An exhibition hall space planning generation method based on artificial intelligence

Abstract

The application discloses a kind of exhibition hall space planning generation methods based on artificial intelligence, comprising the following steps: obtaining exhibition hall space basic data and exhibition hall function demand data and pre-processing;Initial text description of exhibition hall layout is generated by autoregressive generation mode;Exhibition hall layout graph structure is obtained based on initial text description;Exhibition hall layout state vector is generated according to exhibition hall layout graph structure, and initial state sequence is obtained;Initial state sequence is iteratively updated, and the optimized exhibition hall layout state sequence is obtained;Exhibition hall layout result is generated according to the optimized exhibition hall layout state sequence and three-dimensional scene generation is carried out.The application can automatically generate two-dimensional plane and three-dimensional layout scheme under the exhibition hall planning scene of diversified exhibition area demand, complex audience flow line and multiple security constraints coexisting by introducing large language model driven exhibition hall layout generation, which significantly improves the intelligent level and design efficiency of exhibition hall space planning.

Description

Technical Field

[0001] This invention relates to the field of exhibition hall space planning technology, and in particular to an artificial intelligence-based method for generating exhibition hall space plans. Background Technology

[0002] Current exhibition space planning mainly relies on manual experience and two-dimensional design software. Designers manually divide functional areas, arrange exhibits, and design passageways based on architectural floor plans using tools such as CAD. This method is highly dependent on the designer's personal experience, and the process of refining and modifying the plan is time-consuming and labor-intensive. It is difficult to generate multiple alternative layouts for comparison in a timely manner. At the same time, traditional methods often start only from planar geometric relationships and lack systematic modeling of the actual visitor's path, behavior, and safety evacuation constraints. As a result, the layout often relies on subjective judgment in terms of visitor flow continuity, exhibit visibility, and crowd capacity, and lacks quantifiable optimization criteria.

[0003] With the development of artificial intelligence and building information technology, existing technologies have developed exhibition hall layout assistance tools based on rule engines, heuristic algorithms, or simple machine learning models. However, these tools typically only make local adjustments to a given template and fail to model the entire exhibition hall as a sequence of states that evolves with the visitor process. They also fail to fully utilize natural language descriptions to quickly obtain complex functional requirements and constraints. Existing technologies have a low degree of integration of multi-source information such as exhibition area functions, area requirements, number of exhibits, visitor flow, and safety evacuation. They are unable to automatically generate structured layout representations and lack the ability to iteratively optimize layout states based on selective state space models and integrate them with the three-dimensional layout results. This limits the level of intelligence, automation, and refinement in exhibition hall space planning. Summary of the Invention

[0004] To address the technical problems existing in the background art, this invention proposes an artificial intelligence-based method for generating exhibition hall space planning.

[0005] This invention proposes an artificial intelligence-based method for generating exhibition hall space planning, comprising the following steps: S1. Obtain basic data of the exhibition hall space and functional requirements data of the exhibition hall and preprocess them to obtain the layout generation input feature vector; S2. Input the layout generation input feature vector into a pre-trained large-scale language model to generate an initial text description of the exhibition hall layout through an autoregressive generation method. S3. Based on the initial text description, perform semantic parsing to generate a set of functional areas, a set of edges, a set of area attributes, and a set of edge attributes to obtain the exhibition hall layout diagram structure; S4. Generate the exhibition hall layout state vector based on the exhibition hall layout diagram structure, and arrange the exhibition hall layout state vectors of all visited locations in the order of visit to form an initial state sequence; S5. The initial state sequence is iteratively updated based on the selective state space model to obtain the optimized exhibition hall layout state sequence. S6. Generate the exhibition hall floor plan layout result based on the optimized exhibition hall layout state sequence, and generate a 3D scene based on the exhibition hall floor plan layout result.

[0006] Preferably, the basic data of the exhibition hall space and the functional requirements data of the exhibition hall are acquired and preprocessed to obtain the layout generation input feature vector, as follows: Acquire basic spatial data for the exhibition hall, including the architectural floor plan, floor height information, and permitted exhibition area information. Establish a Cartesian coordinate system based on the architectural floor plan, and discretize the planar boundaries of the exhibition hall to obtain a set of boundary points. ,in, and Representing boundary points respectively In the Cartesian coordinate system, the set B is used to represent the planar boundary of the exhibition hall; The floor height information of the exhibition hall is quantified to obtain a set of floor height parameters. Based on the permitted exhibition area information, the permitted exhibition areas are labeled in a Cartesian coordinate system to obtain the set of permitted exhibition areas. ,in, This represents the floor height parameter of the r-th region. This represents the subset of regions where the s-th exhibition area is allowed to lie in the plane; Obtain and categorize the functional requirements data of the exhibition hall to obtain a set of exhibition area types. Exhibition area requirements Collection of exhibits Audience flow constraint set and the set of safety evacuation constraints ,in, This indicates the type identifier for the q-th type of exhibition area. Indicates the type of exhibition area The corresponding target area parameters, Indicates the type of exhibition area The corresponding number of exhibits This represents the constraint parameter for the r-th audience flow line. This represents the safety evacuation constraint parameter corresponding to the r-th audience flow line constraint; The basic data of the exhibition hall space and the functional requirements data of the exhibition hall are processed into feature vectors to construct spatial feature vectors and requirement feature vectors. The spatial feature vectors and requirement feature vectors are then concatenated in a predetermined order to form the layout generation input feature vector. .

[0007] Preferably, the layout generation input feature vector is input into a pre-trained large-scale language model to generate an initial text description of the exhibition hall layout through an autoregressive generation method, as follows: Generate input feature vectors for the layout. A linear transformation is performed to obtain a conditional vector, which is then input into a pre-trained large-scale language model. The pre-trained large-scale language model uses GPT-5, and a hidden state sequence is obtained through multi-layer transformation. , ; Based on hidden state sequences The initial layout description word sequence is generated using an autoregressive generation method. , ,in, This represents the u-th word in the output sequence, and the autoregressive generation satisfies the conditional probability decomposition: ; in, Indicates that given existing output tokens and layout to generate input feature vectors Generate the u-th output word under the condition of The conditional probability; The word sequence describing the initial layout Decoding is performed to obtain the initial text description of the exhibition hall layout. .

[0008] Preferably, the initial text description Includes collections of exhibition area types Each exhibition area type Functional area names and exhibition area requirements Target area parameters Corresponding target area description and exhibit quantity set Parameters of the number of exhibits Corresponding descriptions of the number of exhibits and the set of constraints on visitor flow. Audience Flow Constraint Parameters The corresponding audience flow sequence description and the set of safety evacuation constraints Medium safety evacuation constraint parameters The corresponding safety evacuation requirements are described.

[0009] Preferably, based on the initial text description, semantic parsing is performed to generate a set of functional regions, a set of edges, a set of region attributes, and a set of edge attributes to obtain the exhibition hall layout structure, as follows: Initial text description Sentence segmentation is performed to obtain a sentence sequence. , ,in, This represents the i-th sentence; For sentence sequence Each sentence in Semantic segmentation is performed to obtain a set of semantic segments. ; Based on the set of exhibition area types Exhibition area requirements Collection of exhibits Audience flow constraint set and the set of safety evacuation constraints For semantic fragment sets Semantic parsing is performed on each semantic fragment to obtain a set of region-related information triples. : ; in, This indicates the functional area identifier referred to by the semantic fragment. Indicates the functional area type. Indicates the target area parameter. This parameter represents the number of exhibits corresponding to that functional area. Construct a set of functional regions R based on the set of triples C containing region-related information. For each functional region in the set R... Establish regional attribute vectors ,in, , Representation and Functional Areas Corresponding exhibition area types, Representation and Functional Areas The corresponding target area parameters, Representation and Functional Areas The corresponding exhibit quantity parameter will be used to generate all area attribute vectors. Combined to form a set of regional attributes ; From the initial text description The audience flow sequence is described in the parsing, and the functional area access sequence is constructed. Access sequence based on functional area Construct an adjacency matrix for adjacent access relationships Among them, when the functional areas in the audience flow With functional areas When functional areas are adjacent and have a direct connection in the access sequence, let Otherwise And based on the adjacency matrix Obtain the edge set , ; Based on audience flow constraint set and safe evacuation constraint set set of opposite edges Each edge in Constructing edge attribute vectors Obtain the set of edge attributes , ,in, At least include parameters related to audience flow constraints. The corresponding streamline direction parameters and safety evacuation constraint parameters Corresponding evacuation requirement parameters; Set the functional region set R and the edge set Regional attribute set and edge attribute set Combine to generate the exhibition hall layout structure. , .

[0010] Preferably, an exhibition hall layout state vector is generated based on the exhibition hall layout diagram, and the exhibition hall layout state vectors of all visited locations are arranged in the order of visit to form an initial state sequence, as follows: Based on the exhibition hall layout structure Set of functional areas Perform spatial parameter initialization for each functional area. Determine the coordinates of the reference point To obtain the set of reference points , ; For each functional area Determine the set of boundary contour points , ,in, Indicates functional area The u-th boundary point, and The coordinate components of the boundary points are represented based on the boundary point set. Calculation functional area Geometric area parameters and with the region attribute vector Target area parameters By combining the features, we obtain the area feature vector. , ; For each functional area Based on the set of boundary contour points Calculate the width parameter of the minimum bounding rectangle and height parameters and obtain shape parameters , ,in, , , Indicates functional area The aspect ratio shape parameters will As a functional area Shape characteristics; Based on edge set and edge attribute set For each edge Get the channel length parameter of the corresponding channel and channel width parameters Construct channel feature vectors ,in, Indicates functional area With functional areas The center distance between them Indicates the connection function area With functional areas The channel width parameter; Based on the set of permitted exhibition areas Set of floor height parameters Audience flow constraint set and the set of safety evacuation constraints For each functional area Determine the corresponding floor height parameters and allowed exhibition area index parameters , ,in, Used to indicate functional areas The permitted exhibition area And based on preset population density parameters Calculation functional area Crowd capacity parameters The population density parameter Based on expert experience, : ; Each functional area The functional areas are formed by combining the reference point coordinates, width parameters, height parameters, shape parameters, geometric area parameters, target area parameters, exhibit quantity parameters, floor height parameters, and crowd capacity parameters. state vector : ; Functional area access sequence and edge set For each access location v, determine the corresponding functional area and the set of edges directly connected to the functional area. For each edge Obtain the corresponding channel feature vector Then perform aggregation operations to obtain the edge aggregation vector. ; Functional area access sequence Construct a showroom layout state vector for each access location v. The exhibition hall layout state vector is composed of the corresponding state vector. and edge aggregation vector composition, Arrange the exhibition hall layout state vectors of all visited locations in the order of visit to form an initial state sequence. , .

[0011] Preferably, the initial state sequence is iteratively updated based on a selective state-space model to obtain an optimized exhibition hall layout state sequence, as follows: Based on the exhibition hall layout structure and the initial state sequence Construct the optimization objective function ; For each iteration step and access location Initialize the hidden state vector sequence and define the selection vector sequence ,in, Used to indicate the layout state vector Are the dimensions in the iteration step? The middle part has been updated; In each iteration step Inside, for access location The current layout state vector Hidden state vector and selection vector Input a selective state-space model, which is based on the Mamba architecture, to obtain the updated hidden state vector. and candidate layout state vector ; Based on selection vector Update the layout state vector to obtain the iteration steps. Layout state vector The update rules are as follows: ; in, This indicates element-wise multiplication. This represents a vector where each component is 1. After updating all access locations, the latest state sequence is obtained. Calculate the corresponding objective function value. If satisfied and ,in, The convergence threshold, To minimize the number of iterations, the latest state sequence is... As an optimized exhibition hall layout state sequence .

[0012] Preferably, the optimization objective function consists of objective terms related to audience flow continuity constraints, regional area constraints, and passage width constraints, and is expressed as: ; in, , , These are the weighting coefficients. This indicates the target terms related to the continuity constraint of the audience flow. This represents the objective terms related to the area constraint of the region. This indicates the target items related to the passage width constraint; Audience flow continuity constraint related objective items Functional area access sequence and channel length parameters Perform the calculation: ; in, Indicates the state sequence The corresponding functional area is accessed at position v in the layout. Access Location Functional areas The center distance parameter between them This represents the preset target streamline step size parameter; Passage width constraint related target items Based on the edge set and channel feature vectors Perform the calculation: ; in, Indicates the state sequence Connecting functional areas under the corresponding layout With functional areas The channel width parameter, The minimum passage width threshold is set based on expert experience; Area-related target items The calculation method is as follows: ; in, Indicates the state sequence Functional areas under corresponding layout The geometric area parameters, Representation and Functional Areas The corresponding target area parameters.

[0013] Preferably, the exhibition hall floor plan is generated based on the optimized exhibition hall layout state sequence, and a 3D scene is generated based on the exhibition hall floor plan, as follows: Functional area access sequence and functional area set For each functional area From the optimized exhibition hall layout state sequence Extract the corresponding region state vector The region state vector Represented as: ; in, and Indicates functional area Optimization of reference point coordinate components in a Cartesian coordinate system and Indicates functional area Optimize the width and height parameters. Indicates functional area Optimized shape parameters, Indicates functional area Optimize geometric area parameters, Indicates functional area Target area parameters, Indicates functional area The number of exhibits Indicates functional area The corresponding floor height parameters, Indicates functional area Optimize the crowd capacity parameters; Based on the region state vector Includes optimized reference point coordinates Width parameters and height parameters In a Cartesian coordinate system, each functional area is represented by Construct a set of axis-aligned rectangle boundary points : ; in, , , , ; For all functional areas The set of rectangular boundary points is summarized to form the set of area boundaries for the exhibition hall's floor plan. , Each element is formed by connecting four boundary points in sequence to create a closed polygon, which is used to represent the functional areas in the exhibition hall's floor plan. The area occupied by the plane; Based on edge set And optimize reference point coordinates , For each edge Construct a set of channel polyline segments: ; in, , , This indicates the connection between functional areas in the exhibition hall's floor plan. With functional areas The channel segment; For all edges The corresponding set of broken line segments of the passageway are summarized to form the passageway set of the exhibition hall floor plan. : ; in, Each element Used to indicate connecting passages between functional areas in the floor plan of an exhibition hall; Based on the set of permitted exhibition areas and channel set Construct a connectivity matrix for the exhibition hall's floor plan. ,in: ; in, Indicates functional area With functional areas Effective connectivity signage in the exhibition hall layout; Set of region boundaries Channel set and connectivity matrix Combine to generate the exhibition hall floor plan. , And generate a 3D scene based on the exhibition hall floor plan; A three-dimensional rectangular coordinate system is established based on the exhibition hall floor plan. The exhibition hall floor plan is used as the horizontal reference plane, and a height direction is set. According to the rectangular boundary vertices and planar positions of each functional area and the corresponding floor height, each functional area is stretched along the height direction in the three-dimensional rectangular coordinate system to generate a three-dimensional polyhedron composed of a bottom surface, a top surface, and side surfaces, to represent the three-dimensional spatial occupancy range of each functional area. According to the connectivity between functional areas and the start and end positions of the passage centerline and the planar width parameters, a passage block with a predetermined height and width is generated in the three-dimensional rectangular coordinate system and is connected to or intersects with the boundary of the corresponding functional area, to represent the three-dimensional passage space between functional areas. Within the three-dimensional space of each functional area, several exhibition equipment arrangement anchor points are determined according to the center position, planar boundary, floor height, and number of exhibits of the functional area. The exhibition equipment arrangement anchor points include planar position and height position, which are used to indicate the arrangement position of the exhibition equipment in three-dimensional space. For each exhibit placement anchor point, the exhibit type and geometric dimension parameters are assigned based on the exhibition area type and exhibit quantity of the functional area to which the anchor point belongs. An exhibit object associated with the exhibit placement anchor point is generated, forming an exhibit object set in each functional area. The three-dimensional polyhedra of all functional areas, the passage blocks between all functional areas, and the entire exhibit object set are summarized to form a three-dimensional layout object set for the exhibition hall. The geometric information of each three-dimensional object, as well as attribute information such as exhibition area type, target area, exhibit quantity, floor height, and number of visitors, are organized according to a preset data structure to construct a three-dimensional layout model data for the exhibition hall space planning. This data is used to output the three-dimensional layout model of the exhibition hall space planning and its corresponding digital planning data.

[0014] The exhibition hall space planning and generation method proposed in this invention has the following beneficial technical effects: 1. This application deeply integrates a large language model with a selective state space model, unifying the exhibition space planning process into a closed-loop chain of "natural language requirements → structured layout diagram → layout state sequence → optimization results". This achieves a direct mapping from the curator's natural language intent to a quantifiable spatial structure. The large language model is not only used to generate the initial layout text, but also works in conjunction with feature vectors composed of exhibition area type, area requirements, number of exhibits, visitor flow lines, and safety evacuation constraints to automatically construct the graph structure and layout state representation of functional areas and their connectivity. This breaks through the limitations of existing technologies that rely on fixed templates or single rule engines, enabling complex requirements to enter the subsequent optimization stage in a unified data form, significantly improving the intelligence and automation of scheme generation.

[0015] 2. This application utilizes a state-space model with a selection mechanism to iteratively update the layout state sequence, explicitly incorporating multiple constraints such as visitor flow continuity, area matching, passage width, visibility, and safe evacuation into a unified optimization framework. By selectively updating different state dimensions, it achieves fine-tuning of key spatial parameters. Furthermore, the optimization results are simultaneously mapped to two-dimensional planar layout data and three-dimensional layout models, realizing the integrated output of "structured optimization results - 2D drawings - 3D scene". Compared with existing solutions, this application has significant innovations in the demand acquisition method, layout representation, and dynamic optimization mechanism. It can not only automatically generate multiple candidate solutions that meet multiple constraints, but also improve the utilization rate of exhibition hall space and the consistency and controllability of visitor experience design. Attached Figure Description

[0016] Figure 1 This is a flowchart of an artificial intelligence-based exhibition hall space planning and generation method according to the present invention. Detailed Implementation

[0017] Embodiments of the present invention are described in detail below. Examples of these embodiments are illustrated in the accompanying drawings, wherein the same or similar symbols denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0018] like Figure 1 The method for generating exhibition hall space planning based on artificial intelligence, as shown, includes the following steps: S1. Obtain basic data of the exhibition hall space and functional requirements data of the exhibition hall and preprocess them to obtain the layout generation input feature vector; In an optional embodiment, basic data of the exhibition hall space and functional requirements data of the exhibition hall are acquired and preprocessed to obtain the layout generation input feature vector, as follows: Acquire basic spatial data for the exhibition hall, including the architectural floor plan, floor height information, and permitted exhibition area information. Establish a Cartesian coordinate system based on the architectural floor plan, and discretize the planar boundaries of the exhibition hall to obtain a set of boundary points. ,in, and Representing boundary points respectively In the Cartesian coordinate system, the set B is used to represent the planar boundary of the exhibition hall; The floor height information of the exhibition hall is quantified to obtain a set of floor height parameters. Based on the permitted exhibition area information, the permitted exhibition areas are labeled in a Cartesian coordinate system to obtain the set of permitted exhibition areas. ,in, This represents the floor height parameter of the r-th region. This represents the subset of regions where the s-th exhibition area is allowed to lie in the plane; Obtain and categorize the functional requirements data of the exhibition hall to obtain a set of exhibition area types. Exhibition area requirements Collection of exhibits Audience flow constraint set and the set of safety evacuation constraints ,in, This indicates the type identifier for the q-th type of exhibition area. Indicates the type of exhibition area The corresponding target area parameters, Indicates the type of exhibition area The corresponding number of exhibits This represents the constraint parameter for the r-th audience flow line. This represents the safety evacuation constraint parameter corresponding to the r-th audience flow line constraint; The basic data of the exhibition hall space and the functional requirements data of the exhibition hall are processed into feature vectors to construct spatial feature vectors and requirement feature vectors. The spatial feature vectors and requirement feature vectors are then concatenated in a predetermined order to form the layout generation input feature vector. .

[0019] S2. Input the layout generation input feature vector into a pre-trained large-scale language model to generate an initial text description of the exhibition hall layout through an autoregressive generation method. In an optional embodiment, the layout generation input feature vector is input into a pre-trained large-scale language model to generate an initial text description of the exhibition hall layout through an autoregressive generation method, as follows: Generate input feature vectors for the layout. A linear transformation is performed to obtain a conditional vector, which is then input into a pre-trained large-scale language model. The pre-trained large-scale language model uses GPT-5, and a hidden state sequence is obtained through multi-layer transformation. , ; Based on hidden state sequences The initial layout description word sequence is generated using an autoregressive generation method. , ,in, This represents the u-th word in the output sequence, and the autoregressive generation satisfies the conditional probability decomposition: ; in, Indicates that given existing output tokens and layout to generate input feature vectors Generate the u-th output word under the condition of The conditional probability; The word sequence describing the initial layout Decoding is performed to obtain the initial text description of the exhibition hall layout. ; In an optional embodiment, the initial text description Includes collections of exhibition area types Each exhibition area type Functional area names and exhibition area requirements Target area parameters Corresponding target area description and exhibit quantity set Parameters of the number of exhibits Corresponding descriptions of the number of exhibits and the set of constraints on visitor flow. Audience Flow Constraint Parameters The corresponding audience flow sequence description and the set of safety evacuation constraints Medium safety evacuation constraint parameters The corresponding safety evacuation requirements are described.

[0020] S3. Based on the initial text description, perform semantic parsing to generate a set of functional areas, a set of edges, a set of area attributes, and a set of edge attributes to obtain the exhibition hall layout diagram structure; In an optional embodiment, semantic parsing is performed based on the initial text description to generate a set of functional regions, a set of edges, a set of region attributes, and a set of edge attributes, thereby obtaining the exhibition hall layout diagram structure, as follows: Initial text description Sentence segmentation is performed to obtain a sentence sequence. , ,in, This represents the i-th sentence; For sentence sequence Each sentence in Semantic segmentation is performed to obtain a set of semantic segments. ; Based on the set of exhibition area types Exhibition area requirements Collection of exhibits Audience flow constraint set and the set of safety evacuation constraints For semantic fragment sets Semantic parsing is performed on each semantic fragment to obtain a set of region-related information triples. : ; in, This indicates the functional area identifier referred to by the semantic fragment. Indicates the functional area type. Indicates the target area parameter. This parameter represents the number of exhibits corresponding to that functional area. Construct a set of functional regions R based on the set of triples C containing region-related information. For each functional region in the set R... Establish regional attribute vectors ,in, , Representation and Functional Areas Corresponding exhibition area types, Representation and Functional Areas The corresponding target area parameters, Representation and Functional Areas The corresponding exhibit quantity parameter will be used to generate all area attribute vectors. Combined to form a set of regional attributes ; From the initial text description The audience flow sequence is described in the parsing, and the functional area access sequence is constructed. Access sequence based on functional area Construct an adjacency matrix for adjacent access relationships Among them, when the functional areas in the audience flow With functional areas When functional areas are adjacent and have a direct connection in the access sequence, let Otherwise And based on the adjacency matrix Obtain the edge set , ; Based on audience flow constraint set and safe evacuation constraint set set of opposite edges Each edge in Constructing edge attribute vectors Obtain the set of edge attributes , ,in, At least include parameters related to audience flow constraints. The corresponding streamline direction parameters and safety evacuation constraint parameters Corresponding evacuation requirement parameters; Set the functional region set R and the edge set Regional attribute set and edge attribute set Combine to generate the exhibition hall layout structure. , .

[0021] S4. Generate the exhibition hall layout state vector based on the exhibition hall layout diagram structure, and arrange the exhibition hall layout state vectors of all visited locations in the order of visit to form an initial state sequence; In an optional embodiment, an exhibition hall layout state vector is generated based on the exhibition hall layout diagram structure, and the exhibition hall layout state vectors of all visited locations are arranged in the order of visit to form an initial state sequence, as follows: Based on the exhibition hall layout structure Set of functional areas Perform spatial parameter initialization for each functional area. Determine the coordinates of the reference point To obtain the set of reference points , ; For each functional area Determine the set of boundary contour points , ,in, Indicates functional area The u-th boundary point, and The coordinate components of the boundary points are represented based on the boundary point set. Calculation functional area Geometric area parameters and with the region attribute vector Target area parameters By combining the features, we obtain the area feature vector. , ; For each functional area Based on the set of boundary contour points Calculate the width parameter of the minimum bounding rectangle and height parameters and obtain shape parameters , ,in, , , Indicates functional area The aspect ratio shape parameters will As a functional area Shape characteristics; Based on edge set and edge attribute set For each edge Get the channel length parameter of the corresponding channel and channel width parameters Construct channel feature vectors ,in, Indicates functional area With functional areas The center distance between them Indicates the connection function area With functional areas The channel width parameter; Based on the set of permitted exhibition areas Set of floor height parameters Audience flow constraint set and the set of safety evacuation constraints For each functional area Determine the corresponding floor height parameters and allowed exhibition area index parameters , ,in, Used to indicate functional areas The permitted exhibition area And based on preset population density parameters Calculation functional area Crowd capacity parameters The population density parameter Based on expert experience, : ; Each functional area The functional areas are formed by combining the reference point coordinates, width parameters, height parameters, shape parameters, geometric area parameters, target area parameters, exhibit quantity parameters, floor height parameters, and crowd capacity parameters. state vector : ; Functional area access sequence and edge set For each access location v, determine the corresponding functional area and the set of edges directly connected to the functional area. For each edge Obtain the corresponding channel feature vector Then perform aggregation operations to obtain the edge aggregation vector. ; Functional area access sequence Construct a showroom layout state vector for each access location v. The exhibition hall layout state vector is composed of the corresponding state vector. and edge aggregation vector composition, Arrange the exhibition hall layout state vectors of all visited locations in the order of visit to form an initial state sequence. , .

[0022] S5. The initial state sequence is iteratively updated based on the selective state space model to obtain the optimized exhibition hall layout state sequence. In an optional embodiment, the initial state sequence is iteratively updated based on a selective state-space model to obtain an optimized exhibition hall layout state sequence, as follows: Based on the exhibition hall layout structure and the initial state sequence Construct the optimization objective function ; For each iteration step and access location Initialize the hidden state vector sequence and define the selection vector sequence ,in, Used to indicate the layout state vector Are the dimensions in the iteration step? The middle part has been updated; In each iteration step Inside, for access location The current layout state vector Hidden state vector and selection vector Input a selective state-space model, which is based on the Mamba architecture, to obtain the updated hidden state vector. and candidate layout state vector ; Based on selection vector Update the layout state vector to obtain the iteration steps. Layout state vector The update rules are as follows: ; in, This indicates element-wise multiplication. This represents a vector where each component is 1. After updating all access locations, the latest state sequence is obtained. Calculate the corresponding objective function value. If satisfied and ,in, The convergence threshold, To minimize the number of iterations, the latest state sequence is... As an optimized exhibition hall layout state sequence ; In an optional embodiment, the optimization objective function consists of objective terms related to audience flow continuity constraints, regional area constraints, and passage width constraints, and is expressed as follows: ; in, , , These are the weighting coefficients. This indicates the target terms related to the continuity constraint of the audience flow. This represents the objective terms related to the area constraint of the region. This indicates the target items related to the passage width constraint; Audience flow continuity constraint related objective items Functional area access sequence and channel length parameters Perform the calculation: ; in, Indicates the state sequence The corresponding functional area is accessed at position v in the layout. Access Location Functional areas The center distance parameter between them This represents the preset target streamline step size parameter; Passage width constraint related target items Based on the edge set and channel feature vectors Perform the calculation: ; in, Indicates the state sequence Connecting functional areas under the corresponding layout With functional areas The channel width parameter, The minimum passage width threshold is set based on expert experience; Area-related target items The calculation method is as follows: ; in, Indicates the state sequence Functional areas under corresponding layout The geometric area parameters, Representation and Functional Areas The corresponding target area parameters.

[0023] S6. Generate the exhibition hall floor plan layout result based on the optimized exhibition hall layout state sequence, and generate a 3D scene based on the exhibition hall floor plan layout result; In an optional embodiment, an exhibition hall floor plan layout result is generated based on the optimized exhibition hall layout state sequence, and a 3D scene is generated based on the exhibition hall floor plan layout result, as follows: Functional area access sequence and functional area set For each functional area From the optimized exhibition hall layout state sequence Extract the corresponding region state vector The region state vector Represented as: ; in, and Indicates functional area Optimization of reference point coordinate components in a Cartesian coordinate system and Indicates functional area Optimize the width and height parameters. Indicates functional area Optimized shape parameters, Indicates functional area Optimize geometric area parameters, Indicates functional area Target area parameters, Indicates functional area The number of exhibits Indicates functional area The corresponding floor height parameters, Indicates functional area Optimize the crowd capacity parameters; Based on the region state vector Includes optimized reference point coordinates Width parameters and height parameters In a Cartesian coordinate system, each functional area is represented by Construct a set of axis-aligned rectangle boundary points : ; in, , , , ; For all functional areas The set of rectangular boundary points is summarized to form the set of area boundaries for the exhibition hall's floor plan. , Each element is formed by connecting four boundary points in sequence to create a closed polygon, which is used to represent the functional areas in the exhibition hall's floor plan. The area occupied by the plane; Based on edge set And optimize reference point coordinates , For each edge Construct a set of channel polyline segments: ; in, , , This indicates the connection between functional areas in the exhibition hall's floor plan. With functional areas The channel segment; For all edges The corresponding set of broken line segments of the passageway are summarized to form the passageway set of the exhibition hall floor plan. : ; in, Each element Used to indicate connecting passages between functional areas in the floor plan of an exhibition hall; Based on the set of permitted exhibition areas and channel set Construct a connectivity matrix for the exhibition hall's floor plan. ,in: ; in, Indicates functional area With functional areas Effective connectivity signage in the exhibition hall layout; Set of region boundaries Channel set and connectivity matrix Combine to generate the exhibition hall floor plan. , And generate a 3D scene based on the exhibition hall floor plan; A three-dimensional rectangular coordinate system is established based on the exhibition hall floor plan. The exhibition hall floor plan is used as the horizontal reference plane, and a height direction is set. According to the rectangular boundary vertices and planar positions of each functional area and the corresponding floor height, each functional area is stretched along the height direction in the three-dimensional rectangular coordinate system to generate a three-dimensional polyhedron composed of a bottom surface, a top surface, and side surfaces, to represent the three-dimensional spatial occupancy range of each functional area. According to the connectivity between functional areas and the start and end positions of the passage centerline and the planar width parameters, a passage block with a predetermined height and width is generated in the three-dimensional rectangular coordinate system and is connected to or intersects with the boundary of the corresponding functional area, to represent the three-dimensional passage space between functional areas. Within the three-dimensional space of each functional area, several exhibition equipment arrangement anchor points are determined according to the center position, planar boundary, floor height, and number of exhibits of the functional area. The exhibition equipment arrangement anchor points include planar position and height position, which are used to indicate the arrangement position of the exhibition equipment in three-dimensional space. For each exhibit placement anchor point, the exhibit type and geometric dimension parameters are assigned based on the exhibition area type and exhibit quantity of the functional area to which the anchor point belongs. An exhibit object associated with the exhibit placement anchor point is generated, forming an exhibit object set in each functional area. The three-dimensional polyhedra of all functional areas, the passage blocks between all functional areas, and the entire exhibit object set are summarized to form a three-dimensional layout object set for the exhibition hall. The geometric information of each three-dimensional object, as well as attribute information such as exhibition area type, target area, exhibit quantity, floor height, and number of visitors, are organized according to a preset data structure to construct a three-dimensional layout model data for the exhibition hall space planning. This data is used to output the three-dimensional layout model of the exhibition hall space planning and its corresponding digital planning data.

[0024] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for generating exhibition hall space planning based on artificial intelligence, characterized in that, Includes the following steps: S1. Obtain and classify the functional requirements data of the exhibition hall to obtain the set of exhibition area types, the set of exhibition area area requirements, the set of exhibit quantities, the set of visitor flow constraints, and the set of safety evacuation constraints; obtain the layout generation input feature vector; S2. Input the layout generation input feature vector into a pre-trained large-scale language model to generate an initial text description of the exhibition hall layout through an autoregressive generation method. S3. Based on the initial text description, perform semantic parsing to generate a set of functional areas, a set of edges, a set of area attributes, and a set of edge attributes to obtain the exhibition hall layout diagram structure; S4. Generate the exhibition hall layout state vector based on the exhibition hall layout diagram structure, and arrange the exhibition hall layout state vectors of all visited locations in the order of visit to form an initial state sequence; S5. The initial state sequence is iteratively updated based on the selective state space model to obtain the optimized exhibition hall layout state sequence. S6. Generate the exhibition hall floor plan layout result based on the optimized exhibition hall layout state sequence, and generate a 3D scene based on the exhibition hall floor plan layout result.

2. The exhibition hall space planning and generation method based on artificial intelligence according to claim 1, characterized in that, Acquire and preprocess basic data of the exhibition hall space and functional requirements data of the exhibition hall to obtain the layout and generate the input feature vector, as follows: Acquire basic data of the exhibition hall space, including the architectural floor plan of the exhibition hall, the floor height information of the exhibition hall, and the information on the permitted exhibition area. Establish a Cartesian coordinate system based on the architectural floor plan of the exhibition hall, and discretize the planar boundary of the exhibition hall to obtain a set of boundary points. The floor height information of the exhibition hall is quantified to obtain a set of floor height parameters. Based on the information of the allowed exhibition area, the area is marked in a plane rectangular coordinate system to obtain a set of allowed exhibition areas. The basic data of the exhibition hall space and the data of the exhibition hall functional requirements are processed into feature vectors to construct spatial feature vectors and requirement feature vectors. The spatial feature vectors and requirement feature vectors are then concatenated in a predetermined order to form the layout generation input feature vector.

3. The exhibition hall space planning and generation method based on artificial intelligence according to claim 1, characterized in that, The layout generation input feature vector is input into a pre-trained large-scale language model, which generates an initial text description of the exhibition hall layout through an autoregressive generation method, as follows: The layout-generated input feature vector is linearly transformed to obtain a conditional vector. The conditional vector is then input into a pre-trained large-scale language model, and the hidden state sequence is obtained through multi-layer transformation. The initial layout description word sequence is generated based on the hidden state sequence using an autoregressive generation method. The initial text description of the exhibition hall layout is obtained by decoding the word sequence of the initial layout description.

4. The exhibition hall space planning and generation method based on artificial intelligence according to claim 3, characterized in that, The initial text description includes the functional area name for each exhibition area type in the exhibition area type set, the target area description corresponding to the target area parameter in the exhibition area area requirement set, the exhibit quantity description corresponding to the exhibit quantity parameter in the exhibit quantity set, the audience flow order description corresponding to the audience flow constraint parameter in the audience flow constraint set, and the safety evacuation requirement description corresponding to the safety evacuation constraint parameter in the safety evacuation constraint set.

5. The exhibition hall space planning and generation method based on artificial intelligence according to claim 1, characterized in that, Based on the initial text description, semantic parsing is performed to generate a set of functional regions, a set of edges, a set of region attributes, and a set of edge attributes, thus obtaining the exhibition hall layout structure, as follows: The initial text description is segmented into sentences to obtain a sentence sequence, and each sentence in the sentence sequence is further divided into semantic segments to obtain a set of semantic segments; Based on the set of exhibition area types, the set of exhibition area area requirements, the set of exhibit quantities, the set of audience flow constraints, and the set of safety evacuation constraints, semantic parsing is performed on each semantic segment in the set of semantic segments to obtain a set of triplets of region-related information. Construct a set of functional regions based on a set of triples of region-related information, establish a region attribute vector for each functional region in the set of functional regions, and combine all region attribute vectors to form a set of region attributes. Parse the audience flow sequence description from the initial text description, construct the functional area access sequence, build an adjacency matrix of adjacent access relationships based on the functional area access sequence, and obtain the edge set based on the adjacency matrix. Based on the audience flow constraint set and the safety evacuation constraint set, an edge attribute vector is constructed for each edge in the edge set to obtain the edge attribute set; The functional area set, edge set, area attribute set, and edge attribute set are combined to generate the exhibition hall layout diagram structure.

6. The exhibition hall space planning and generation method based on artificial intelligence according to claim 1, characterized in that, Based on the exhibition hall layout diagram, generate exhibition hall layout state vectors, and arrange the exhibition hall layout state vectors of all visited locations in the order of visit to form an initial state sequence, as follows: Based on the exhibition hall layout diagram, the spatial parameters of the functional area set are initialized, and the reference point coordinates are determined for each functional area to obtain the reference point set. For each functional area, a set of boundary contour points is determined. The geometric area parameters of the functional area are calculated based on the set of boundary points and combined with the target area parameters in the area attribute vector to obtain the area feature vector. For each functional area, the width and height parameters of the minimum bounding rectangle are calculated based on the set of boundary contour points, and the shape parameters are obtained, which will be used as the shape features of the functional area. Based on the edge set and edge attribute set, for each edge, obtain the channel length parameter and channel width parameter of the corresponding channel, and construct the channel feature vector; Based on the set of allowed exhibition areas, the set of floor height parameters, the set of audience flow constraints, and the set of safety evacuation constraints, the corresponding floor height parameters and allowed exhibition area index parameters are determined for each functional area, and the crowd capacity parameters of the functional area are calculated based on the preset crowd density parameters. The reference point coordinates, width parameters, height parameters, shape parameters, geometric area parameters, target area parameters, exhibit quantity parameters, floor height parameters, and crowd capacity parameters of each functional area are combined to form the state vector of the functional area. Based on the functional area access sequence and edge set, for each access position, the corresponding functional area and the set of edges directly connected to the functional area are determined. For each edge, the corresponding channel feature vector is obtained and the aggregation operation is performed to obtain the edge aggregation vector. For each access location in the functional area access sequence, a showroom layout state vector is constructed. The showroom layout state vector consists of a corresponding state vector and an edge aggregation vector. The showroom layout state vectors of all access locations are arranged in the access order to form an initial state sequence.

7. The exhibition hall space planning and generation method based on artificial intelligence according to claim 1, characterized in that, The initial state sequence is iteratively updated based on the selective state-space model to obtain the optimized exhibition hall layout state sequence, as follows: An optimization objective function is constructed based on the exhibition hall layout diagram and the initial state sequence. For each iteration step and the visited position, initialize the hidden state vector sequence and define the selection vector sequence; In each iteration step, for the access position, the current layout state vector, hidden state vector and selection vector are input into the selective state space model to obtain the updated hidden state vector and candidate layout state vector. The layout state vector is updated based on the selection vector. After updating all access positions, the latest state sequence is obtained. The corresponding optimization objective function value is calculated. If the absolute value of the difference between the optimization objective function value of the current iteration step and the optimization objective function value of the previous iteration step is less than or equal to the convergence threshold and the current iteration step number is greater than or equal to the minimum iteration step number, then the latest state sequence is taken as the optimized exhibition hall layout state sequence.

8. The exhibition hall space planning and generation method based on artificial intelligence according to claim 1, characterized in that, The optimization objective function consists of objective terms related to audience flow continuity constraints, regional area constraints, and passage width constraints.

9. The exhibition hall space planning and generation method based on artificial intelligence according to claim 1, characterized in that, The exhibition hall floor plan is generated based on the optimized exhibition hall layout state sequence, and a 3D scene is generated based on the exhibition hall floor plan, as follows: Based on the functional area access sequence and the functional area set, the corresponding area state vector is extracted from the optimized exhibition hall layout state sequence for each functional area. Based on the coordinates of the optimized reference points, the width parameter, and the height parameter contained in the region state vector, an axis-aligned rectangular boundary point set is constructed for each functional region in a Cartesian coordinate system. The set of rectangular boundary points of all functional areas is summarized to form the set of area boundaries for the exhibition hall's floor plan; Based on the edge set and the optimized reference point coordinates, a channel polyline segment set is constructed for each edge; The set of all the channel polyline segments corresponding to all edges is summarized to form the channel set of the exhibition hall floor plan; Based on the set of allowed exhibition areas and the set of passageways, a connectivity matrix for the exhibition hall layout is constructed. The exhibition hall floor plan is generated by combining the set of regional boundaries, the set of channels, and the connectivity matrix, and a 3D scene is generated based on the floor plan.

Images