Intelligent layout optimization method and system for high-rise building construction operation platform

Abstract

The invention belongs to the field of aerial building machinery, and discloses an intelligent layout optimization method for a high-rise building construction operation platform, which comprises the following steps of: establishing a data set responded by key components of the high-rise building construction operation platform through a multi-software joint simulation environment; a DSR algorithm is developed by embedding physical priori knowledge and dimensional analysis, and modeling of a high-rise building construction operation platform is simplified; and obtaining an optimal design scheme of the high-rise building construction operation platform layout by using an intelligent optimization method. The method has been verified in a practical project applying a high-rise building construction operation platform in China. The invention shows the potential of man-machine cooperation in the process of improving the design efficiency, optimizing the spatial layout and promoting the design of the high-rise building construction operation platform by the intelligent design technology.

Description

Technical Field

[0001] This invention belongs to, but is not limited to, the field of high-rise building construction technology, and particularly relates to a method and system for intelligent layout optimization of high-rise building construction operation platforms. Background Technology

[0002] High-rise buildings offer a practical solution for optimizing land resources and maximizing development potential within limited space, playing a vital role in urban development and improving residents' living standards. However, high-rise buildings are characterized by difficult construction organization, high construction risks, and significant environmental impacts. A survey on scaffolding accidents at construction sites revealed that scaffolding-related accidents account for 70% of all deaths and injuries on construction sites. This reflects that traditional integral lifting scaffolding and hydraulic climbing formwork commonly used in construction are no longer sufficient to meet the construction needs of high-rise buildings. Therefore, there is an urgent need for safe and efficient construction equipment to meet the rapid development of high-rise buildings worldwide.

[0003] Against this backdrop, aerial high-rise building construction operation platforms (hereinafter referred to as "high-rise building construction operation platforms") have emerged to achieve safe and efficient construction of high-rise buildings. This large-scale construction equipment functions like an integrated aerial construction factory, providing a safe and reliable platform for construction workers even in strong winds and heavy rain. The structure of the high-rise building construction operation platform forms a vertical construction assembly line: workers complete rebar tying, formwork installation, and concrete pouring inside, greatly improving construction efficiency while reducing labor costs. Compared with traditional climbing formwork, this new equipment offers advantages including strong load-bearing capacity, high construction efficiency, strong adaptability, multi-equipment integration, and intelligent operation.

[0004] Due to the long-term nature of high-altitude operations, the structural safety of high-rise building construction work platforms must be considered during the design process. The layout design of high-rise building construction work platforms, especially the top steel platform, is influenced by a series of factors. First, structural safety must be a primary consideration, requiring the selection of a reasonable main truss and lateral support layout scheme. Second, the support columns must be positioned according to the planar structure of the core tube. Third, the steel platform truss layout must adapt to the construction requirements of the high-rise building construction work platform to prevent conflicts between various work processes and equipment. Furthermore, the design methodology for high-rise building construction work platforms should follow the modular principle, which helps simplify system modules, improve structural flexibility, minimize material usage, and support sustainable development. However, existing design methods for high-rise building construction work platforms rely on expert knowledge and experience, and are characterized by being time-consuming, inefficient, and conservative. Therefore, optimizing the layout design process of high-rise building construction work platform steel platforms through the application of advanced intelligent design algorithms has become an urgent need.

[0005] Recently, an increasing number of studies have explored the application of intelligent design methods in the design process of high-rise building construction work platforms. These advanced methods utilize artificial intelligence (AI) to improve the design efficiency, structural safety, and economic cost of high-rise building construction work platforms. However, current research mainly focuses on the predefined layout of steel platforms for optimizing the design of high-rise building construction work platforms. This method lacks the necessary flexibility, making it difficult to adapt to different project requirements and conditions, thus limiting its applicability. To address the challenges of modeling difficulties and low efficiency in scheme generation and optimization caused by structural complexity during the design process, this invention proposes a layout design method based on intelligent optimization algorithms.

[0006] Based on the above analysis, the urgent technical problems that need to be solved in the existing technology are:

[0007] (1) How to effectively improve the modeling efficiency of complex structures of high-rise building construction operation platforms;

[0008] (2) How to transform the design problem of high-rise building construction operation platform under various working conditions into an optimization problem;

[0009] (3) How to propose an automated and intelligent design framework for the layout of construction operation platforms for high-rise buildings. Summary of the Invention

[0010] To address the problems existing in the prior art, this invention provides a method and system for intelligent layout optimization of high-rise building construction operation platforms.

[0011] This invention is implemented as follows: a method and system for intelligent layout optimization of high-rise building construction operation platforms, characterized by an intelligent optimization design technology and system for the layout of high-rise building construction operation platforms considering physical constraints. The method specifically includes:

[0012] S1: Establish a dataset of responses of key components of a high-rise building construction operation platform through a multi-software co-simulation environment;

[0013] S2: A deep symbolic regression algorithm is developed by embedding prior physical knowledge and dimensional analysis to simplify the modeling of high-rise building construction operation platforms;

[0014] S3: Use intelligent optimization methods to obtain the optimal design scheme for the layout of construction operation platforms in high-rise buildings.

[0015] Furthermore, S1 includes:

[0016] (1) Parametric modeling of key components of high-rise building construction operation platform mainly involves modeling the structural components of the high-rise building construction operation platform, namely the steel platform and support columns. In order to simplify the modeling process, the scaffolding system, formwork system and ancillary facilities are considered as loads applied to the main structure. The parametric modeling process is carried out using Grasshopper.

[0017] (2) Collect structural component response data from finite element calculations, convert the parametric model into a physical model, and perform batch calculations.

[0018] Furthermore, S2 considers the standardization method of dimensional analysis, including:

[0019] (1) Based on the characteristics of the experimental data, construct a dimension matrix and arrange the three indices in the dimensional expression of each feature into a column. Then, use the Toeplitz matrix to calculate the difference in dimension indices between adjacent columns and quantify the dimensional differences between different features.

[0020] (2) Select a scaling vector M, which represents the median or mean of the data for each feature, take the logarithm of each element in the scaling vector, and then use the Toeplitz matrix to quantify the scaling differences between the feature data.

[0021] (3) Determine the scaling factor x = [x, y, z] T These correspond to the powers of the basic units of length, time, and mass, respectively;

[0022] (4) The scaling factor x is used to adjust the unit of each feature. The dimensionality index of each feature is multiplied by the corresponding scaling factor to obtain the normalization coefficient. The normalization coefficient X' is then used to adjust the unit of each feature. i It is applied to the original data to obtain normalized features under the new unit system.

[0023] Further, step S2, which uses physically constrained deep symbol generation via a recurrent neural network (RNN), includes the following steps:

[0024] (1) Use a recurrent neural network to output probability vectors in an autoregressive manner;

[0025] (2) Use prior constraints to reduce the search space;

[0026] (3) Use reinforcement learning (RL) algorithms to train RNNs to generate expressions that more closely meet the expected criteria. Once a preorder traversal is sampled, the corresponding symbolic expression is instantiated and evaluated using a reward function.

[0027] Furthermore, S3, which involves intelligent optimization of layout design through a simplified model, includes the following steps:

[0028] (1) Structural layout design analysis of high-rise building construction operation platform: Based on the planar layout of the core tube, a parametric model of the high-rise building construction operation platform is established. The design variables include the variable positions of the main truss, secondary truss, transverse bracing and support columns. Then, Swallow is used to configure material properties, section specification and external load. Subsequently, Tunny, an optimization component using Optuna in Grasshopper, is used to define the optimization problem and drive the intelligent design process. Finally, users can analyze the importance of parameters during the optimization process and evaluate the structural response of the optimal solution by examining the optimization results in detail.

[0029] (2) Layout optimization design based on NGSA-III: NSGA-III is selected for optimization tasks.

[0030] The following provides a set of four system claims, in particular claim 1 is the independent claim, and the remaining claims are additional technical features subordinate to claim 1. Each claim is limited to the description of technical features and does not involve technical effects or functional descriptions, and the scope of protection of the independent claim is maximized.

[0031] This invention also provides a method and system for intelligent layout optimization of high-rise building construction operation platforms, the system comprising:

[0032] (1) Data acquisition module, used to build response datasets of key components of high-rise building construction operation platform in a multi-software joint simulation environment;

[0033] (2) Deep symbolic regression module, which embeds prior physical knowledge and dimensional analysis to generate normalized symbolic expressions;

[0034] (3) Intelligent optimization module, used to drive the layout design process of high-rise building construction operation platform according to the normalized symbolic expression.

[0035] Furthermore, the data acquisition module includes:

[0036] (1) Parametric modeling unit, which parametrically models the steel platform and support columns in the key components of the high-rise building construction operation platform. The parametric modeling uses Grasshopper tool, in which the scaffolding system, formwork system and auxiliary facilities act as loads on the main structure.

[0037] (2) Physical model conversion unit, which converts the parameterized model into a physical model;

[0038] (3) Finite element analysis unit performs batch calculations on the physical model to collect structural component response data.

[0039] Furthermore, the deep symbolic regression module includes a dimensional analysis submodule, characterized in that the dimensional analysis submodule includes:

[0040] (1) Construct a dimension matrix based on the characteristics of the experimental data, arrange the three exponents in the dimensional expression of each feature into a column, and use the Toeplitz matrix to calculate the exponent difference between adjacent columns;

[0041] (2) Select a scale vector representing the median or average of each feature data, take the logarithm of each element, and use the Toeplitz matrix to quantify the scale difference between each feature data;

[0042] (3) Determine the scaling factor vector x = [x, y, z]^T used to adjust each feature unit, which corresponds to the basic units of length, time, and mass;

[0043] (4) Adjust the dimension index of each feature using the scaling factor to obtain the normalization coefficient, and normalize the original data accordingly.

[0044] Furthermore, the intelligent optimization module includes:

[0045] (1) High-rise building construction operation platform parameter model construction unit, based on the core tube size, establishes a parameter model including the layout of main truss, secondary truss, transverse support and support column;

[0046] (2) Material and load configuration unit, using Swallow to configure material properties, cross-sectional parameters and external loads;

[0047] (3) A layout design unit based on the Optuna optimization component in Grasshopper, which uses Tunny to define optimization problems and drive the intelligent optimization process;

[0048] (4) The optimization task was carried out using NSGA-III.

[0049] Based on the above technical solutions and the technical problems solved, the advantages and positive effects of the technical solution to be protected by this invention are as follows:

[0050] First, this invention proposes an automatic optimization method for the layout design of a high-rise building construction operation platform based on deep symbolic regression (DSR), parametric modeling, and multi-objective optimization (MOO). Specifically, firstly, a structural response dataset is established for the key components of the high-rise building construction operation platform using a multi-software co-simulation environment. Subsequently, a physically constrained deep symbolic regression (PCDSR) algorithm is proposed to explore the structural mechanical properties and establish a simplified model of the high-rise building construction operation platform. Finally, an automated optimization framework integrating parametric modeling and MOO is constructed to achieve intelligent and efficient layout design. This method has been verified in a real-world project in China that utilizes a high-rise building construction operation platform. The results show that: (1) the PCDSR formula can accurately predict and explain the mechanical properties of Bailey trusses, and the R of the formula is... 2 Above 0.97. (2) The optimized scheme saves 20.38% of steel compared to the existing scheme, and the average structural deformation is reduced by 49.76%. (3) Due to the combination of physical constraints, PCDSR exhibits exceptional robustness in the field of fitting multi-level noise formulas. The scientific contribution lies in integrating physical prior knowledge and dimensional analysis into DSR, thereby providing an intelligent design framework for the effective layout design of high-rise building construction operation platforms. This study demonstrates the potential of intelligent design technology in improving design efficiency, optimizing spatial layout, and promoting human-machine collaboration in the design process of high-rise building construction operation platforms.

[0051] Secondly, the current layout design of high-rise building construction work platforms typically relies on traditional manual design, which is time-consuming, labor-intensive, and makes it difficult to find the optimal solution. Furthermore, the structural layout design of high-rise building construction work platforms mainly depends on the core tube planar form of the high-rise building, offering significant flexibility. Therefore, it is necessary to integrate intelligent design methods and concepts into the design phase of high-rise building construction work platforms. Currently, there is limited research and technology related to intelligent design of high-rise building construction work platforms. Published literature mainly focuses on predefined structural layouts, which lack adaptability, leaving a gap in related research. To fill this technological gap in the field of intelligent design for high-rise building construction work platforms, this invention proposes an intelligent optimization design technology for the layout of high-rise building construction work platforms that considers physical constraints. This technology combines parametric modeling and intelligent optimization algorithms, automatically generating layout schemes for high-rise building construction work platforms that simultaneously meet structural safety and economic requirements based on actual project needs. This effectively improves the work efficiency during the design phase, liberates the productivity of designers, significantly enhances the intelligence level of human-computer interaction in high-rise building construction work platforms, and elevates my country's high-rise building design and construction technology to a higher level of development.

[0052] Traditional high-rise building construction platform design suffers from three major problems: First, the complex structure leads to extensive and time-consuming iterative modeling, resulting in insufficient flexibility. Second, the high degree of personalization in the design process means that the structural layout of high-rise building construction platforms primarily relies on the core tube planar form, requiring different layout schemes for different projects. Third, the reliance on manual operation means that current structural layouts depend heavily on expert experience, resulting in limited design data, a tendency to get stuck in local optima, and difficulty in obtaining the optimal solution. To address these three problems, this invention proposes an intelligent optimization design technology for high-rise building construction platform layouts that considers physical constraints. First, the PCDSR method is used to simplify the complex structure of the high-rise building construction platform, obtaining a simplified model and improving modeling efficiency. Second, parametric modeling technology enables rapid batch generation of design schemes. Finally, based on intelligent optimization algorithms, considering actual constraints, the system automatically finds the optimal structural layout scheme that balances safety and economy, achieving fully automated intelligent generation of schemes throughout the high-rise building construction platform design process. Attached Figure Description

[0053] Figure 1 This is a framework diagram of the automatic optimization layout design method for high-rise building construction operation platform based on deep symbolic regression provided in the embodiments of the present invention;

[0054] Figure 2 This is the main system of the high-rise building construction operation platform provided in the embodiments of the present invention;

[0055] Figure 3 This is the internal structure of the steel platform at the construction site and the high-rise building construction operation platform provided in the embodiments of the present invention;

[0056] Figure 4 The key components of the high-rise building construction operation platform provided in this embodiment of the invention are: (a) main truss, (b) secondary truss, and (c) support column;

[0057] Figure 5 This is the parametric modeling of the truss structure provided in the embodiments of the present invention;

[0058] Figure 6 This is an automated calculation workflow in the integrated software environment provided in the embodiments of the present invention;

[0059] Figure 7 The following are the calculation schemes for the structural response dataset provided in the embodiments of the present invention: (a) Scheme 1, (b) Scheme 2, (c) Scheme 3, and (d) Scheme 4;

[0060] Figure 8 This is the expression generation workflow provided in the embodiments of the present invention;

[0061] Figure 9 The following are the parametric models of the simplified model of the high-rise building construction operation platform provided in the embodiments of the present invention: (a) secondary truss, (b) main truss, (c) column, and (d) lateral support;

[0062] Figure 10 This invention provides an automated optimization framework for the layout design of high-rise building construction operation platforms.

[0063] Figure 11 These are on-site construction photos of the high-rise building construction operation platform provided in this embodiment of the invention: (a) exterior, (b) platform, (c) internal structure;

[0064] Figure 12 These are the dimensions and cross-section of the Bailey truss provided in the embodiments of the present invention;

[0065] Figure 13 The training datasets from the FEM model provided in this embodiment of the invention are: (a) Scenario 1, (b) Scenario 2, (c) Scenario 3, and (d) Scenario 4;

[0066] Figure 14 The reward allocation during the training process provided in the embodiments of the present invention are: (a) Scenario 1, (b) Scenario 2, (c) Scenario 3, (d) Scenario 4;

[0067] Figure 15 The Pareto fronts generated by the formulas provided in the embodiments of the present invention are: (a) Scenario 1, (b) Scenario 2, (c) Scenario 3, and (d) Scenario 4;

[0068] Figure 16 The training process of the model provided in the embodiments of the present invention is as follows: (a) a main truss with vertical load, (b) a main truss with horizontal load, (c) a secondary truss with vertical load, and (d) a secondary truss with horizontal load.

[0069] Figure 17 The structural variations of Combo-1 provided in this embodiment of the invention are: (a) the original model and (b) the simplified model;

[0070] Figure 18 The structural variations of Combo-2 provided in this embodiment of the invention are: (a) the original model and (b) the simplified model;

[0071] Figure 19 The structural variations of Combo-3 provided in this embodiment of the invention are: (a) the original model and (b) the simplified model;

[0072] Figure 20The structural variations of Combo-4 provided in this embodiment of the invention are: (a) the original model and (b) the simplified model. Detailed Implementation

[0073] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0074] like Figure 1 As shown in the figure, this invention provides a method and system for intelligent layout optimization of high-rise building construction operation platforms. The method specifically includes:

[0075] S1: Establish a dataset of responses of key components of a high-rise building construction operation platform through a multi-software co-simulation environment;

[0076] S2: The DSR algorithm is developed by embedding physical prior knowledge and dimensional analysis to simplify the modeling of high-rise building construction operation platforms;

[0077] S3: Use intelligent optimization methods to obtain the optimal design scheme for the layout of construction operation platforms in high-rise buildings.

[0078] The main systems of high-rise building construction operation platforms are Figure 2 High-rise building construction platform systems mainly include steel platform systems, scaffolding systems, support systems, and power systems. The steel platform serves as a crucial workspace for various construction activities and provides storage areas for equipment and materials. Workers utilize the scaffolding system to perform tasks such as rebar tying, formwork installation, and concrete pouring. Construction loads on the platform are transferred to the support system, which is the primary component for transmitting vertical forces. The power system, integrated within the support columns, uses hydraulic cylinders to lift the platform to new heights.

[0079] In existing technologies, the layout of high-rise building construction platforms largely relies on empirical formulas and static design parameters. This results in the stress analysis of key platform components often being performed using decentralized finite element models, making global collaborative optimization impossible in a multi-software environment. During actual construction, the coupling effect between the steel platform and the scaffolding system can lead to stress concentration and deformation accumulation. Traditional methods, when considering the coupling of construction loads and time-varying working conditions, often suffer from increased safety margin assessment errors due to data gaps or insufficient modeling accuracy. Furthermore, the lack of standardized data interface formats between various professional software programs leads to a disconnect in data acquisition regarding platform stress and dynamic system response, resulting in extended design cycles and difficulty in iterative updates to adapt to the ever-changing on-site working environment.

[0080] From a physical modeling perspective, existing technologies for constructing mechanical models of high-rise building construction platforms often employ progressively refined finite element meshes. As the platform height and number of components increase, the model's degrees of freedom expand rapidly, leading to a surge in computational resource consumption. Particularly in the coupled analysis of the support and power systems, it is necessary to simultaneously consider the nonlinear stiffness characteristics of the hydraulic cylinders and the natural vibration frequencies of the scaffolding system. This interdisciplinary modeling process is both time-consuming and prone to introducing numerical instability, creating uncertainty for subsequent optimization. Furthermore, dimensional and parameter redundancy issues make it difficult to effectively characterize the response function in the initial stages, hindering engineers from directly grasping the overall force distribution from a macroscopic perspective.

[0081] In existing integrated designs of scaffolding and steel platforms, the longitudinal load transfer of supporting columns and the horizontal wind load are often simplified as static concentrated loads, failing to fully utilize modern sensing technology to monitor platform vibration and tilt changes in real time, thus limiting the implementation of dynamic adjustment strategies. Simultaneously, the matching relationship between the driving power and platform stiffness of the hydraulic jacking system at different construction heights has not been systematically analyzed, leading to jacking delays or advances at some construction nodes, making it difficult to ensure platform level accuracy and construction continuity. Because traditional layout optimization focuses only on a single constraint (such as span or height) and lacks a comprehensive evaluation of multiple objectives (safety, economy, and construction efficiency), the optimization results suffer from insufficient adaptability in field applications.

[0082] To address the aforementioned technical bottlenecks, the proposed method first establishes a dataset of key component responses covering different construction heights, load conditions, and support schemes by coupling finite element analysis tools with a dynamic simulation platform within a multi-software co-simulation environment. During this process, high-frequency sampling combined with time-domain integration is used to obtain the steel platform stiffness matrix, scaffolding system mode shape parameters, and hydraulic cylinder response curves. This ensures the dataset closely approximates real-world conditions and provides high-quality samples for black-box model training.

[0083] Building upon this foundation, this invention introduces the Deep Symbolic Regression (DSR) algorithm, embedding prior physical knowledge and dimensional analysis into the model construction process. By normalizing the geometric dimensions, material elastic modulus, and hydrodynamic parameters of each platform component and mapping them to a symbolic expression space, the DSR algorithm can automatically discover the simplest mathematical expression to describe the response behavior of key components under different load conditions. This approach not only effectively compresses the model's degrees of freedom but also preserves its interpretability, facilitating engineers' understanding of the platform's stress mechanisms from a physical perspective and significantly reducing the computational resources required for subsequent simulation verification.

[0084] During the optimization phase, a multi-objective evolutionary algorithm (such as Non-Dominated Sorting Genetic Algorithm II, NSGA-II) is used to perform a global search for the platform layout. Optimization variables include the steel platform span, scaffold node density, support column arrangement, and the installation location of the hydraulic jacking system. The objective function comprehensively considers indicators such as maximum deflection limits, platform vibration control, construction costs, and real-time jacking efficiency. By decoupling the DSR model from the dataset and embedding it into the optimization loop, the algorithm only needs to perform rapid calculations on the symbolic expression in each iteration, without repeatedly calling high-precision finite element simulations, thus significantly improving the convergence speed. The final Pareto front set provides engineers with a series of balanced solutions between safety and economy, facilitating customized selection based on on-site construction conditions.

[0085] This intelligent layout optimization method acquires high-precision response data from multiple software environments, simplifies the model using physical priors and dimensional analysis, and achieves global optimization by combining a multi-objective evolutionary algorithm. This not only overcomes the drawbacks of traditional modeling, such as massive computational demands and difficulties in iterative updates, but also provides an interpretable, scalable, and easily applicable optimization process for the layout of high-rise building construction platforms. This method can effectively improve construction efficiency and economic benefits while ensuring the overall structural safety of the platform, providing a new technical path for the design and construction management of high-rise building platforms in the industry.

[0086] Figure 3 This paper describes the internal structure of a high-rise building construction platform. The steel platform is assembled using Bailey trusses, with the main truss consisting of four Bailey trusses arranged side-by-side and reinforced with stiffening members. The secondary and outer trusses are made of individual Bailey trusses. The formwork and scaffolding systems are suspended from the secondary trusses, while the outer trusses are used to improve the overall rigidity of the structure and ensure its stability. The steel platform is fixed to the core tube by support columns. The integrated hydraulic cylinder system within these columns consists of multiple cylinders forming the power mechanism, capable of lifting the platform to new heights. The steel platform primarily bears building loads, scaffolding loads, formwork loads, and wind loads.

[0087] S1 includes:

[0088] (1) Parametric modeling of key components of high-rise building construction operation platform, mainly focusing on the main components of high-rise building construction operation platform, namely steel platform and support columns, as shown in the figure. Figure 4The scaffolding system, formwork system, and ancillary facilities are considered equivalent to loads applied to the main structure. To achieve rapid modeling, the parametric modeling process in this invention is performed using Grasshopper. Grasshopper's intuitive parametric modeling interface provides designers with a platform to interact with the algorithm, facilitating the immediate visualization of the algorithm-generated design solutions. Assuming the supporting columns are standardized components, the model can be used directly without further parameterization. Therefore, this paper mainly studies the parametric modeling of main beams and secondary beams. The only geometric variable is the longitudinal length of the Bailey truss. By parameterizing this variable, Bailey trusses of different lengths can be quickly generated. It is worth noting that, as... Figure 5 The length of the Bailey truss varies discretely, with the variation step being the length of a single Bailey segment.

[0089] (2) Collect structural component responses from finite element calculations, convert the parametric model into a physical model, and perform batch calculations. To begin this process, various parameters must be configured, including material properties, section assignments, external loads, and boundary constraints. Swallow supports data transfer from Grasshopper to SAP2000. Specifically, it can be used to define structural properties such as section properties, loads, and node constraints, bind these properties to the geometric model in Grasshopper, assemble them into a structural analysis model, and then import the model into SAP2000 for calculation using the SAP2000 API interface.

[0090] With the GH_CPython plugin, users can develop custom components using Python code in Grasshopper, effectively establishing connections between Bailey trusses and automatically saving configuration data. Furthermore, the powerful Design of Experiments (DOE) software JMP is used to generate samples. Specifically, cluster-based optimal space filling (OSF) technology has proven to be an effective sample generation method. The input samples generated by DOE are then processed by a finite element method (FEM) simulation system to calculate the desired output. By integrating the above methods, it is possible to automate the processes of model building, data transmission, analysis, and data acquisition. In summary, the automated calculation workflow for the layout design of high-rise building construction operation platforms is shown below. Figure 6 .

[0091] To comprehensively evaluate the mechanical response of the main truss and secondary trusses, this invention designs four different loading conditions. These conditions include: (1) a vertical load applied to the secondary truss; (2) a lateral load applied to the secondary truss; (3) a vertical load applied to the main truss; and (4) a lateral load applied to the main truss. The direction and area of ​​uniformly distributed load application are shown in the figure. Figure 7The blue area represents the region bearing the load. The boundary conditions of the truss are set as simply supported constraints.

[0092] S2, considering the standardization method of dimensional analysis, includes:

[0093] (1) Construct a dimension matrix based on the characteristics of the experimental data. The dimension of any feature can be expressed as a product of powers of the dimensions of the basic physical quantities. In civil engineering, commonly used basic physical quantities include length (L), time (T), and mass (M). Therefore, dim x i Represented as x i The characteristic dimensions can be expressed as the product of powers of L, T, and M, as shown below.

[0094]

[0095] Arrange the three exponents in the dimensional expression of each feature into a column. By doing this for all features, the dimension matrix of all parameters of the problem can be obtained as follows.

[0096] Subsequently, the Toeplitz matrix is ​​used to calculate the difference in dimensionality indices between adjacent columns, thereby quantifying the dimensional differences between different features.

[0097] (2) Select a scaling vector M, which represents the median or mean of the data for each feature. Take the logarithm of each element in the scaling vector, and then use the Toeplitz matrix to quantify the scaling differences between the feature data.

[0098] (3) Determine the scaling factor x = [x, y, z] T These correspond to the powers of the basic units of length, time, and mass, respectively. To calculate these scaling factors, a system of linear equations (7) needs to be solved. The generalized inverse matrix A' is calculated by... * The unique minimum norm solution of x can be obtained, as shown in equation (8).

[0099] A′·x=M′ (7)

[0100] x = [x, y, z] T =A * ·M' (8)

[0101] (4) The scaling factor x is used to adjust the units of each feature. Specifically, the dimensionality index of each feature is multiplied by the corresponding scaling factor to obtain the normalization coefficient, as shown below:

[0102] L i =x·α 1i ,T i =y·α 2i M i =z·α3i (9)

[0103] The normalization coefficient X' i Applied to the original data to obtain normalized features under the new unit system, as shown below:

[0104]

[0105] Through the steps described above, the size differences between adjacent features are quantified, and a suitable unit scaling factor is found by solving a system of linear equations. These scaling factors are then applied to adjust the units of each feature, ensuring that the feature data values ​​in the new unit system fall within a reasonable range. Therefore, this process enhances the stability and accuracy of model training.

[0106] S2, which uses physically constrained deep symbol generation of a recurrent neural network (RNN), includes the following steps:

[0107] (1) A recurrent neural network (RNN) is used to output a probability vector in an autoregressive manner. The RNN output is used to generate a vector through a softmax layer. This vector defines the probability distribution used to select a symbol, conditioned on previously selected symbols. The softmax layer transforms the RNN output into a probability distribution for selecting the next symbol, which depends on all previously generated symbols. When sampling symbols, the RNN input is the representation of the previously sampled symbols. The expression generation workflow is as follows: Figure 8 As shown.

[0108] (2) Use prior constraints to reduce the search space. The specific constraints are as follows: minimum and maximum length of the expression; operands of operators cannot all be constants; operands of unary operators cannot be inverse operations; descendants of triangular operators cannot be triangular operators.

[0109] (3) Train the RNN using a reinforcement learning (RL) algorithm to generate expressions that more closely meet the expected criteria. Once a preorder traversal is sampled, the corresponding symbolic expression is instantiated and evaluated using a reward function.

[0110] For specific mechanical and physical problems, prior knowledge can be incorporated into the expression generation process. Specifically, based on prior mechanical and physical knowledge, the generated expression must include certain physical variables. When a generated expression lacks important physical variables, it can be considered that the expression does not conform to reality. By assigning a reward of 0, expressions that do not meet the requirements can be automatically filtered out. Therefore, the method based on physical constraints improves the fitting accuracy and physical meaning of deep symbolic regression.

[0111] The pseudocode for PCDSR is shown in Algorithm 1.

[0112]

[0113] S3, which involves intelligent optimization of layout design through a simplified model, includes the following steps:

[0114] (1) Analysis of the integrated structural layout design of construction operation platform for high-rise buildings

[0115] (1.1) Based on the plan layout of the core tube's outer wall, the location for the Bailey truss can be determined, as shown in the figure. Figure 9 (a).

[0116] (1.2) The four main trusses are arranged in mutually perpendicular directions. The movement steps of these main trusses follow modularity rules to ensure that their movement range does not exceed that of the outermost Bailey truss. The decision of whether to include the corresponding main truss and its position is considered a design variable, as shown. Figure 9 (b)

[0117] (1.3) Since the support columns are typically placed beneath the main trusses, two main trusses are selected from two vertical directions. Under each truss, there are four support columns. The decision regarding the inclusion of these support columns and their locations is considered a design variable, as shown. Figure 9 (c)

[0118] (1.4) Lateral bracing is used to distribute the lateral load borne by the outer truss. Therefore, it is necessary to install lateral bracing along the exterior of the steel platform. The decision on whether to include such lateral bracing and its location is considered a design variable, as shown. Figure 9 (d)

[0119] To achieve automated and intelligent design of the layout of construction operation platforms for high-rise buildings, this invention proposes an optimization framework, such as... Figure 10 The automated optimization process comprises six stages: visual modeling of a simplified finite element model for high-rise building construction operation platforms; definition of external loads; establishment of an automated calculation framework; finite element simulation and analysis under load, multi-objective optimization considering safety and economy; and post-processing by the designer.

[0120] In summary, a parametric model of the high-rise building construction platform is established based on the core tube dimensions. Design variables include the layout of the main trusses, secondary trusses, lateral braces, and support columns. Then, Swallow is used to configure material properties, section specifications, and external loads. The integration of GH_CPython facilitates the definition of constraints and elastic supports, enabling seamless data transfer between the parametric model and SAP2000. Subsequently, the Tunny plugin is used to define the optimization problem and drive the intelligent design process. Finally, users can analyze the importance of parameters during optimization and evaluate the structural response of the optimal solution by thoroughly examining the optimization results.

[0121] (2) Layout design based on NGSA-III with intelligent optimization

[0122] Tunny supports a variety of optimization algorithms. In this invention, NSGA-III is chosen for the optimization task because this algorithm enhances the ability to efficiently handle multi-objective problems with a large number of objectives, provides a well-distributed Pareto front, and ensures robust convergence to the optimal solution. NSGA-III is a multi-objective optimization algorithm based on the reference point method, emphasizing that population members should be non-dominated and close to a specified reference point. Compared to the diversity preservation mechanism of NSGA-II, the most important modification of NSGA-III is that it uses uniformly distributed reference points to maintain population diversity.

[0123] Specifically, the core of NSGA-III lies in its reference-point-based non-dominated ranking mechanism. The computational process includes hyperplane creation, reference point establishment and standardization, and selection operations. First, all solutions are evaluated to determine their objective values, and an ideal point is determined as the minimum of all objectives for a given population. Then, the objective value of each solution is adjusted relative to this ideal point. Using the ASF function, extreme points are identified to define the hyperplane. Reference points are uniformly distributed on the unit simplex. In the standardized objective space, each individual is associated with the nearest reference line, determined by its proximity to these reference points. Selection is performed by ranking individuals based on their distance from their associated reference lines, prioritizing the closest individuals. This simplified approach ensures efficient and effective multi-objective optimization.

[0124] The design objectives of high-rise building construction work platforms include various indicators, such as structural deformation, structural stress, and steel consumption. In this invention, the objective is to minimize structural deformation and steel consumption while ensuring structural safety through optimized structural layout. Therefore, the optimization objective is to minimize structural deformation and steel consumption. To ensure structural safety, constraints are imposed, ensuring that the maximum deformation of the design scheme must be below a safety threshold, the maximum internal force must be within allowable limits, and the maximum stress must not exceed the safety threshold. Based on this principle, the high-rise building construction work platform layout optimization problem can be formulated as follows.

[0125] Minimize F(x) = [δ(x), m(x)]

[0126] stg1(x) = max(δ(x)) - δ safety ≤0

[0127] g2(x) = max(Q(x)) - Q safety ≤0

[0128] g2(x) = max(M(x)) - Msafety ≤0

[0129] g3(x) = max(σ(x)) - σ safety ≤0 (16)

[0130] Here, x represents the design scheme, δ(x) and m(x) represent the structural deformation and steel consumption, respectively. Q(x), M(x), and σ(x) are the structural shear force, bending moment, and stress. δ safety Q safety M safety and σ safety These refer to the safety thresholds for structural deformation, shear force, moment, and stress, respectively.

[0131] To verify the effectiveness of the proposed design framework, data from a real high-rise building construction platform project in China was used for demonstration. This high-rise building project is located in Wuhan, China. The building has 80 floors above ground and a height of 380 meters.

[0132] (I) Case Background

[0133] The Yangtze Center, located in Wuhan, China, consists of an 80-story tower and four basement levels. The tower's concrete structure reaches a height of 380 meters, with standard floor heights ranging from 4.4 to 4.5 meters. Notably, the building utilizes a fourth-generation high-rise construction platform, achieving rapid construction through standardized design, an industrialized production environment, integrated assembly construction technology, and intelligent equipment integration. Compared to climbing formwork systems, this reduces the number of construction workers by 50% and increases work efficiency by 30%. (See diagram) Figure 11 The high-rise building construction platform for this project has a plan dimension of approximately 35 meters and 30 meters, and a facade height of 23 meters. It comprises three structural layers: a rebar tying layer, a concrete pouring layer, and a concrete curing layer. Weighing 800 tons, the platform is supported by 12 lightweight anchor points and can withstand hurricanes up to Category 14. The steel platform system is constructed from standardized components, particularly utilizing widely available Type 200 standard Bailey trusses. Figure 12 The dimensions and cross-sections of various Bailey elements are described.

[0134] This invention employs two objectives to evaluate the structural safety and cost-effectiveness of high-rise building construction platforms: maximum structural deformation and steel consumption. To further ensure structural safety, constraints require that structural deformation, internal forces, and stresses must not exceed their respective safety thresholds.

[0135] Table 1 Design parameters in the optimization framework

[0136]

[0137]

[0138] Based on the interaction and boundary conditions of the finite element model, the contact relationship of the Bailey bridge panels in the steel platform is set as hinged, and the constraint points on the support columns are designated as elastic constraints, constraining displacements in the X, Y, and Z directions. The stiffness of the constraints is determined according to the actual situation. High-rise building construction platforms mainly bear stacking loads, scaffolding loads, formwork loads, wind loads, etc. Detailed load values ​​and application locations are listed in Table 2. According to the specifications, the load effect combination selected in this invention is shown in Table 3.

[0139] Table 2 Loads on construction operation platforms for high-rise buildings.

[0140]

[0141] Table 3 Load combinations for the finite element model of the construction platform for high-rise buildings.

[0142]

[0143] (II) Model Establishment

[0144] Geometric models of primary and secondary trusses with spans ranging from 6m to 32m were established using parametric techniques. Table 4 lists four calculation schemes for key components of the high-rise building construction platform. Specifically, according to the design load specifications for high-rise building construction platforms, the vertical load range of the primary truss was set to [0, 30] kN / m, and the horizontal load range was set to [0, 5] kN / m. For the secondary trusses, the vertical load range was set to [0, 10] kN / m, and the horizontal load range was set to [0, 2] kN / m. Four different datasets were generated using orthogonal experimental design, each containing 500 samples. The transformation from the geometric model to the physical model was achieved through a combined simulation using Grasshopper and SAP2000. The connections between Bailey beams were modeled as pin joints. Each dataset includes independent variables such as span and load, with the dependent variable being the maximum deformation of the truss. The training dataset was used to train the PCDSR algorithm, enabling accurate prediction of structural behavior under various load conditions. The training dataset is shown in [the table]. Figure 13 middle.

[0145] Table 4. Scenario settings for finite element calculation.

[0146]

[0147] The obtained dataset was input into PCDSR, and the hyperparameters of the algorithm were set as shown in Table 5. The reward distribution for the four scenarios during training is shown in the figure. Figure 14The density represents the Gaussian kernel density of the reward for the generated expression. As the number of training iterations increases, the reward of the expression generated by PCDSR gradually increases, eventually converging to 1, indicating that the algorithm has reached convergence and the generated expression can accurately reflect the physical characteristics of the truss structure. Figure 15 This represents the Pareto front of the expression, where the horizontal axis represents the complexity of the expression, and the vertical axis represents the fit of the expression to the dataset. Surprisingly, the beam deformation ω=cqL 4 The expression / EI appears at the Pareto front in four different scenarios, with a fitting accuracy R0. 2 The value consistently exceeds 0.97. This high accuracy indicates that the deformation of a Bailey truss under uniform load can be equivalently simulated as the deformation of a simply supported beam. Therefore, the equivalent bending stiffness of the Bailey structure can be obtained through fitting. This step helps to equate the complex structure of a high-rise building construction platform to a simplified beam element structure, thereby simplifying the structural modeling process.

[0148] Table 5. Hyperparameters of deep symbolic regression

[0149]

[0150] In this invention, the prediction accuracy of PCDSR was evaluated to highlight the superiority of the algorithm, and the results are shown below. Figure 16 Together with Table 6. To verify the effectiveness of the simplified model of the high-rise building construction platform, a comparative experiment was conducted between the original model and the simplified model. The contour plot of the finite element model is shown in the figure. Figures 17 to 20 The errors are summarized in Table 7. Subsequently, an automated intelligent design framework was applied under these conditions. The degree of improvement compared to existing solutions is listed in Table 8. Finally, the roles of physical prior knowledge and dimensional analysis constraints in PCDSR were investigated. The main findings are summarized below.

[0151] (1) PCDSR can effectively generate deformation formulas for truss structures and derive equivalent mechanical properties from these formulas. The Pareto front formula for truss deformation can be obtained through PCDSR, and the optimal formula can be determined based on structural mechanics. Based on this, the equivalent moment of inertia of the cross section under the four scenarios is 184200.988 cm. 4 213.736cm 4 824465.212cm 4 and 27883.929cm 4 It can be observed that the main truss has significantly stronger bending resistance than the secondary truss. To simplify the complexity, the truss cross-section is equivalent to a rectangular cross-section with dimensions of 30.57 mm high and 897.52 mm wide, and 157.50 mm high and 856.43 mm wide. Figure 16As shown, deformation formulas for different scenarios are used to calculate truss deformations in the training data. The results show that the formula obtained from deep symbolic regression has excellent performance in predicting various deformations of Bailey trusses, with R0... 2 The values ​​are 0.989, 0.977, 0.991, and 0.999. In summary, the deformation formula can simplify complex truss structures into rectangular cross-section beams, laying the foundation for establishing a simplified model of high-rise building construction operation platforms.

[0152] Table 6 Equivalent section moment of inertia and equivalent beam dimensions

[0153]

[0154] (2) The deformation of the simplified model under various load combinations is consistent with that of the original model, indicating that the simplified model can serve as a simplified replacement for more complex models. Specifically, to verify the calculation accuracy of the simplified model of the high-rise building construction platform, a finite element model was established in SAP2000 based on the width b and height h of the equivalent beam, as shown in Table 6. The structural deformation of the simplified model under different loads was compared with that of the original model. Figures 17 to 20 The structural deformations of the simplified and original models are presented. It can be observed that the deformation of the simplified model is consistent with that of the original model. The maximum deformations of each model are 85.36 mm and 83.69 mm, respectively, meeting safety requirements. Under other load combinations, the errors between the original and simplified models are 5.57%, -0.264%, and -3.07%, respectively, all less than 6%. The results show that the simplified model proposed in this study can be used to replace more complex models to improve computational efficiency while maintaining sufficient accuracy. This method is particularly useful in situations where computational resources are limited or rapid response is required. It is worth emphasizing that traditional modeling methods require at least 6 hours to build the finite element model for inverse design. In contrast, the simplified model proposed here only requires 30 minutes for parametric modeling. Furthermore, by adjusting design parameters, a large number of design schemes can be flexibly generated through parametric techniques. Therefore, in the initial design stage, compared with traditional methods, this method reduces modeling time by more than 90%, significantly improving the efficiency of high-rise building construction platform layout design and paving the way for rapid intelligent optimization.

[0155] Table 7 Comparison between the original model and the simplified model

[0156]

[0157] In the design of construction operation platforms for high-rise buildings, a multi-objective optimization framework based on NSGA-III was introduced, which significantly improved structural performance and material utilization. By simultaneously minimizing the maximum structural deformation and steel consumption, the optimized scheme saved 20.38% of steel compared to the traditional design (from 273.02t to 217.39t), and reduced the maximum structural deformation by an average of 49.76%, fully demonstrating the algorithm's advantages in optimization efficiency and result quality.

[0158] The Pareto front obtained through multi-objective optimization provides designers with a set of non-dominant optimal solutions, which can be flexibly selected according to site conditions and construction requirements. Among all candidate schemes, a structural deformation threshold of 96 mm and a reference steel consumption of 273.08 t are set as benchmarks to quickly locate the target range. If the construction focuses on weight reduction, the low-steel-consumption scheme can be prioritized; if higher displacement control is required, further compression deformation can be performed within the allowable weight increase range.

[0159] Taking one of the most representative solutions as an example, the platform controlled the maximum deformation to 49.20 mm using only 217.89 tons of steel, and the stress levels of each key component were within the safe threshold. This is thanks to the strict setting of stress constraints during the optimization process, ensuring lightweight design without sacrificing structural safety.

[0160] A detailed comparison of four working conditions (Combo-1 to Combo-4) further confirmed the algorithm's effectiveness: the maximum deformation of the original scheme was 85.36mm, 81.20mm, 75.68mm, and 78.91mm, respectively, which was reduced to 42.39mm, 37.96mm, 42.59mm, and 38.09mm after optimization, corresponding to optimization rates of 50.34%, 53.25%, 43.72%, and 51.73%. This result demonstrates the robustness of the optimization framework under multiple scenarios and load combinations.

[0161] Sensitivity analysis revealed significant differences in the impact of different components on the objective function: column arrangement was the most critical factor for deformation control, while the main truss design was the primary determinant of steel consumption. This finding provides targeted improvement directions for subsequent design phases, helping engineers concentrate limited optimization resources on the most influential aspects.

[0162] In constructing structural response prediction models, deep symbolic regression (PC-DSR) incorporating prior mechanical knowledge significantly improves training efficiency. Compared to models without prior knowledge, it achieves the same prediction accuracy (R²). 2The required training cycles (≈0.94) are reduced from 243 to 79, improving efficiency by approximately 70%. Simultaneously, dimensional constraints make the generation formula simpler and the physical meaning clearer, further enhancing the model's interpretability and providing reliable theoretical support for intelligent design processes.

[0163] It should be noted that embodiments of the present invention can be implemented in hardware, software, or a combination of both. The hardware portion can be implemented using dedicated logic; the software portion can be stored in memory and executed by a suitable instruction execution system, such as a microprocessor or dedicated-design hardware. Those skilled in the art will understand that the above-described devices and methods can be implemented using computer-executable instructions and / or included in processor control code, for example, such code provided on a carrier medium such as a disk, CD, or DVD-ROM, a programmable memory such as read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The devices and modules of the present invention can be implemented by hardware circuitry such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, or programmable hardware devices such as field-programmable gate arrays, programmable logic devices, etc., or by software executed by various types of processors, or by a combination of the above-described hardware circuitry and software, such as firmware.

[0164] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications, equivalent substitutions, and improvements made by those skilled in the art within the scope of the technology disclosed in the present invention, and within the spirit and principles of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for intelligent layout optimization of construction operation platforms for high-rise buildings, characterized in that, Includes the following steps: S1: Constructing a response dataset of key components for a high-rise building construction operation platform in a multi-software co-simulation environment; S2: Embedding prior physical knowledge and dimensional analysis to develop a deep symbolic regression algorithm to simplify the construction of high-rise building construction operation platform models; S3: Use intelligent optimization methods to determine the layout design scheme of the construction operation platform for high-rise buildings.

2. The method according to claim 1, characterized in that, Step S1 includes: (1) Parametric modeling of key components of high-rise building construction operation platform. The key components mainly include steel platform and support column. In the file, scaffolding system, formwork system and auxiliary facilities are applied as loads to the main structure. The parametric modeling uses Grasshopper tool. (2) The parametric model is converted into a physical model, and the response data of the structural components are collected in batches using finite element calculation.

3. The method according to claim 1, characterized in that, Step S2 employs a dimensional analysis standardization method, including: (1) Construct a dimension matrix based on the characteristics of the experimental data, in which the three exponents in the dimensional expression of each feature are arranged into columns, and the exponent difference between adjacent columns is calculated using the Toeplitz matrix; (2) Select a scale vector M representing the median or average of each feature data, take the logarithm of each element of the scale vector, and then use the Toeplitz matrix to quantify the scale difference between the feature data. (3) Determine the scaling factor x = [x, y, z]^T, which corresponds to the basic unit indices of length, time and mass, respectively; (4) Adjust the units of each feature using the scaling factor x, that is, multiply the dimension index of each feature by the corresponding scaling factor to obtain the normalization coefficient, and apply the normalization coefficient to the original data to form normalized features under the new unit system.

4. The method according to claim 1, characterized in that, Step S2 further includes physical constraint deep symbol generation based on a recurrent neural network, the step comprising: (1) Use a recurrent neural network to output a probability vector in an autoregressive manner; (2) Apply prior constraints to reduce the expression search space; (3) A recurrent neural network is trained using a reinforcement learning algorithm. Symbolic expressions are generated during the sampling preorder traversal, and the instantiated symbolic expressions are evaluated using a reward function.

5. The method according to claim 1, characterized in that, Step S3 involves intelligent optimization of the layout design through a simplified model, including: (1) Construct a parametric model of a high-rise building construction operation platform based on the core tube dimensions, wherein the design variables include the layout of the main truss, secondary truss, transverse support and support columns; (2) Use Swallow to configure material properties, section parameters and external loads, and use the Tunny optimization component based on Optuna in Grasshopper to define the optimization problem to drive the design process; (3) Select NGSA-III to perform intelligent optimization tasks.

6. The method according to claim 1, characterized in that, The response dataset of key components of the high-rise building construction operation platform constructed in step S1 includes structural response data collected during batch finite element calculations after the parametric model is converted into a physical model. The following is a set of four system claims, in which claim 1 is the independent claim, and the remaining claims are additional technical features subordinate to claim 1. Each claim is limited to the description of technical features and does not involve technical effects or functional descriptions, and the scope of protection of the independent claim is maximized.

7. A system for automatically optimizing the layout design of high-rise building construction operation platforms based on deep symbolic regression, characterized in that, The system includes: (1) Data acquisition module, used to build response datasets of key components of high-rise building construction operation platform in a multi-software joint simulation environment; (2) Deep symbolic regression module, which embeds prior physical knowledge and dimensional analysis to generate normalized symbolic expressions; (3) Intelligent optimization module, used to drive the layout design process of high-rise building construction operation platform according to the normalized symbolic expression.

8. The system according to claim 7, characterized in that, The data acquisition module includes: (1) Parametric modeling unit, which parametrically models the steel platform and support columns in the key components of the high-rise building construction operation platform. The parametric modeling uses Grasshopper tool, in which the scaffolding system, formwork system and auxiliary facilities act as loads on the main structure. (2) Physical model conversion unit, which converts the parameterized model into a physical model; (3) Finite element analysis unit performs batch calculations on the physical model to collect structural component response data.

9. The system according to claim 7, characterized in that, The deep symbolic regression module includes a dimensional analysis submodule, characterized in that the dimensional analysis submodule includes: (1) Construct a dimension matrix based on the characteristics of the experimental data, arrange the three exponents in the dimensional expression of each feature into a column, and use the Toeplitz matrix to calculate the exponent difference between adjacent columns; (2) Select a scale vector representing the median or average of each feature data, take the logarithm of each element, and use the Toeplitz matrix to quantify the scale difference between each feature data; (3) Determine the scaling factor vector x = [x, y, z]^T used to adjust each feature unit, which corresponds to the basic units of length, time, and mass; (4) Adjust the dimension index of each feature using the scaling factor to obtain the normalization coefficient, and normalize the original data accordingly.

10. The system according to claim 7, characterized in that, The intelligent optimization module includes: (1) High-rise building construction operation platform parameter model construction unit, based on the core tube size, establishes a parameter model including the layout of main truss, secondary truss, transverse support and support column; (2) Material and load configuration unit, using Swallow to configure material properties, cross-sectional parameters and external loads; (3) A layout design unit based on the Optuna optimization component in Grasshopper, which uses Tunny to define optimization problems and drive the intelligent optimization process; (4) The optimization task was carried out using NSGA-III.

Images