Building indoor plane layout optimization design method based on machine learning
By using a machine learning-based modular design approach, a reasonable building interior floor plan is generated, which solves the problem of unreasonable design caused by relying on human experience. It achieves self-learning and optimization, reducing the number of design revisions.
Patent Information
- Application Number
- CN202511502561.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-10-21
- Publication Date
- 2026-02-06
AI Technical Summary
The existing interior floor plan design of buildings relies on human experience, resulting in unreasonable design results that require multiple revisions and lack rationality.
It employs machine learning-based multimodal data fusion, real-time interaction, building interior layout selection, decision-making, and AI self-learning modules. It generates building interior floor plans through BIM models, multi-party opinion integration, and a hierarchical optimization engine, supporting self-learning and feedback optimization.
It enhances the self-learning and optimization capabilities of building interior floor plan design, shortens the design process, and reduces repeated revisions of design schemes.
Smart Images
Figure CN121479880A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The application relates to the technical field of intelligent design of buildings, in particular to a building indoor plane layout optimization design method based on machine learning. BACKGROUND
[0002] The building indoor plane layout refers to a design process of systematically planning and organizing the interior space of a building in the horizontal direction, and the core is to scientifically arrange functional areas, moving lines, structures and equipment and other elements through a two-dimensional plane graph. Before planning and decoration of the building interior is required in the field of buildings, the building indoor plane layout is usually designed in advance, component design drawings are made, and decoration is handled by construction personnel according to the design drawings.
[0003] However, the existing building indoor plane layout design method has the following problems: mainly through the experience judgment of designers and the combination of customer needs, the building indoor plane layout is designed and planned through special design software. This design method relying on human subjective experience has great limitations, is easily affected by the supervisor and knowledge reserve of the designer, and is prone to cause unreasonable design results. Moreover, the design drawings often need to be modified several times to determine a usable scheme, and the rationality of the building indoor plane layout design is insufficient. Therefore, a corresponding technical solution is needed to solve the existing technical problems. SUMMARY
[0004] The purpose of the present application is to provide a building indoor plane layout optimization design method based on machine learning, which solves the problem that the building indoor plane layout is designed and planned through special design software mainly through the experience judgment of designers and the combination of customer needs. This design method relying on human subjective experience has great limitations, is easily affected by the supervisor and knowledge reserve of the designer, and is prone to cause unreasonable design results. Moreover, the design drawings often need to be modified several times to determine a usable scheme, and the rationality of the building indoor plane layout design is insufficient. This technical problem, the present application designs a building indoor plane layout optimization design method based on autonomous learning. The optimization design method includes five modules: a multi-modal data fusion module, a real-time interaction module, a building indoor plane layout selection module, a building indoor plane layout decision module and an AI data collection self-learning module. The plane layout design of the building indoor plane layout, information interaction, scheme generation, scheme determination and self-learning of the building indoor plane layout design are realized, and are fed back to the multi-modal data fusion module. The self-learning and optimization ability of the plane layout optimization design is improved, the process of the building indoor plane layout optimization design is shortened, and the problem of repeated correction of the design scheme caused by information difference is reduced.
[0005] To achieve the above object, the application provides the following technical scheme: a building indoor plane layout optimization design method based on machine learning, which comprises five modules, namely a multi-modal data fusion module, a real-time interaction module, a building indoor plane layout selection module, a building indoor plane layout decision module and an AI data acquisition self-learning module, wherein the multi-modal data fusion module comprises a BIM model unit, a multi-party opinion data integration unit and a hierarchical optimization engine unit, the BIM model unit is used for analyzing and integrating building structure data and providing a design platform, the multi-party opinion data integration unit is used for integrating designer data and customer opinion data, and the hierarchical optimization engine unit is used for optimizing the building indoor plane layout, the BIM model unit, the multi-party opinion data integration unit and the hierarchical optimization engine unit jointly generate a building indoor plane model, the real-time interaction module is used for real-time data sharing and opinion exchange between designers and customers, the building indoor plane layout selection module generates a building indoor plane layout scheme for selection, the building indoor plane layout decision module selects the scheme provided by the building indoor plane layout selection module, and the AI data acquisition self-learning module autonomously learns the entire design process and feeds back data.
[0006] As a preferred mode of the application, the hierarchical optimization engine unit comprises a global planning layer and a local refinement layer, the global planning layer generates a functional partition scheme by using a PPO algorithm, and the specific PPO algorithm is as follows:
[0007] , wherein, is an action probability ratio, is usually 0.2, the local refinement layer optimizes furniture placement based on a graph attention network and processes spatial adjacency constraints.
[0008] As a preferred mode of the application, the real-time interaction module is a development parameterization adjustment interface, which supports users to modify partition boundaries through gesture operation and realizes AR visual preview of the layout scheme by deploying a lightweight YOLO model.
[0009] As a preferred mode of the application, the building indoor plane layout selection module generates building indoor plane layout schemes in combination with the multi-modal data fusion module and the real-time interaction module, and the building indoor plane layout schemes are three or more.
[0010] As a preferred mode of the application, the building indoor plane layout decision module mainly considers customer opinions and secondarily considers designer opinions to make an optimal scheme.
[0011] As a preferred mode of the present application, the AI data acquisition self-learning module is based on machine learning, receives and integrates the data of the building indoor layout optimization design process, realizes model iteration optimization, realizes knowledge base updating, environment data acquisition and model retraining, adjusts the learning path in real time based on user behavior data, and feeds back to the building indoor layout, and feeds back the learning data results to the multi-modal data fusion module to optimize the multi-modal data fusion module.
[0012] Compared with the prior art, the present application has the following advantages:
[0013] The present application designs a building indoor layout optimization design method based on autonomous learning, which includes five modules: multi-modal data fusion module, real-time interaction module, building indoor layout selection module, building indoor layout decision module and AI data acquisition self-learning module, which realizes the plane layout design, information interaction, scheme generation, scheme determination and self-learning of building indoor layout design, and feeds back to the multi-modal data fusion module, improves the self-learning and optimization ability of the plane layout optimization design, shortens the process of building indoor layout optimization design, and reduces the problem of repeated correction of design scheme caused by information difference. BRIEF DESCRIPTION OF DRAWINGS
[0014] Figure 1 The method system diagram of the present application is shown in the figure;
[0015] Figure 2 The multi-modal data fusion module of the present application is shown in the figure. DETAILED DESCRIPTION
[0016] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.
[0017] Please refer to Figures 1-2The application provides a technical solution: a building indoor plane layout optimization design method based on machine learning, which comprises five modules, namely, a multi-modal data fusion module, a real-time interaction module, a building indoor plane layout selection module, a building indoor plane layout decision module and an AI data acquisition self-learning module. The multi-modal data fusion module comprises a BIM model unit, a multi-party opinion data integration unit and a hierarchical optimization engine unit. The BIM model unit analyzes and integrates building structure data and provides a design platform. The multi-party opinion data integration unit integrates designer data and customer opinion data. The hierarchical optimization engine unit optimizes the building indoor plane layout. The BIM model unit, the multi-party opinion data integration unit and the hierarchical optimization engine unit jointly generate a building indoor plane model. The real-time interaction module provides real-time data sharing and opinion exchange for designers and customers. The building indoor plane layout selection module generates a building indoor plane layout scheme for selection. The building indoor plane layout decision module selects the scheme provided by the building indoor plane layout selection module. The AI data acquisition self-learning module autonomously learns the entire design process and feeds back data. Through the five modules, the plane layout design of the building indoor plane layout, information interaction, scheme generation, scheme determination and self-learning of the building indoor plane layout design are realized, and the data is fed back to the multi-modal data fusion module, the self-learning and optimization capability of the plane layout optimization design is improved, the process of the building indoor plane layout optimization design is shortened, and the problem of repeated correction of the design scheme caused by information difference is reduced.
[0018] Further improvement, as shown in Figure 1 , the hierarchical optimization engine unit comprises a global planning layer and a local refinement layer. The global planning layer generates a functional partition scheme by using a PPO algorithm. The specific PPO algorithm is as follows:
[0019] , wherein, is the action probability ratio, is usually 0.2. The local refinement layer optimizes furniture placement based on a graph attention network and processes spatial adjacency constraints.
[0020] Further improvement, as shown in Figure 1 , the real-time interaction module is a parameterized adjustment interface, which supports users to modify partition boundaries through gesture operation and to realize AR visual preview of the layout scheme by deploying a lightweight YOLO model.
[0021] Further improvement, as shown in Figure 1 , the building indoor plane layout selection module generates a building indoor plane layout scheme in combination with the multi-modal data fusion module and the real-time interaction module. The building indoor plane layout scheme is three or more.
[0022] Further improvement, as shown in Figure 1As shown, the building indoor layout decision module is mainly based on customer opinions and supplemented by designer opinions to make the optimal scheme.
[0023] Specifically, the AI data collection self-learning module is based on machine learning, receives and integrates the data of the building indoor layout optimization design process, realizes model iteration optimization, realizes knowledge base updating, environment data collection and model retraining, adjusts the learning path in real time based on user behavior data, and feeds back to the building indoor layout, and feeds back the learning data results to the multi-modal data fusion module to optimize the multi-modal data fusion module.
[0024] In the description of the present application, it should be understood that the terms "coaxial", "bottom", "one end", "top", "middle", "the other end", "upper", "one side", "top", "inner", "front", "central", "both ends" and the like indicate the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as limiting the present application.
[0025] In addition, the terms "first", "second", "third", "fourth" are only for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the technical features indicated, so that the features limited by "first", "second", "third", "fourth" can be explicitly or implicitly included at least one of the features.
[0026] In the present application, unless otherwise explicitly specified and limited, the terms "mounting", "setting", "connecting", "fixing", "threading" and the like should be understood broadly, for example, it can be fixedly connected, or it can be detachably connected, or it can be integrated; it can be mechanically connected, or it can be electrically connected; it can be directly connected, or it can be indirectly connected through an intermediate medium; it can be the internal communication of two elements or the interaction relationship between two elements, unless otherwise explicitly limited, the above-mentioned terms in the present application can be understood according to the specific meaning of the above-mentioned terms in the present application by the person skilled in the art.
[0027] Finally, it should be pointed out that: the above-mentioned only for the preferred embodiments of the present application, and does not limit the present application, although the present application has been described in detail with reference to the foregoing embodiments, for the person skilled in the art, it still can modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part of the technical features. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.
Claims
1. A method for optimizing the interior floor plan layout of buildings based on machine learning, characterized by: This design method comprises five modules: a multimodal data fusion module, a real-time interaction module, a building interior layout selection module, a building interior layout decision module, and an AI data acquisition self-learning module. The multimodal data fusion module includes a BIM model unit, a multi-party opinion data integration unit, and a hierarchical optimization engine unit. The BIM model unit analyzes and integrates building structural data and provides a design platform. The multi-party opinion data integration unit integrates designer data and client feedback data. The hierarchical optimization engine unit optimizes the building interior layout. The BIM model unit, multi-party opinion data integration unit, and hierarchical optimization engine unit jointly generate a building interior layout model. The real-time interaction module allows designers and clients to share data and exchange opinions in real time. The building interior layout selection module generates selectable building interior layout schemes. The building interior layout decision module selects from the schemes provided by the building interior layout selection module. The AI data acquisition self-learning module autonomously learns throughout the design process and provides data feedback.
2. The building interior floor plan optimization design method based on machine learning according to claim 1, characterized in that: The hierarchical optimization engine unit includes a global planning layer and a local refinement layer. The global planning layer uses the PPO algorithm to generate a functional partitioning scheme. The specific PPO algorithm is as follows: ,in, The probability ratio of actions. Typically set to 0.2, the local refinement layer optimizes furniture placement based on graph attention networks and handles spatial adjacency constraints.
3. The building interior floor plan optimization design method based on machine learning according to claim 1, characterized in that: The real-time interactive module is designed to develop a parameterized adjustment interface, allowing users to modify partition boundaries and deploy lightweight YOLO models to achieve AR visualization previews of layout schemes through gesture operations.
4. The building interior floor plan optimization design method based on machine learning according to claim 1, characterized in that: The building interior floor plan selection module, in conjunction with the multimodal data fusion module and the real-time interaction module, generates three or more building interior floor plan layout schemes.
5. The building interior floor plan optimization design method based on machine learning according to claim 1, characterized in that: The building interior layout decision module prioritizes customer feedback and supplements it with designer opinions to produce the optimal solution.
6. The building interior floor plan optimization design method based on machine learning according to claim 1, characterized in that: The AI data acquisition self-learning module is supported by machine learning. It receives and integrates data from the process of optimizing the design of building interior floor plans, realizes model iterative optimization, updates the knowledge base, collects environmental data and retrains the model, adjusts the learning path in real time based on user behavior data and feeds it back to the building interior floor plan, and feeds the learning data results back to the multimodal data fusion module to optimize the multimodal data fusion module.