Multi-disciplinary conflict decision optimization method based on bim model

Abstract

The application discloses a multi-specialty conflict decision optimization method based on a BIM model, which comprises the following steps: constructing a constraint library suitable for a target project, binding the structured parameters in the constraint library with corresponding components in a BIM model, then combining automatic detection of conflict components, quantitative ordering of specialty constraint weights and secondary ordering rule optimization, and iteratively generating multiple sets of candidate schemes in the range of a constraint threshold with priority as the guide, and finally updating the compliance verification and full-link responsibility traceability chain of the final scheme, effectively solving the pain points of non-uniform constraint parameters, subjective priority determination, lack of constraint guidance in scheme decision and difficulty in defining decision responsibility in traditional BIM multi-specialty collaboration conflicts, significantly improving the scientificity and accuracy of conflict decision, and ensuring that high-priority specialty hard constraints are satisfied first, and the decision-making process is traceable and compliance can be verified.

Description

Technical Field

[0001] This invention relates to the field of BIM system management technology, and in particular to a multi-disciplinary conflict decision optimization method based on a BIM model. Background Technology

[0002] BIM technology, as a core digital tool for multi-disciplinary collaboration in building engineering, has been widely applied throughout the design and construction process. Based on the BIM platform, each discipline (architecture, structure, MEP, etc.) builds its own model, integrates these models to detect conflicts, and then negotiates modifications to address issues such as spatial collisions and parameter discrepancies, ultimately forming a unified construction basis. However, in practical applications, this process often stalls due to the "hard constraints" of each discipline. Architectural professionals must adhere to requirements such as ceiling height and planning regulations; structural professionals are constrained by load-bearing safety and seismic standards; and MEP professionals must ensure system energy efficiency and pipeline compliance. Any modification to the design by any party may exceed its own hard constraints, leading to compliance risks or additional costs. Consequently, when conflicts arise, each discipline is generally "unwilling and afraid to make changes," significantly reducing collaboration efficiency and even causing delays and rework.

[0003] While existing BIM models can achieve conflict detection and model integration, they fail to provide effective solutions for collaborative conflicts under hard constraints. The core reasons are as follows:

[0004] First, BIM models can only identify geometric conflicts and cannot digitally represent the hard constraints of various disciplines (such as structural load-bearing limits and mechanical current thresholds). This makes it difficult to quickly determine the root cause and extent of the conflict, requiring repeated manual verification and resulting in extremely high communication costs. Second, it lacks a hard constraint priority assessment mechanism, and conflict resolution relies solely on manual negotiation. There is no scientific basis to support compromise decisions between disciplines, which can easily lead to deadlocks where everyone sticks to their own opinions, ultimately requiring approval from high-level managers. Third, it lacks intelligent adaptation capabilities based on hard constraints, only indicating conflicts but failing to generate feasible modification solutions. Manual adjustments can easily trigger new constraint conflicts, further exacerbating the "unwillingness to modify." Finally, it has not established accountability and process control for hard constraints, and the compliance of modified constraints lacks verification, leading to a refusal to cooperate between disciplines due to ambiguous responsibilities. Summary of the Invention

[0005] To address the aforementioned issues, this invention proposes a multi-disciplinary conflict decision optimization method based on BIM models, aiming to solve the problems of inconsistent constraint parameters, subjective priority determination, and lack of constraint guidance in scheme decision-making in existing BIM models during multi-disciplinary collaborative conflicts.

[0006] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows:

[0007] A multi-disciplinary conflict decision optimization method based on BIM model includes the following steps:

[0008] A constraint library based on the target project is constructed. Hard constraint parameters for each discipline are configured in the constraint library according to the applicable conditions of the target project. After the hard constraint parameters are standardized, structured parameters are generated, and the structured parameters are bound to the corresponding components in the BIM model.

[0009] The system automatically detects conflicting components in the BIM model, extracts the corresponding structured parameters from the constraint library, quantifies and calculates the professional constraint weight scores corresponding to each conflicting component, and aggregates them to generate a priority ranking table based on professional constraints. If there are professional constraints with the same score in the priority ranking table, the scores are adjusted according to the preset secondary ranking rules until the scores of each professional constraint are different.

[0010] Based on the priority ranking table, at least two candidate schemes are generated iteratively within the threshold range of each professional constraint corresponding to the structured parameters, and the key indicators of each candidate scheme are marked for manual decision-making. After the decision-making, the final scheme is generated. For professional constraints with higher scores in the priority ranking table, the corresponding hard constraint parameters of the professional are given priority when generating candidate schemes. The iterative process has a maximum number of iterations or a scheme convergence threshold as a termination condition.

[0011] The final solution is subjected to compliance verification. After confirming that the unconstrained threshold has been exceeded, the entire chain information of the final solution from conflict detection to decision implementation is recorded, a responsibility traceability chain is generated, and the final solution and the responsibility traceability chain are updated to the BIM model.

[0012] In some implementations, the constraint library includes an industry standard library, a project-customized library, and a professional custom library.

[0013] In some implementations, the standardization of the hard constraint parameters includes: using a natural language processing model to automatically parse the raw text data from industry standard libraries, project customization libraries, and professional custom libraries; mapping the parsed semantic information to the corresponding structured parameters according to the dimensions of the structured parameters; and pushing parameters with a parsing confidence level lower than a preset threshold or mapping failures to manual review, correction, or completion.

[0014] In some implementations, the dimensions of the structured parameters include the major, constraint library type, constraint type, indicator name, constraint threshold, weight factor, and verification rules. The constraint type includes security compliance, project requirements, cost control, and efficiency optimization.

[0015] In some implementations, after the hard constraint parameters are standardized, the following steps are also included: automatically comparing the structured parameters of each professional custom library with the industry standard library and project customization library in the constraint library for compliance. If parameter conflicts are found, the conflicting items are marked and pushed to the corresponding professional designer. After the parameters are corrected and confirmed, they are updated to the professional custom library.

[0016] In some implementations, the automatic detection of conflicting components in the BIM model includes: using an OBB collision detection algorithm to identify overlapping conflicts in the BIM model space, and then using a rule reasoning engine based on the structured parameters to predict potential conflicts that are not spatially overlapping but violate constraint thresholds. The rule base of the rule reasoning engine is built based on industry standards, project requirements and historical conflict data, and the overlapping conflicts and the potential conflicts have different priority calculation weight coefficients.

[0017] In some implementations, if the automatically detected conflicting component is not bound to the corresponding structured parameters, a completion reminder is automatically sent to the corresponding professional designer. If the structured parameters are not completed within a specified time, non-safety compliance constraints are processed according to the default lowest priority defined in the constraint library; safety compliance constraints immediately trigger a manual supervision process, and priority is determined after the parameters are completed. Before the parameters are completed, the generation of solutions for the conflict is suspended.

[0018] In some implementations, the process of generating the candidate solution includes first forming a preliminary solution for resolving the conflict based on an iterative algorithm for selecting the conflict type of the conflicting component, and then optimizing the preliminary solution to generate the candidate solution.

[0019] In some implementations, based on the preliminary scheme, a multi-objective optimization function is constructed with the optimization objectives of maximizing constraint satisfaction rate, minimizing implementation cost, and minimizing project schedule impact. Then, an optimization algorithm suitable for discrete-continuous mixed variable space is used to iteratively search the optimal solution space of the multi-objective optimization function. Each candidate solution corresponds to an optimization variant of the preliminary scheme. During the iteration process, a crowding ranking strategy is used to retain non-dominated solutions. Finally, at least two optimal schemes are selected as candidate schemes.

[0020] In some implementations, the key metrics cover constraint satisfaction rate, implementation cost, and project timeline impact.

[0021] The beneficial effects of this invention are as follows: By constructing a constraint library adapted to the target project, binding the structured parameters in the constraint library with the corresponding components in the BIM model, and then combining automatic detection of conflicting components, quantitative sorting of professional constraint weights, and optimization of secondary sorting rules, multiple candidate schemes are iteratively generated within the constraint threshold range based on priority. Coupled with the final scheme compliance verification and the update of the full-link responsibility traceability chain, this invention effectively solves the pain points of inconsistent constraint parameters, subjective priority judgment, lack of constraint guidance in scheme decision-making, and difficulty in defining decision-making responsibility in traditional BIM multi-professional collaboration conflicts. It significantly improves the scientificity and accuracy of conflict decision-making, ensuring that high-priority professional hard constraints are satisfied first and that the entire decision-making process is traceable and compliance is verifiable. Attached Figure Description

[0022] Figure 1 This is a flowchart illustrating the multi-disciplinary conflict decision optimization method based on a BIM model disclosed in an embodiment of the present invention. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of this invention clearer and more explicit, the content of this invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, only the parts relevant to this invention are shown in the accompanying drawings, not all of them.

[0024] This embodiment proposes a multi-disciplinary conflict decision optimization method based on a BIM model, such as... Figure 1 As shown, it includes the following steps:

[0025] Step 1: Construct a constraint library based on the target project. Configure hard constraint parameters for each discipline in the constraint library according to the applicable conditions of the target project. After standardizing the hard constraint parameters, generate structured parameters and bind the structured parameters to the corresponding components in the BIM model.

[0026] Step 1 aims to complete the digital construction of the constraint library in the project initiation phase, laying the foundation for subsequent automated conflict decision-making. Specifically, the constraint library in this solution includes an industry standard library, a project-customized library, and a professional custom library. The industry standard library provides general, mandatory basic constraints, representing the minimum compliance standard for all hard constraint parameters and ensuring the solution does not violate industry regulations and technical standards. The project-customized library provides constraints specific to the target project, covering contractual agreements, client requirements, and project-specific scenarios. These are project-level hard indicators beyond industry standards, primarily serving the client. The professional custom library provides core technical constraints within each discipline of the design unit, derived from the experience accumulated by professional designers. These indicators are not explicitly stated in industry standards or project contracts but directly affect the performance of the professional system, serving as a technical guarantee for ensuring the normal functioning of the professional system.

[0027] During the construction of the constraint library, hard constraint parameters need to be standardized to facilitate large-scale invocation in subsequent steps. In one example, the standardization of hard constraint parameters includes: using a natural language processing model to automatically parse the raw text data from industry standard libraries, project-customized libraries, and professional custom libraries; mapping the parsed semantic information to the corresponding structured parameters according to the dimensions of the structured parameters; and pushing parameters with a parsing confidence level lower than a preset threshold (e.g., 90%, which can be flexibly adjusted by project technicians according to the accuracy requirements of the target project) or mapping failures to manual review, correction, or completion.

[0028] In the aforementioned standardization process, Natural Language Processing (NLP) models were used as the primary tool for semantic recognition and feature extraction. Pre-trained models based on the Transformer architecture, such as BERT or RoBERTa, which possess excellent recognition performance, can be employed. However, before practical application of this solution, general-purpose pre-trained models require training on data from the architectural field. Specifically, a domain corpus is constructed from general architectural engineering texts, including national standards, industry specifications, standard atlases, and academic papers. Experts then annotate the hard constraint descriptions within this corpus, focusing on the dimensions of structured parameters. Specifically, the dimensions of these structured parameters include the relevant professional field, constraint library type, constraint type, indicator name, constraint threshold, weight factor, and verification rules. Constraint types include safety compliance, project requirements, cost control, and efficiency optimization.

[0029] During the structured parameter annotation process, weight factors are automatically generated based on the constraint type, with fixed dimension weights as follows: safety compliance (0.4) > project requirements (0.3) > cost control (0.2) > efficiency optimization (0.1). In one example, if an industry standard states that "the net height of a shop's interior should not be less than 2.8m," the annotation process would specify the following: profession: architecture; constraint library type: project-customized library; constraint type: project requirement; indicator name: net height of a shop's interior; constraint threshold: ≥2.8m; weight factor: 0.3; and verification rules: a red warning is triggered if the height is below 2.8m, and a yellow warning is triggered if the height is between 2.8m and 2.9m. Finally, the annotated dataset is subjected to supervised fine-tuning on a selected pre-trained model to obtain a natural language processing model specifically designed for parsing hard constraint parameters. Even better, the fine-tuned natural language processing model can be packaged into an independent service.

[0030] After the above annotation and training tasks, the natural language processing model possesses a relatively accurate ability to recognize semantics in the architectural field. Taking the standardization task of the industry standard library as an example, if the project contract appendix "Requirements for Mechanical and Electrical Systems of Commercial Complexes" is imported into the natural language processing model, the model will automatically parse it according to seven dimensions: professional field, constraint library type, constraint type, indicator name, constraint threshold, weight factor, and verification rules, and automatically identify the hard constraint parameters in the text. For example, the contract clause "Energy efficiency ratio (COP) of air conditioning duct system ≥ 4.2, duct air velocity ≤ 8m / s" will generate structured parameters after mapping, specifically including: professional field: mechanical and electrical, constraint library type: project customized library, constraint type: project requirement, indicator name: energy efficiency ratio (COP) of duct system, constraint threshold: ≥ 4.2, weight factor: 0.3, verification rule: if COP is lower than 4.2, the scheme will be automatically recalculated and added to the project customized library.

[0031] Optionally, after standardizing the hard constraint parameters, the process also includes: automatically comparing the structured parameters of each professional custom library with the industry standard library and project customization library in the constraint library for compliance. If parameter conflicts are found, the conflicting items are marked and pushed to the corresponding professional designer. After the parameters are corrected and confirmed, they are updated back into the professional custom library. This step aims to address the loopholes in the hard constraint parameters of each profession and to perform compliance verification on the parameters in the professional custom library.

[0032] Step 2: Automatically detect conflicting components in the BIM model, extract corresponding structured parameters from the constraint library, quantify and calculate the professional constraint weight scores for each conflicting component, and aggregate them to generate a priority ranking table based on professional constraints. If there are professional constraints with the same score in the priority ranking table, adjust the scores according to preset secondary ranking rules until the scores of each professional constraint are different. The secondary ranking rules are preset by the project manager based on the core requirements of the project. Optionally, the ranking priority is as follows: constraint type (safety and compliance > project requirements > cost control > efficiency optimization) > conflict level (level 1 conflict > level 2 conflict) > number of core constraints (more > fewer). The adjustment method is to add a correction value of 0.1 times the base score to the original score of the professional constraints with higher ranking.

[0033] Step 2 aims to associate the original conflict detection results of the BIM model with the structured parameters in the constraint library, and generate a corresponding priority ranking table by setting rules in advance, so as to replace manual negotiation and judgment of the hard constraint parameters that should be satisfied first in the current conflict of the target project.

[0034] Specifically, the automatic detection of conflicting components in the BIM model includes: using the OBB collision detection algorithm to identify overlapping conflicts in the BIM model space, and then using a rule reasoning engine based on structured parameters to predict potential conflicts that are not spatially overlapping but violate constraint thresholds. The rule base of the rule reasoning engine is built based on industry standards, project requirements and historical conflict data, and overlapping conflicts and potential conflicts are assigned different priorities and weight coefficients.

[0035] The following example illustrates the collision between a smoke exhaust duct, a secondary beam, and a suspended ceiling. Specifically, the OBB collision detection algorithm detects overlap between the duct, the suspended ceiling, and the secondary beam, generates a conflict event ID, and automatically extracts all structured parameters of the involved components, as shown in Table 1 below.

[0036] Table 1. Structural parameters of the components involved.

[0037] First, the constraint threshold of "safe duct spacing ≥ 0.15m" is retrieved and compared with the current actual value "distance between duct and secondary beam 0.08m". It is confirmed that the constraint threshold is not met. Considering the geometric overlap and parameter violations, it is officially judged as "Level 1 conflict (spatial overlap conflict)" and the priority calculation weight coefficient is set to 1.0.

[0038] Then, the rule base (built based on industry standards, project requirements, and historical conflict data, including multiple manually set inference rules) is invoked to retrieve the core electromechanical parameters from Table 1 (constraint threshold "duct velocity ≤ 10 m / s", current actual value "9.2 m / s", weight factor "0.1"). If the "duct shifting around the beam by 0.6 m" solution is adopted to resolve the spatial conflict, the duct length needs to be increased by 0.8 m, and a 90° bend needs to be added. Based on the fluid mechanics formula (resistance loss = λ × L / D × v² / 2g), the inference is... The original duct airflow was Q = 28000 m³ / h, the original duct diameter was DN1000, the frictional resistance coefficient was λ = 0.02, the original duct length was L0 = 12 m, the new duct length was ΔL = 0.8 m, the local resistance coefficient of the 90° elbow was ξ = 1.5, and the original air velocity was v0 = 9.2 m / s. The calculation process showed that the total resistance loss increased by 15%, and the flow velocity inside the duct was proportional to the square root of the resistance loss, so v1 = v0 × √(1.15) = 9.2 × 1.072 = 10.6 m / s, exceeding the constraint threshold of ≤10 m / s. Finally, based on the parameter threshold and the reasoning results, a "secondary conflict (potential constraint conflict)" was determined, and the priority calculation weight coefficient was set to 0.7.

[0039] In this scheme, a weighted summation algorithm is used to calculate the weight scores of each professional constraint. In one example, the total score of a professional constraint = Σ(single constraint weight factor × constraint satisfaction score) × the sum of conflict weight coefficients. The constraint satisfaction score is determined by the degree of matching between the actual value and the threshold (10 points for meeting, 8 points for being close to the threshold, 6 points for not meeting the threshold, and 0 points for exceeding the threshold). It should be noted that the above definition of "close to the threshold" is: the actual value is within ±5% of the constraint threshold and does not exceed the threshold; "not meeting the threshold" is defined as: the actual value exceeds ±5% but does not exceed the threshold; and "exceeding the threshold" is defined as: the actual value violates the core requirements of the constraint threshold. The above scoring criteria can be adjusted by the project manager according to the accuracy requirements of the target project. The weight factors within each professional area are used to distinguish the relative importance of different constraint types. They do not need to meet the normalization requirement of a sum of 1; they only need to reflect the priority differences of constraint types such as safety compliance and project requirements within the same professional area, ensuring that the score calculation aligns with the core needs of the professional area. Furthermore, if a profession has at least one core constraint involving the superposition of Level 1 and Level 2 conflicts in the current conflict event, the sum of the overall conflict weight coefficients for that profession is the sum of the weight coefficients calculated from the two priority levels. The conflict weight coefficient only supports the superposition of Level 1 conflict (1.0) + Level 2 conflict (0.7) across levels, and the coefficient after superposition is 1.7; in addition, conflicts of the same level are not superimposed (e.g., if there are multiple Level 1 or Level 2 conflicts, the coefficient still takes the single value of the corresponding level), and there is no upper limit to the superposition of conflict coefficients, calculated according to the actual number of cross-level conflicts. Specifically, the criteria for determining the above core constraints are: a weight factor ≥ 0.3 and a constraint type that is safety compliance or project requirement. Core constraints are the key constraints for conflict decision-making, and non-core constraints (weight factor < 0.3 and not safety compliance or project requirement) do not participate in the determination of the superposition of conflict weight coefficients. In addition, if only a single conflict level is involved (e.g., only Level 1 conflict), the coefficient is 1.0. Based on the above example, the calculation process of the professional constraint weight score is as follows:

[0040] (1) Mechanical and electrical engineering (component involved: air duct - DN1000-007)

[0041] Safety and compliance category (duct safety distance): weighting factor 0.4, actual value 0.08m < threshold 0.15m (not met), satisfaction score 6 points;

[0042] Cost control category (single component modification cost): weight factor 0.2, no modification (meets constraints), satisfaction score 10 points;

[0043] Efficiency optimization category (duct air velocity): weight factor 0.1, actual value 9.2m / s < threshold 10m / s (compliant), satisfaction score 10 points;

[0044] Single major basic score = (0.4×6) + (0.2×10) + (0.1×10) = 2.4 + 2 + 1 = 5.4 points;

[0045] The sum of conflict weight coefficients = Level 1 conflict 1.0 + Level 2 conflict 0.7 = 1.7;

[0046] The final score for the Mechanical and Electrical Engineering major is 5.4 × 1.7 = 9.18 points.

[0047] (2) Structural Engineering (Components involved: Secondary beam - L800×400-05)

[0048] Safety and compliance category (secondary beam deflection): weighting factor 0.4, actual value 22mm < threshold 24mm (compliant), satisfaction score 10 points;

[0049] Safety and compliance category (seismic resistance level): weight factor 0.4, after completion it is level III (compliant), satisfaction score 10 points;

[0050] Single major basic score = (0.4 × 10) + (0.4 × 10) = 4 + 4 = 8 points;

[0051] The sum of conflict weight coefficients = 1.0 for conflicts involving only Level 1;

[0052] The final score for structural engineering is 8 × 1.0 = 8.0 points.

[0053] (3) Architectural specialty (component involved: ceiling - area B-012)

[0054] Project Requirements (Ceiling Height): Weighting factor 0.3, actual value 2.9m = threshold 2.9m (compliant), satisfaction score 10 points;

[0055] Safety and compliance category (ceiling fire resistance rating): weighting factor 0.4, actual value 1.0h = threshold 1.0h (compliant), satisfaction score 10 points;

[0056] Single major basic score = (0.3 × 10) + (0.4 × 10) = 3 + 4 = 7 points;

[0057] The sum of conflict weight coefficients = 1.0 for conflicts involving only Level 1;

[0058] The final score for architecture is 7 × 1.0 = 7.0 points.

[0059] Even better, if the automatically detected conflicting component is not bound to the corresponding structured parameters, a completion reminder is automatically sent to the corresponding professional designer. If the structured parameters are not completed within a specified time (e.g., 24 hours), non-safety compliance constraints are processed according to the default lowest priority defined in the constraint library; safety compliance constraints immediately trigger a manual supervision process, and priority is determined after the parameters are completed. Before completion, the generation of solutions for the conflict is suspended. In one example, the completion reminder includes the conflict event ID, the ID of the involved component, the name of the missing parameter, and the constraint type.

[0060] Step 3: Based on the priority ranking table, at least two candidate schemes are generated iteratively within the threshold range of each professional constraint corresponding to the structured parameters. The key indicators of each candidate scheme are marked for manual decision-making. After the decision is made, the final scheme is generated. For professional constraints with higher scores in the priority ranking table, the corresponding hard constraint parameters are given priority when generating candidate schemes. The iteration process is set with a maximum number of iterations or a scheme convergence threshold as the termination condition.

[0061] The process of generating candidate solutions includes first selecting an iterative algorithm based on the conflict type of conflicting components to form a preliminary solution for resolving the conflict, and then optimizing the preliminary solution to generate candidate solutions.

[0062] In one example, the initial plan is as follows:

[0063] For the spatial overlap conflict between "duct-secondary beam / ceiling" (pipeline-related spatial conflict): the A* path search algorithm is used to optimize the duct space layout and generate "Solution 1: the duct is moved 0.6m to the left of the secondary beam".

[0064] To address the potential constraint conflict (parameter-based conflict) of "duct wind speed": a genetic algorithm is used to iteratively adjust the duct cross-section within the constraint threshold (wind speed ≤ 10 m / s) to generate "Solution 2: duct diameter expanded to DN1100".

[0065] Regarding the conflict between "ceiling and air duct" (a conflict among building components): a parametric adjustment algorithm is used to generate "Solution 3: Local reduction of ceiling height by 0.12m".

[0066] All the preliminary solutions described above satisfy the requirement of "core hard constraint satisfaction rate ≥ 95%", and proceed to the second stage of optimization. Specifically, based on the preliminary solutions, a multi-objective optimization function is constructed with the optimization objectives of maximizing constraint satisfaction rate, minimizing implementation cost, and minimizing project schedule impact. According to the basic algorithm used when generating the preliminary solutions, a genetic algorithm suitable for discrete-continuous mixed variable spaces is adapted and used for unified solution solving, iteratively searching the optimal solution space of the multi-objective optimization function. The genetic algorithm is compatible with the output results of basic algorithms such as A* path search and parameterized adjustment. Decision variables (discrete / continuous) of different conflict types are uniformly encoded and iterated. Each candidate solution corresponds to an optimized variant of the preliminary solution. During the iteration process, a crowding ranking strategy is used to retain non-dominated solutions, and finally, at least two optimal solutions are selected as candidate solutions.

[0067] In one example, a multi-objective optimization function is constructed: Objective 1 (maximizing constraint satisfaction rate), Objective 2 (minimizing implementation cost), and Objective 3 (minimizing project schedule impact). A genetic algorithm is used to solve the function. The layout path of the ductwork is treated as a discrete decision variable, while the duct cross-sectional dimensions and ceiling height reduction are treated as continuous variables and uniformly encoded. An iteration termination condition is set: maximum number of iterations = The algorithm was tested 50 times. The convergence threshold was set at ≥98% for the core hard constraint parameters (during the iterative optimization of candidate algorithms, when the number of hard constraint parameters satisfied by an algorithm reaches or exceeds 98% of the total number of core hard constraint parameters, the algorithm is considered to have converged) and the implementation cost fluctuation was ≤5%. The crossover probability was set to 0.8, the mutation probability to 0.05, and the population size to 40. The crowding ranking strategy was used to retain non-dominated solutions. Finally, two optimal candidate algorithms were selected, and key indicators were marked for manual decision-making. These key indicators covered the constraint satisfaction rate, implementation cost, and project schedule impact. The constraint satisfaction rate was the weighted score rate of the core constraints of each discipline. The constraint satisfaction rate = Σ(single constraint weight factor × satisfaction score) ÷ Σ(single constraint weight factor × full score of 10) × 100%, as shown in Table 2.

[0068] Table 2 Candidate Solutions

[0069] Note: The calculation process for the 95.9% weighted satisfaction rate in the architectural field is as follows: Σ(single constraint weight factor × satisfaction score) ÷ Σ(single constraint weight factor × 10 points) × 100% = (0.3 × 8 + 0.4 × 10) ÷ (0.3 × 10 + 0.4 × 10) × 100% = 6.4 ÷ 7 × 100% ≈ 95.9%. Among them, the ceiling height of 2.78m, after confirmation by the owner, is counted as 8 points according to the "approach threshold".

[0070] Step 4: Perform compliance verification on the final solution. After confirming that the unconstrained threshold has been exceeded, record the entire chain of information from conflict detection to decision implementation of the final solution, generate a responsibility traceability chain, and update the final solution and responsibility traceability chain to the BIM model.

[0071] In step 4, the compliance verification includes basic verification and related verification. The basic verification includes comparing the parameters of the final solution (e.g., solution 3) with the threshold values ​​of the structural parameters of the components in question (as shown in Table 1). The ceiling height is 2.78m (with written confirmation from the owner, meeting the project's customized library constraints), secondary beam deflection is 22mm (≤24mm), and duct velocity is 8.6m / s (≤10m / s), with no constraint threshold exceeding the detection limit. The related verification includes: after the final solution is updated to the BIM model, with the conflicting component as the center, all related components (including all components from architecture, structure, and MEP specialties, such as sprinkler heads, emergency lighting, evacuation signs, pipelines, supports, etc.) within the related verification range (e.g., 5m, which can be flexibly adjusted by project technicians according to component type) are checked for new conflicts using the same judgment criteria as in step 2. If no new conflicts are found, the related verification passes.

[0072] The above embodiments are merely illustrative of the technical concept and features of the present invention, and are intended to enable those skilled in the art to understand the content of the present invention and implement it accordingly. They should not be construed as limiting the scope of protection of the present invention. All equivalent changes or modifications made based on the essence of the content of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A multi-disciplinary conflict decision optimization method based on a BIM model, characterized in that, Includes the following steps: A constraint library based on the target project is constructed. Hard constraint parameters for each discipline are configured in the constraint library according to the applicable conditions of the target project. After the hard constraint parameters are standardized, structured parameters are generated, and the structured parameters are bound to the corresponding components in the BIM model. The system automatically detects conflicting components in the BIM model, extracts the corresponding structured parameters from the constraint library, quantifies and calculates the professional constraint weight scores corresponding to each conflicting component, and aggregates them to generate a priority ranking table based on professional constraints. If there are professional constraints with the same score in the priority ranking table, the scores are adjusted according to the preset secondary ranking rules until the scores of each professional constraint are different. Based on the priority ranking table, at least two candidate schemes are generated iteratively within the threshold range of each professional constraint corresponding to the structured parameters, and the key indicators of each candidate scheme are marked for manual decision-making. After the decision-making, the final scheme is generated. For professional constraints with higher scores in the priority ranking table, the corresponding hard constraint parameters of the professional are given priority when generating candidate schemes. The iterative process has a maximum number of iterations or a scheme convergence threshold as a termination condition. The final solution is subjected to compliance verification. After confirming that the unconstrained threshold has been exceeded, the entire chain information of the final solution from conflict detection to decision implementation is recorded, a responsibility traceability chain is generated, and the final solution and the responsibility traceability chain are updated to the BIM model.

2. The multi-disciplinary conflict decision optimization method based on BIM model as described in claim 1, characterized in that, The constraint library includes an industry standard library, a project customization library, and a professional custom library.

3. The multi-disciplinary conflict decision optimization method based on BIM model as described in claim 2, characterized in that, The standardization of the hard constraint parameters includes: using a natural language processing model to automatically parse the raw text data from the industry standard library, project customization library, and professional custom library; mapping the parsed semantic information to the corresponding structured parameters according to the dimensions of the structured parameters; and pushing parameters with a parsing confidence level lower than a preset threshold or mapping failure to manual review, correction, or completion to human review.

4. The multi-disciplinary conflict decision optimization method based on BIM model as described in claim 3, characterized in that, The dimensions of the structured parameters include the major, constraint library type, constraint type, indicator name, constraint threshold, weight factor, and verification rules. The constraint types include security compliance, project requirements, cost control, and efficiency optimization.

5. The multi-disciplinary conflict decision optimization method based on BIM model as described in claim 4, characterized in that, After the hard constraint parameters are standardized, the method also includes: automatically comparing the structured parameters of each professional custom library with the industry standard library and project customization library in the constraint library for compliance. If a parameter conflict is found, the conflict item is marked and pushed to the corresponding professional designer. After the parameter is corrected and confirmed, it is updated to the professional custom library.

6. The multi-disciplinary conflict decision optimization method based on BIM model as described in claim 1, characterized in that, The automatic detection of conflicting components in the BIM model includes: using an OBB collision detection algorithm to identify overlapping conflicts in the BIM model space, and then using a rule reasoning engine based on the structured parameters to predict potential conflicts that are not spatially overlapping but violate constraint thresholds. The rule base of the rule reasoning engine is built based on industry standards, project requirements and historical conflict data, and the overlapping conflicts and the potential conflicts have different priority calculation weight coefficients.

7. The multi-disciplinary conflict decision optimization method based on BIM model as described in claim 1, characterized in that, If the automatically detected conflicting component is not bound to the corresponding structured parameters, a completion reminder will be automatically sent to the corresponding professional designer. If the structured parameters are not completed within the specified time, non-safety and compliance constraints will be processed according to the default lowest priority defined in the constraint library. Safety and compliance constraints will immediately trigger a manual oversight process. Priority will be determined after all parameters are completed. The generation of solutions for the conflict will be suspended until all parameters are completed.

8. The multi-disciplinary conflict decision optimization method based on BIM model as described in claim 1, characterized in that, The process of generating the candidate solution includes first selecting an iterative algorithm based on the conflict type of the conflicting component to form a preliminary solution for resolving the conflict, and then optimizing the preliminary solution to generate the candidate solution.

9. The multi-disciplinary conflict decision optimization method based on BIM model as described in claim 8, characterized in that, Based on the preliminary scheme, a multi-objective optimization function is constructed with the objectives of maximizing constraint satisfaction rate, minimizing implementation cost, and minimizing project schedule impact. Then, an optimization algorithm suitable for discrete-continuous mixed variable space is used to iteratively search the optimal solution space of the multi-objective optimization function. Each candidate solution corresponds to an optimization variant of the preliminary scheme. During the iteration process, a crowding ranking strategy is used to retain non-dominated solutions. Finally, at least two optimal schemes are selected as candidate schemes.

10. The multi-disciplinary conflict decision optimization method based on a BIM model as described in claim 1, characterized in that, The key indicators cover constraint satisfaction rate, implementation cost, and project timeline impact.

Images