat the dependency analysis process described above until all dependencies between task units in the leaf nodes are identified and represented as directed edges in the DAG, thus obtaining the DAG.

[0014] Preferably, the step of calculating the interdependence level of the edges of the DAG based on the cluster linkage of tasks in the DAG to obtain a weighted DAG specifically involves:

[0015] For each task dependency edge in the DAG, identify the medical cluster units corresponding to task X and task Y, where task X depends on task Y; obtain a preset cluster linkage rule base, which defines the linkage influence factor between different combinations of medical cluster units, and the influence factor is calculated based on the linkage type and linkage tightness;

[0016] Based on the complexity of task Y and the influence factor of the cluster linkage rule base between task X and task Y, the basic score of mutual dependence of dependent edges is calculated; the basic score of mutual dependence of all dependent edges is normalized, and each dependent edge is converted into a mutual dependence key level according to four preset fixed percentile intervals to obtain the weighted DAG.

[0017] Preferably, the step of calculating the confidence level of the generated code at the current stage based on the multi-view node attributes and preset medical geometry rules specifically involves:

[0018] Extract the preset node attributes of the mind tree nodes corresponding to the BIM components completed in the current BIM component code generation stage, compare the actual geometric parameters of the generated BIM components with the medical standard indicators in the preset node attributes, and calculate the standard penalty value.

[0019] Based on the preset medical geometric rule library, a geometric penalty term is obtained based on whether the generated BIM components have geometric logic conflicts or are unreasonable; the confidence level is calculated based on 1 - standard penalty term - geometric penalty term.

[0020] Preferably, the threshold calculated based on the current cluster unit complexity is specifically:

[0021] According to the formula C=(N) r ×W r )+(N e ×W e )+(N i ×W i Calculate the complexity of the cluster unit to which the BIM component belongs, where N r W represents the number of rooms within a cluster unit. r N represents the room weights. e W represents the number of medical devices within a cluster unit. e N represents the device weight. i W represents the number of interfaces or connection pipelines between this cluster unit and other cluster units. i For interconnection weight;

[0022] The threshold is determined using a piecewise linear function based on the complexity C.

[0023] In a second aspect of the present invention, a BIM model building system for rapid construction of clustered medical buildings is provided, characterized in that the system comprises the following modules:

[0024] A planning agent is used to construct a thought tree based on design requirements. The node attributes include multiple attributes that represent cross-building function associations and medical standard requirements. The thought tree is pruned using a medical prototype library and compliance constraints. A Directed Acyclic Graph (DAG) is generated based on the leaf nodes of the thought tree. The interdependence level of the DAG edges is calculated based on the cluster linkage of tasks in the DAG to obtain a weighted DAG. The planning agent determines task priorities and resource allocation based on the weighted DAG and generates an execution sequence.

[0025] The BIM code generation agent is used to obtain the execution sequence. For each leaf node task, the generation of BIM component code is divided into multiple stages. At the end of each stage, the confidence of the generated code in the current stage is calculated based on the multi-view node attributes and the preset medical geometry rules. When the confidence is lower than the threshold calculated based on the current cluster unit complexity, the reasoning path is adjusted and the thought chain process of the corresponding component is restarted.

[0026] The inspection and correction agent is used to perform inspection and correction on the generated BIM component code set when the confidence score of the BIM code generation agent report continues to be lower than the preset lower limit, or when a specific key task cluster in the weighted DAG is completed; all corrected BIM component codes are collected to construct a BIM model of the clustered medical building.

[0027] Preferably, the pruning of the mind tree using a pharmaceutical prototype library and compliance constraints specifically involves:

[0028] Traverse each node of the mind tree and extract the design task information corresponding to the node;

[0029] The matching degree of the design task information is calculated with the functions, spatial topology relationships and essential components of each standard medical functional unit stored in the medical prototype library. When the matching degree of a node is lower than the preset matching threshold, the node and all its child nodes are removed.

[0030] For each remaining node, extract the medical regulatory requirements contained in the node attributes and check their compliance with the current national and local medical building design regulations stored in the compliance constraint database. If the number of violations of any node exceeds the preset number of violations, remove the node and all its child nodes.

[0031] Preferably, the leaf node-based task generation of the DAG based on the mind tree specifically involves:

[0032] Define a BIM modeling task unit for each leaf node of the mind tree; for any two BIM modeling task units A and B, when the completion of task A is a prerequisite for the start of task B, or the output data of task A is the input data of task B, then create a directed edge from task A to task B in the DAG.

[0033] Repeat the dependency analysis process described above until all dependencies between task units in the leaf nodes are identified and represented as directed edges in the DAG, thus obtaining the DAG.

[0034] Preferably, the step of calculating the interdependence level of the edges of the DAG based on the cluster linkage of tasks in the DAG to obtain a weighted DAG specifically involves:

[0035] For each task dependency edge in the DAG, identify the medical cluster units corresponding to task X and task Y, where task X depends on task Y; obtain a preset cluster linkage rule base, which defines the linkage influence factor between different combinations of medical cluster units, and the influence factor is calculated based on the linkage type and linkage tightness;

[0036] Based on the complexity of task Y and the influence factor of the cluster linkage rule base between task X and task Y, the basic score of mutual dependence of dependent edges is calculated; the basic score of mutual dependence of all dependent edges is normalized, and each dependent edge is converted into a mutual dependence key level according to four preset fixed percentile intervals to obtain the weighted DAG.

[0037] Preferably, the step of calculating the confidence level of the generated code at the current stage based on the multi-view node attributes and preset medical geometry rules specifically involves:

[0038] Extract the preset node attributes of the mind tree nodes corresponding to the BIM components completed in the current BIM component code generation stage, compare the actual geometric parameters of the generated BIM components with the medical standard indicators in the preset node attributes, and calculate the standard penalty value.

[0039] Based on the preset medical geometric rule library, a geometric penalty term is obtained based on whether the generated BIM components have geometric logic conflicts or are unreasonable; the confidence level is calculated based on 1 - standard penalty term - geometric penalty term.

[0040] Preferably, the threshold calculated based on the current cluster unit complexity is specifically:

[0041] According to the formula C=(N) r ×W r )+(N e ×W e )+(N i ×W iCalculate the complexity of the cluster unit to which the BIM component belongs, where N r W represents the number of rooms within a cluster unit. r N represents the room weights. e W represents the number of medical devices within a cluster unit. e N represents the device weight. i W represents the number of interfaces or connection pipelines between this cluster unit and other cluster units. i For interconnection weight;

[0042] The threshold is determined using a piecewise linear function based on the complexity C.

[0043] This invention assigns node attributes representing cross-building functional relationships and medical regulatory requirements to thought tree nodes, and uses a medical prototype library and compliance constraints for pruning, ensuring that the unique functional requirements and regulations of medical buildings are incorporated from the early stages. Based on the confidence threshold adjusted by the complexity of cluster units, the modeling of medical areas is more rigorous, while simpler areas can be quickly approved, improving the accuracy and speed of code generation. Moreover, by calculating the mutual dependency level for task dependency edges in the DAG, modeling tasks that are important to the overall function and safety of the cluster can be prioritized, and real-time detection based on confidence scores avoids the inefficiency of global checks. Attached Figure Description

[0044] Figure 1 This is a flowchart of Example 1;

[0045] Figure 2 This is a diagram of a mind tree;

[0046] Figure 3 This is a schematic diagram of a DAG;

[0047] Figure 4 This is a schematic diagram illustrating the relationship between the threshold and the complexity of the cluster unit. Detailed Implementation

[0048] In the embodiments of this invention, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a specific manner to facilitate understanding.

[0049] It is understood that the term "embodiment" used throughout the specification means that a specific feature, structure, or characteristic related to an embodiment is included in at least one embodiment of this application. Therefore, various embodiments throughout the specification do not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It is understood that in the various embodiments of this application, the sequence number of each process does not imply the order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0050] In this invention, unless otherwise specified, the same or similar parts between the various embodiments can be referred to each other. In the various embodiments of this invention, and in the various implementation methods / methods / implementations within each embodiment, unless otherwise specified or logically conflicting, the terminology and / or descriptions between different embodiments and between the various implementation methods / methods / implementations within each embodiment are consistent and can be mutually referenced. The technical features in different embodiments and the various implementation methods / methods / implementations within each embodiment can be combined according to their inherent logical relationships to form new embodiments, implementation methods, methods, or implementation approaches. The embodiments described below do not constitute a limitation on the scope of protection of this application.

[0051] Figure 1 A flowchart of the first embodiment of the present invention is shown. Figure 1 The method for rapid construction of BIM models for clustered medical buildings includes the following steps:

[0052] S1. Construct a mind tree based on design requirements. The attributes of the nodes include multiple attributes that represent cross-building function associations and medical standard requirements. Use a medical prototype library and compliance constraints to prune the mind tree. The planning agent generates a DAG based on the leaf nodes of the mind tree. Calculate the interdependence level of the DAG edges based on the cluster linkage of tasks in the DAG to obtain a weighted DAG. The planning agent determines task priorities and resource allocation based on the weighted DAG and generates an execution sequence.

[0053] Build a mind tree based on design requirements, such as Figure 2As shown. The nodes of the mind tree not only define the hierarchy of the design task, but also include attributes such as the accessibility requirements of medical processes to different building areas and the special spatial specifications of specific medical equipment. For example, the initial design requirement is "to build a comprehensive medical cluster including outpatient, emergency and inpatient departments". The requirement is decomposed into the planning level of floors, departments and even specific rooms. Each node has cross-building function and specification attributes. The mind tree may contain solutions that do not conform to the characteristics of medical buildings or do not meet specific specifications. It is necessary to remove non-compliant and infeasible branches of the mind tree. In one embodiment, the medical prototype library and compliance constraints are constructed as an OWL ontology. The inference engine is used to identify and remove nodes and their branches in the mind tree that have logical conflicts with the medical building ontology model or cannot meet semantic constraints. In another embodiment, the pruning of the mind tree using the medical prototype library and compliance constraints is specifically as follows:

[0054] Traverse each node of the mind tree and extract the design task information corresponding to the node;

[0055] The matching degree of the design task information is calculated with the functions, spatial topology relationships and essential components of each standard medical functional unit stored in the medical prototype library. When the matching degree of a node is lower than the preset matching threshold, the node and all its child nodes are removed.

[0056] For each remaining node, extract the medical regulatory requirements contained in the node attributes and check their compliance with the current national and local medical building design regulations stored in the compliance constraint database. If the number of violations of any node exceeds the preset number of violations, remove the node and all its child nodes.

[0057] For example, a node in the mind tree represents an advanced imaging diagnostic center, with a sub-node called a superconducting MRI examination room cluster. The design task information of the superconducting MRI examination room cluster is matched against standard MRI examination units in a medical prototype library. If the prototype library requires an MRI room to include an operating room, equipment room, and shielded room, and the initial plan under the current superconducting MRI examination room cluster node lacks an equipment room, resulting in a match rate below a preset threshold (e.g., 70%), then the superconducting MRI examination room cluster node and all its sub-task nodes will be removed. If the match passes, the node's medical compliance requirements are then checked, such as national regulations requiring MRI rooms to meet specific electromagnetic shielding standards. If the compliance constraint database shows that the shielding materials used in the current node's plan do not conform to the latest national standards, and the total number of violations exceeds the violation limit (e.g., one critical violation), the node will also be removed.

[0058] The planning agent generates a Directed Acyclic Graph (DAG) based on the leaf nodes of a thought tree, such as... Figure 3As shown. In one embodiment, a DAG that considers the coordination of construction phases is generated by iteratively adjusting the task order to minimize the estimated potential conflict points of parallel construction by different disciplines within the cluster. In another embodiment, the planning agent generates the DAG based on the leaf node tasks of the thought tree, specifically:

[0059] Define a BIM modeling task unit for each leaf node of the mind tree; for any two BIM modeling task units A and B, when the completion of task A is a prerequisite for the start of task B, or the output data of task A is the input data of task B, then create a directed edge from task A to task B in the DAG.

[0060] Repeat the dependency analysis process described above until all dependencies between task units in the leaf nodes are identified and represented as directed edges in the DAG, thus obtaining the DAG.

[0061] Specifically, assume the leaf nodes of the mind tree include "Installation of the main air conditioning system pipe for Area A of Operating Room 1 in Building 1" and "Interior finishing of Operating Room 1 in Area A of Operating Room 1 in Building 1". Since the installation of the main air conditioning system pipe is a prerequisite for the interior finishing of the operating room, that is, the completion of the pipe installation task is a prerequisite for the start of the interior finishing task, a directed edge is created in the DAG from "Installation of the main air conditioning system pipe for Area A of Operating Room 1 in Building 1" to "Interior finishing of Operating Room 1 in Area A of Operating Room 1 in Building 1". This process is repeated for all leaf nodes; for example, "Laying of anesthetic gas pipelines" will also point to "Interior finishing of Operating Room 1 in Area A of Operating Room 1 in Building 1".

[0062] To enable the planning agent to make better decisions that better meet project requirements, a weighted DAG is obtained by further calculating the interdependence level of edge computation in the DAG based on the clustering and linkage of tasks in the DAG. In one embodiment, a graph neural network model is used. Specifically, the attributes of each task in the DAG are input, and the GNN model outputs the criticality level of each dependent edge. The higher the level, the greater the impact on the overall progress or functional implementation of the cluster. The planning agent uses the critical path method combined with the criticality level for scheduling, prioritizing the execution of tasks located on the critical path with the highest criticality level, and allocating more virtual computing resources to these tasks. In another embodiment, the step of obtaining a weighted DAG based on the interdependence level of edge computation in the DAG based on the clustering and linkage of tasks in the DAG is specifically as follows:

[0063] For each task dependency edge in the DAG, identify the medical cluster units corresponding to task X and task Y, where task X depends on task Y; obtain a preset cluster linkage rule base, which defines the linkage influence factor between different combinations of medical cluster units, and the influence factor is calculated based on the linkage type and linkage tightness;

[0064] Based on the complexity of task Y and the influence factor of the cluster linkage rule base between task X and task Y, the basic score of mutual dependence of dependent edges is calculated; the basic score of mutual dependence of all dependent edges is normalized, and each dependent edge is converted into a mutual dependence key level according to four preset fixed percentile intervals to obtain the weighted DAG.

[0065] Suppose there's a dependent edge in a Directed Acyclic Graph (DAG). Upstream task Y is "Construction of the sterile supplies storage area in the central supply room," and downstream task X is "Processing of the clean corridor connection in the operating room." Tasks X and Y correspond to the medical cluster units "Central Supply Room" and "Operating Department," respectively. Comparing with the cluster's linkage rule base, we find that "transfer of sterile supplies from the central supply room to the operating department" belongs to the "medical process continuity" linkage type and is a "direct physical connection," such as through a pass-through window or dedicated channel. The linkage impact factor is 0.9. If the complexity score of task Y is 50, then the basic mutual dependency score of this edge is 45. After normalizing the basic scores of all dependent edges, if this score of 45 falls within the highest 25% range, then the mutual dependency criticality level of this edge is Level 1.

[0066] S2, the BIM code generation agent obtains the execution sequence. For each leaf node task, the generation of BIM component code is divided into multiple stages. At the end of each stage, the confidence of the generated code in the current stage is calculated based on the multi-view node attributes and the preset medical geometry rules. When the confidence is lower than the threshold calculated based on the current cluster unit complexity, the reasoning path is adjusted and the thought chain process of the corresponding component is restarted.

[0067] When performing specific modeling tasks, the BIM code generation agent breaks down the generation process of complex medical components into several sub-stages. After each sub-stage is completed, it evaluates whether the currently generated local code / model conforms to the various requirements initially defined in the mind tree nodes and the general geometric logic of medical buildings. The evaluation result is specifically a confidence score. If the confidence score is low after a certain sub-stage is completed, it indicates that there is a deviation in the current generation approach. The agent then adjusts its internal thought chain, such as changing certain design parameters or generating alternative generation strategies, and then regenerates the code for the problematic component or sub-step. The more complex the component, the higher the confidence score requirement.

[0068] In an optional embodiment, calculating the confidence level of the generated code at the current stage based on the multi-view node attributes and preset medical geometry rules specifically involves:

[0069] Extract the preset node attributes of the mind tree nodes corresponding to the BIM components completed in the current BIM component code generation stage, compare the actual geometric parameters of the generated BIM components with the medical standard indicators in the preset node attributes, and calculate the standard penalty value.

[0070] Based on the preset medical geometric rule base, a geometric penalty term is obtained based on whether the generated BIM components have geometric logic conflicts or are unreasonable; the confidence level is calculated based on 1 - standard penalty term - geometric penalty term. In another embodiment, weights are added to the standard penalty term and the geometric penalty term respectively.

[0071] Assume the multiple attributes of the mind tree node corresponding to the exhaust duct specify "the duct must maintain a pressure difference of -15Pa", "the material is 304 stainless steel", and "the diameter is 300mm". The actual generated duct model parameters are a pressure difference design of -14Pa, which violates one rule, increasing the penalty by 0.1. The material is galvanized steel, which also violates one rule, increasing the penalty by another 0.1, for a total of 0.2. The diameter is 300mm, which meets the requirements, so the penalty does not increase. If the medical geometry rule library requires "the duct must not collide with structural beams", and the generated duct collides with a load-bearing beam, the geometry penalty increases by 0.2, and the final confidence level of the exhaust duct code is 0.6. In one embodiment, if the final calculated confidence level is negative, the confidence level is set to 0. Of course, a negative confidence level can also be used; this invention does not specifically limit this. In one embodiment, cluster units correspond to nodes in the DAG, and can be divided into different cluster units according to different granularities. For example, the first phase of the outpatient building can be considered a cluster unit, and the energy supply center can also be considered a cluster unit.

[0072] For complex units, requiring very high confidence levels may cause the BIM code generation agent to fail to meet standards, leading to frequent restarts of the generation process and even making the generation of some complex units impractical. To ensure the smooth progress of the process, in an optional embodiment, the threshold calculated based on the complexity of the current cluster units is specifically:

[0073] According to the formula C=(N) r ×W r )+(N e ×W e )+(N i ×W i Calculate the complexity of the cluster unit to which the BIM component belongs, where N r W represents the number of rooms within a cluster unit. r N represents the room weights. e W represents the number of medical devices within a cluster unit. e N represents the device weight. iis the number of interfaces or the number of connecting pipelines between this cluster unit and other cluster units, W i is the interconnection weight;

[0074] Determine the threshold according to the complexity C using a piecewise linear function.

[0075] For example, to calculate the confidence threshold for code generation of BIM components in a "Neonatal Intensive Care Unit (NICU) cluster unit", calculate the complexity C of the NICU cluster unit. If there are 5 rooms in the NICU with a room quantity weight of 0.3, 10 medical devices with a device weight of 0.4, 4 interfaces with other cluster units, and an interconnection weight of 0.3, then C is 6.7. Further, the preset piecewise linear function is: when C ≤ 10, T = 0.85; when 10 < C ≤ 30, T = 0.85 - ((C - 10) / 20) × 0.10. Since C = 6.7 is less than 10, the confidence threshold T for BIM component generation in this NICU cluster unit is 0.85. Similarly, if the complexity C of a simple "General Outpatient Clinic" cluster unit is 12, then its threshold T is 0.84. Figure 4 Shows a schematic diagram of the threshold and the complexity of the cluster unit. Those skilled in the art should know that the above data are only for illustrating the solution of the present invention and do not constitute a limitation to the present invention.

[0076] S3. When the confidence score reported by the BIM code generation agent continuously falls below the preset lower limit, or when a specific critical task cluster in the weighted DAG is completed, the inspection and correction agent performs inspection and correction on the generated BIM component code set; collect all the corrected BIM component codes and construct a BIM model of the cluster-style medical building.

[0077] If, after the BIM code generation agent has made multiple adjustments and restarts to the thought chain, the confidence of a certain BIM component still cannot reach the lowest acceptable standard, which is beyond the correction ability range of the generation agent itself, the inspection and correction agent will intervene and try to solve this problem. When a predefined set of interrelated and important modeling tasks for the cluster function in the weighted DAG is completed, a stage task is completed. A comprehensive inspection is performed on these generated BIM component code sets. At this time, not only the correctness of individual components is concerned, but also their cooperation and consistency are checked, and then the places that need to be modified are repaired. The verified and optimized BIM component codes are integrated into a complete BIM model. In one embodiment, the agent of the present invention utilizes the understanding, reasoning, and generation capabilities of a large model. Different agents complete corresponding tasks and work together by invoking the large model to jointly construct the BIM model of the cluster-style medical building.

[0078] Specific embodiment two provides a BIM model building system for rapid construction of clustered medical buildings, the system including the following modules:

[0079] A planning agent is used to construct a thought tree based on design requirements. The node attributes include multiple attributes that represent cross-building function associations and medical standard requirements. The thought tree is pruned using a medical prototype library and compliance constraints. A Directed Acyclic Graph (DAG) is generated based on the leaf nodes of the thought tree. The interdependence level of the DAG edges is calculated based on the cluster linkage of tasks in the DAG to obtain a weighted DAG. The planning agent determines task priorities and resource allocation based on the weighted DAG and generates an execution sequence.

[0080] The BIM code generation agent is used to obtain the execution sequence. For each leaf node task, the generation of BIM component code is divided into multiple stages. At the end of each stage, the confidence of the generated code in the current stage is calculated based on the multi-view node attributes and the preset medical geometry rules. When the confidence is lower than the threshold calculated based on the current cluster unit complexity, the reasoning path is adjusted and the thought chain process of the corresponding component is restarted.

[0081] The inspection and correction agent is used to perform inspection and correction on the generated BIM component code set when the confidence score of the BIM code generation agent report continues to be lower than the preset lower limit, or when a specific key task cluster in the weighted DAG is completed; all corrected BIM component codes are collected to construct a BIM model of the clustered medical building.

[0082] Preferably, the pruning of the mind tree using a pharmaceutical prototype library and compliance constraints specifically involves:

[0083] Traverse each node of the mind tree and extract the design task information corresponding to the node;

[0084] The matching degree of the design task information is calculated with the functions, spatial topology relationships and essential components of each standard medical functional unit stored in the medical prototype library. When the matching degree of a node is lower than the preset matching threshold, the node and all its child nodes are removed.

[0085] For each remaining node, extract the medical regulatory requirements contained in the node attributes and check their compliance with the current national and local medical building design regulations stored in the compliance constraint database. If the number of violations of any node exceeds the preset number of violations, remove the node and all its child nodes.

[0086] Preferably, the leaf node-based task generation of the DAG based on the mind tree specifically involves:

[0087] Define a BIM modeling task unit for each leaf node of the mind tree; for any two BIM modeling task units A and B, when the completion of task A is a prerequisite for the start of task B, or the output data of task A is the input data of task B, then create a directed edge from task A to task B in the DAG.

[0088] Repeat the dependency analysis process described above until all dependencies between task units in the leaf nodes are identified and represented as directed edges in the DAG, thus obtaining the DAG.

[0089] Preferably, the step of calculating the interdependence level of the edges of the DAG based on the cluster linkage of tasks in the DAG to obtain a weighted DAG specifically involves:

[0090] For each task dependency edge in the DAG, identify the medical cluster units corresponding to task X and task Y, where task X depends on task Y; obtain a preset cluster linkage rule base, which defines the linkage influence factor between different combinations of medical cluster units, and the influence factor is calculated based on the linkage type and linkage tightness;

[0091] Based on the complexity of task Y and the influence factor of the cluster linkage rule base between task X and task Y, the basic score of mutual dependence of dependent edges is calculated; the basic score of mutual dependence of all dependent edges is normalized, and each dependent edge is converted into a mutual dependence key level according to four preset fixed percentile intervals to obtain the weighted DAG.

[0092] Preferably, the step of calculating the confidence level of the generated code at the current stage based on the multi-view node attributes and preset medical geometry rules specifically involves:

[0093] Extract the preset node attributes of the mind tree nodes corresponding to the BIM components completed in the current BIM component code generation stage, compare the actual geometric parameters of the generated BIM components with the medical standard indicators in the preset node attributes, and calculate the standard penalty value.

[0094] Based on the preset medical geometric rule library, a geometric penalty term is obtained based on whether the generated BIM components have geometric logic conflicts or are unreasonable; the confidence level is calculated based on 1 - standard penalty term - geometric penalty term.

[0095] Preferably, the threshold calculated based on the current cluster unit complexity is specifically:

[0096] According to the formula C=(N) r ×W r )+(N e ×W e )+(N i ×W iCalculate the complexity of the cluster unit to which the BIM component belongs, where N r W represents the number of rooms within a cluster unit. r N represents the room weights. e W represents the number of medical devices within a cluster unit. e N represents the device weight. i W represents the number of interfaces or connection pipelines between this cluster unit and other cluster units. i For interconnection weight;

[0097] The threshold is determined using a piecewise linear function based on the complexity C.

[0098] The above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state disk (SSD)).

[0099] The steps of the methods or algorithms described in the embodiments of this application can be directly embedded in hardware, software units executed by a processor, or a combination of both. The software units can be stored in RAM, flash memory, ROM, EPROM, EEPROM, registers, hard disks, removable disks, CD-ROMs, or any other form of storage medium in the art. Exemplarily, the storage medium can be connected to the processor so that the processor can read information from and write information to the storage medium. Optionally, the storage medium can also be integrated into the processor. The processor and storage medium can be housed in an ASIC.

[0100] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0101] Although this application has been described in conjunction with specific features and embodiments, it is obvious that various modifications and combinations can be made thereto without departing from the spirit and scope of this application. Accordingly, this specification and drawings are merely illustrative descriptions of the application as defined by the appended claims, and are considered to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Thus, if such modifications and modifications fall within the scope of the claims and their equivalents, this application is also intended to include such modifications and modifications.

Claims

1. A method for rapidly constructing BIM models of clustered medical buildings, characterized in that, The method includes the following steps: A mind tree is constructed based on design requirements. The attributes of the nodes include multiple attributes that represent cross-building function associations and medical standard requirements. The mind tree is pruned using a medical prototype library and compliance constraints. The planning agent generates a Directed Acyclic Graph (DAG) based on the leaf nodes of the mind tree. The interdependence level of the tasks in the DAG is calculated as the edges of the DAG based on the cluster linkage of the tasks in the DAG to obtain a weighted DAG. The planning agent determines the task priority and resource allocation based on the weighted DAG and generates an execution sequence. The BIM code generation agent obtains the execution sequence. For each leaf node task, the generation of BIM component code is divided into multiple stages. At the end of each stage, the confidence of the generated code in the current stage is calculated based on the multi-view node attributes and the preset medical geometry rules. When the confidence is lower than the threshold calculated based on the current cluster unit complexity, the reasoning path is adjusted and the thought chain process of the corresponding component is restarted. When the confidence score of the BIM code generation agent report continues to be lower than the preset lower limit, or when the key task cluster in the weighted DAG is completed, the inspection and correction agent performs inspection and correction on the generated BIM component code set; gather all the corrected BIM component codes and construct the BIM model of the clustered medical building. The pruning of the thought tree using a pharmaceutical prototype library and compliance constraints specifically involves: Traverse each node of the mind tree and extract the design task information corresponding to the node; The matching degree of the design task information is calculated with the functions, spatial topology relationships and essential components of each standard medical functional unit stored in the medical prototype library. When the matching degree of a node is lower than the preset matching threshold, the node and all its child nodes are removed. For each remaining node, extract the medical regulatory requirements contained in the node attributes and check their compliance with the current national and local medical building design regulations stored in the compliance constraint database. If the number of violations of any node exceeds the preset number of violations, remove the node and all its child nodes. The confidence level of the generated code at the current stage is calculated based on the multi-view node attributes and preset medical geometric rules, specifically as follows: Extract the preset node attributes of the mind tree nodes corresponding to the BIM components completed in the current BIM component code generation stage, compare the actual geometric parameters of the generated BIM components with the medical standard indicators in the preset node attributes, and calculate the standard penalty value. Based on the preset medical geometric rule library, a geometric penalty value is obtained based on whether the generated BIM components have geometric logic conflicts or are unreasonable; the confidence level is calculated accordingly.

2. The method as described in claim 1, characterized in that, The planning agent generates a Directed Acyclic Graph (DAG) based on the leaf node tasks of the thought tree, specifically as follows: S1, define a BIM modeling task unit for each leaf node of the mind tree; for any two BIM modeling task units A and B, when the completion of task A is a prerequisite for the start of task B, or when the output data of task A is the input data of task B, then create a directed edge from task A to task B in the DAG. Repeat S1 until all dependencies between leaf node task units are identified and represented as directed edges in the DAG, thus obtaining the DAG.

3. The method as described in claim 1, characterized in that, The process of calculating the interdependence level of edges in a DAG based on the clustered linkage of tasks within the DAG to obtain a weighted DAG is as follows: For each task dependency edge in the DAG, identify the medical cluster units corresponding to task X and task Y, where task X depends on task Y; obtain a preset cluster linkage rule base, which defines the linkage influence factor between different combinations of medical cluster units, and the influence factor is calculated based on the linkage type and linkage tightness; Based on the complexity of task Y and the influence factor of the cluster linkage rule base between task X and task Y, the basic score of mutual dependence of dependent edges is calculated; the basic score of mutual dependence of all dependent edges is normalized, and each dependent edge is converted into a mutual dependence key level according to four preset fixed percentile intervals to obtain the weighted DAG.

4. The method as described in claim 1, characterized in that, The threshold calculated based on the current cluster unit complexity is specifically as follows: The complexity of the cluster unit to which the BIM component belongs is calculated according to the formula, where is the number of rooms in the cluster unit, is the room weight, is the number of medical devices in the cluster unit, is the device weight, is the number of interfaces or connecting pipelines between the cluster unit and other cluster units, and is the interconnection weight. The threshold is determined using a piecewise linear function based on the complexity C.

5. A rapid construction BIM model building system for clustered medical buildings, characterized in that, The system includes the following modules: A planning agent is used to construct a thought tree based on design requirements. The node attributes include multiple attributes that represent cross-building function associations and medical standard requirements. The thought tree is pruned using a medical prototype library and compliance constraints. A Directed Acyclic Graph (DAG) is generated based on the leaf nodes of the thought tree. The interdependence level of the DAG edges is calculated based on the cluster linkage of tasks in the DAG to obtain a weighted DAG. The planning agent determines task priorities and resource allocation based on the weighted DAG and generates an execution sequence. The BIM code generation agent is used to obtain the execution sequence. For each leaf node task, the generation of BIM component code is divided into multiple stages. At the end of each stage, the confidence of the generated code in the current stage is calculated based on the multi-view node attributes and the preset medical geometry rules. When the confidence is lower than the threshold calculated based on the current cluster unit complexity, the reasoning path is adjusted and the thought chain process of the corresponding component is restarted. The checking and correction agent is used to check and correct the generated BIM component code set when the confidence score of the BIM code generation agent report is continuously lower than the preset lower limit, or when the key task cluster in the weighted DAG is completed; all corrected BIM component codes are collected to construct a BIM model of the clustered medical building. The pruning of the thought tree using a pharmaceutical prototype library and compliance constraints specifically involves: Traverse each node of the mind tree and extract the design task information corresponding to the node; The matching degree of the design task information is calculated with the functions, spatial topology relationships and essential components of each standard medical functional unit stored in the medical prototype library. When the matching degree of a node is lower than the preset matching threshold, the node and all its child nodes are removed. For each remaining node, extract the medical regulatory requirements contained in the node attributes and check their compliance with the current national and local medical building design regulations stored in the compliance constraint database. If the number of violations of any node exceeds the preset number of violations, remove the node and all its child nodes. The confidence level of the generated code at the current stage is calculated based on the multi-view node attributes and preset medical geometric rules, specifically as follows: Extract the preset node attributes of the mind tree nodes corresponding to the BIM components completed in the current BIM component code generation stage, compare the actual geometric parameters of the generated BIM components with the medical standard indicators in the preset node attributes, and calculate the standard penalty value. Based on the preset medical geometric rule library, a geometric penalty value is obtained based on whether the generated BIM components have geometric logic conflicts or are unreasonable. The confidence level is calculated using the formula "1 - normative penalty value - geometric penalty value".

6. The system as described in claim 5, characterized in that, The leaf node-based task generation of the DAG based on the mind tree specifically involves: S1, Define a BIM modeling task unit for each leaf node of the mind tree; For any two BIM modeling task units A and B, when the completion of task A is a prerequisite for the start of task B, or when the output data of task A is the input data of task B, then create a directed edge from task A to task B in the DAG. Repeat S1 until all dependencies between leaf node task units are identified and represented as directed edges in the DAG, thus obtaining the DAG.

7. The system as described in claim 5, characterized in that, The process of calculating the interdependence level of edges in a DAG based on the clustered linkage of tasks within the DAG to obtain a weighted DAG is as follows: For each task dependency edge in the DAG, identify the medical cluster units corresponding to task X and task Y, where task X depends on task Y; obtain a preset cluster linkage rule base, which defines the linkage influence factor between different combinations of medical cluster units, and the influence factor is calculated based on the linkage type and linkage tightness; Based on the complexity of task Y and the influence factor of the cluster linkage rule base between task X and task Y, the basic score of mutual dependence of dependent edges is calculated; the basic score of mutual dependence of all dependent edges is normalized, and each dependent edge is converted into a mutual dependence key level according to four preset fixed percentile intervals to obtain the weighted DAG.

Images