A method and system for generating a hospital building planning type plan

Abstract

The application relates to the technical field of electric digital data processing, and discloses a hospital building planning plane generation method and system, which comprises the following steps: acquiring an initial topology matrix reflecting function flow line isolation rules and cooperation timeliness indexes and a two-dimensional red line boundary matrix representing boundary working conditions; calculating geometric overlap deviation of current round plane layout data and the two-dimensional red line boundary matrix, and introducing the geometric overlap deviation into coordinate correction calculation as a boundary penalty term to obtain an absolute geometric coordinate matrix of each functional room; when the iteration round reaches a set upper limit and the energy residual error is higher than a set threshold, calling a cell division operator to perform recursive bisection topology division on the two-dimensional red line boundary matrix, and outputting the absolute geometric coordinate matrix; and the application eliminates the calculation convergence deadlock under high-dimensional medical constraints and in-situ eliminates the physical overlap interference of each space unit at the data processing layer.

Description

Technical Field

[0001] This invention relates to a method and system for generating hospital building planning floor plans, belonging to the field of electronic digital data processing technology. Background Technology

[0002] Currently, data processing for building plan generation typically first represents each functional unit as discrete data nodes, then records the association constraints between nodes using a topological adjacency matrix containing various connectivity weights. The digital processor then performs iterative calculations based on this matrix, converting the collaborative relationships between nodes into positional relationships in two-dimensional space, serving as the computational basis for generating the plan layout. In the process of converting topological relationships into geometric coordinates, existing methods usually employ gradient descent, simultaneously calculating the distance and overlap between nodes, and adjusting the update step size of the centroid coordinates of each spatial unit based on the convergence residual of the objective function. Such methods require that the coordinate changes maintain... Continuity and stability are essential for the successful joint solution of multiple constraints. In application scenarios such as hospital buildings with a large number of nodes and complex constraint relationships, different functional rooms may have strong requirements for cooperation and proximity, but may also require strict isolation due to factors such as cleanliness and personnel flow. Traditional methods usually add these constraints of different natures into the same calculation process according to their respective weights. When the number of nodes and the constraint density increase, various constraints are prone to mutual influence, causing the calculation process to oscillate repeatedly, the convergence speed to decrease, or to get stuck in local solutions. This results in a plan layout with overlapping functional room outlines, conflicting positions, or boundary crossings.

[0003] To address the aforementioned issues, existing improvement methods typically employ methods such as increasing the local penalty coefficient or the number of discrete mesh samples to attempt to reduce conflicts between spatial units. However, under irregular rigid boundary conditions, the direct superposition of multiple penalty terms can easily cause the objective function to increase rapidly, resulting in drastic changes in the solution state. The computational matrix may exhibit singularities, leading to instability in the iterative process or even failure to converge. Existing AI-based assisted design methods also suffer from similar problems. For instance, Chinese invention patent application CN121389746A discloses a method and system for generating architectural planning-type planes based on a latent diffusion model. This method generates planar layouts through denoising prediction using the latent diffusion model, offering flexibility in terms of planar form and spatial combination. However, in medical building scenarios, functional flow isolation, departmental collaboration efficiency, and red-line boundaries are all constraints that need to be explicitly satisfied. The latent diffusion model generates results using a probabilistic sampling method, making it difficult to establish a stable and definite correspondence with these physical boundary conditions and flow topology rules.

[0004] Therefore, how to handle conflicting proximity constraints and isolation constraints separately during data processing, and how to continuously correct the coordinate update process based on spatial geometric deviations, so that discrete topological data can be stably converted into an absolute geometric coordinate matrix under irregular boundary conditions, remains a problem that existing hospital building plan generation technology needs to solve. Summary of the Invention

[0005] To address the problems in the background art, the technical solution of the present invention is as follows: A method for generating hospital building planning floor plans, comprising the following steps: Step S1: Obtain the initial topology matrix reflecting the functional streamline isolation rules and collaboration timeliness indicators, as well as the 2D redline boundary matrix characterizing the boundary conditions; Step S2: Within each coordinate derivation iteration cycle, calculate the geometric overlap deviation between the planar layout data generated in the current round and the 2D redline boundary matrix, and correct the coordinate update step size of the current round in real time based on the geometric overlap deviation. Introduce the geometric overlap deviation as a boundary penalty term into the coordinate correction calculation, so that the planar layout data is constrained by an increasing gradient resistance when approaching the boundary of the 2D redline boundary matrix, and calculate the absolute geometric coordinate matrix of each medical department's functional room. Step S3: When the cyclic calculation faces the outer boundary of an irregular polygon and the number of iterations reaches the set upper limit of the number of iterations, and the comprehensive energy residual of the absolute geometric coordinate matrix of each medical department's functional room is still higher than the set convergence threshold, the global continuous gradient optimization calculation is stopped, and the heuristic polygon cavity segmentation operator is called. Based on the area parameters of each medical department's functional node in the initial topology matrix, the 2D red line boundary matrix is ​​recursively bisectioned into topology. The absolute geometric quadrant of the high-efficiency emergency department is locked first. Through the scaling limitation of the local cavity partition boundary, the absolute geometric coordinate matrix of each medical department's functional room is output.

[0006] Preferably, when processing multi-dimensional spatial data sets with vertical stacking of multiple floors, the method further includes the following steps: Step S4, using the vertical traffic flow analysis module to read the vertical traffic core location data between different levels, multiplying the absolute geometric coordinate matrix of each medical department functional room on a single floor by a 3D affine transformation operator, performing topological projection alignment along the vertical axis, using the cleanliness gradient and pedestrian flow overlap probability between adjacent floors as penalty terms to participate in cross-floor data calibration in real time, and outputting the cross-floor calibrated spatial data set.

[0007] Preferably, the real-time reverse correction of the coordinate update step size of the current round based on the geometric overlap deviation includes the following steps: Step S21, calculate the overlap area between the spatial unit bounding box outline of the current round and the 2D red line boundary matrix, and determine the geometric displacement deviation of the current round based on the overlap area; Step S22, dynamically reduce the original update step size using the geometric displacement deviation to obtain the corrected coordinate update step size of the current round.

[0008] Preferably, the heuristic polygonal cell segmentation operator is invoked to recursively perform bisection topological segmentation of the 2D redline boundary matrix based on the area parameters of each medical department functional node in the initial topology matrix, including the following steps: Step S31, retrieve the high-efficiency emergency departments in the initial topology matrix and read the target planning area corresponding to the high-efficiency emergency departments; Step S32, perform orthogonal component segmentation of the 2D redline boundary matrix along the discrete boundary normal vector of the 2D redline boundary matrix to divide it into independent absolute geometric quadrants, and assign the high-efficiency emergency departments to the absolute geometric quadrants that satisfy the target planning area.

[0009] Preferably, the geometric overlap deviation is introduced as a boundary penalty term into the coordinate correction calculation, so that the planar layout data is limited by increasing gradient resistance when approaching the boundary of the 2D redline boundary matrix. This includes the following steps: Step S23, converting the geometric overlap deviation into an inverse correction vector for coordinate adjustment; Step S24, superimposing the inverse correction vector into the global connectivity convergence gradient of the current round, changing the direction of the evolution of the planar layout data, and limiting the planar layout data from exceeding the boundary interval of the 2D redline boundary matrix.

[0010] Preferably, the initial topology matrix includes medical cleanliness classification parameters. In the calculation process derived from the initial topology matrix, the method further includes the following steps: Step S11, identifying the first functional node with clean streamline attributes and the second functional node with contaminated streamline attributes in the initial topology matrix; Step S12, calculating the spatial distance between the first functional node and the second functional node in the absolute geometric coordinate matrix of each medical department functional room, and limiting the spatial distance to be greater than the set safety protection threshold.

[0011] Preferably, the initial topology matrix contains equipment load parameters. In the calculation process derived from the initial topology matrix, the method further includes the following steps: Step S13, extracting heavy equipment nodes in the initial topology matrix whose load values ​​exceed the set weight load standard; Step S14, constraining the coordinate values ​​of the heavy equipment nodes in the absolute geometric coordinate matrix of each medical department functional room to the set bottom base coordinate space.

[0012] Preferably, when the 2D redline boundary matrix has non-right-angled sharp edges, the method further includes the following steps: Step S15, extract the vertices of the boundary polygons of the 2D redline boundary matrix and identify the coordinates of sharp vertices with included angles less than 90°; Step S16, in the derivation stage of the absolute geometric coordinate matrix of each medical department's functional room, set the spatial area around the coordinates of the sharp vertices as a low-priority arrangement quadrant and restrict public department nodes from falling into the low-priority arrangement quadrant.

[0013] Preferably, the initial topology matrix includes emergency response time-related weights. In the calculation process derived from the initial topology matrix, the method further includes the following steps: Step S17, selecting emergency center nodes and operating room nodes in the initial topology matrix whose emergency response time-related weights are greater than a set time threshold; Step S18, in the absolute geometric coordinate matrix of each medical department's functional room, constraining the emergency center nodes and operating room nodes to be on the same geometric floor or to share the same set of vertical kernel projection coordinates. After outputting the absolute geometric coordinate matrix of each medical department's functional room, the method further includes the following steps: Step S5, establishing a 3D solid model for each functional room based on the absolute geometric coordinate matrix of each medical department's functional room; Step S6, performing boundary collision detection on the 3D solid model, and after confirming that the constraints of no flow line collision and absolute boundary non-intersection are met, outputting the final plan layout planning file.

[0014] A system for generating hospital building planning floor plans, comprising a method for generating hospital building planning floor plans, including: The data acquisition module is used to acquire the initial topology matrix reflecting the functional flow line isolation rules and collaboration timeliness indicators, as well as the 2D redline boundary matrix representing the boundary conditions. The step size dynamic correction module, connected to the data acquisition module, is used to calculate the geometric overlap deviation between the planar layout data generated in the current round and the 2D redline boundary matrix in each coordinate derivation iteration cycle. Based on the geometric overlap deviation, the coordinate update step size of the current round is corrected in real time. The geometric overlap deviation is introduced into the coordinate correction calculation as a boundary penalty term, so that the planar layout data is restricted by an increasing gradient resistance when approaching the boundary of the 2D redline boundary matrix. The absolute geometric coordinate matrix of each medical department's functional room is calculated. The local cell segmentation module, connected to the step size dynamic correction module, is used to stop global continuous gradient optimization calculation when the cyclic calculation faces the external boundary of an irregular polygon, the number of iterations reaches the set upper limit, and the comprehensive energy residual of the absolute geometric coordinate matrix of each medical department's functional room is still higher than the set convergence threshold. It then calls the heuristic polygon cell segmentation operator to recursively perform bisection topological segmentation on the 2D redline boundary matrix based on the area parameters of each medical department's functional node in the initial topology matrix. It prioritizes locking the absolute geometric quadrant of high-efficiency emergency departments and outputs the absolute geometric coordinate matrix of each medical department's functional room through the scaling limitation of the local cell partition boundary.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. By separately handling the collaborative proximity relationship and circulation isolation relationship between functional rooms, the interference between constraints of different natures in the same calculation process is avoided, enabling the processor to more clearly identify the impact of various constraints on the floor plan layout; in cases with many departmental nodes and complex constraint relationships, repeated oscillations, convergence stagnation and invalid calculations in coordinate iteration can be reduced, making the transformation process of multidimensional spatial data more stable.

[0016] 2. Based on the geometric overlap deviation between the planar layout data and the 2D red line boundary matrix, the coordinate update step size is corrected in real time, and the geometric overlap deviation is used as a boundary penalty term in the coordinate calculation. This makes the functional rooms more restricted as they get closer to the boundary. As a result, the boundary overshoot and spatial overlap can be corrected in time during the derivation of the absolute geometric coordinate matrix, reducing the intersection of functional room outlines and flow line conflicts. The output results can meet the boundary constraints without relying on post-clipping.

[0017] 3. When global continuous gradient optimization calculations are difficult to converge under the condition of irregular polygonal external boundaries, the 2D red line boundary matrix can be recursively bisected topologically by using a heuristic polygonal cell segmentation operator. This can transform the complex global solution problem into layout calculations within multiple local regions, and prioritize the spatial location of high-efficiency emergency departments. This approach provides a clear subsequent solution path for convergence stagnation states, and can reduce the risk of computational interruption or layout generation failure caused by irregular boundaries. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the derivation logic of the absolute geometric coordinate matrix of this invention. Figure 2 This is a schematic diagram of the collaborative architecture of the functional planning and calculation components of this invention.

[0019] The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0020] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0021] A method for generating hospital building planning floor plans includes the following steps: Step S1: Obtain the initial topology matrix reflecting the functional streamline isolation rules and collaboration timeliness indicators, as well as the 2D redline boundary matrix characterizing the boundary conditions; Step S2: Within each coordinate derivation iteration cycle, calculate the geometric overlap deviation between the planar layout data generated in the current round and the 2D redline boundary matrix, and correct the coordinate update step size of the current round in real time based on the geometric overlap deviation. Introduce the geometric overlap deviation as a boundary penalty term into the coordinate correction calculation, so that the planar layout data is constrained by an increasing gradient resistance when approaching the boundary of the 2D redline boundary matrix, and calculate the absolute geometric coordinate matrix of each medical department's functional room. Step S3: When the cyclic calculation faces the outer boundary of an irregular polygon and the number of iterations reaches the set upper limit of the number of iterations, and the comprehensive energy residual of the absolute geometric coordinate matrix of each medical department's functional room is still higher than the set convergence threshold, the global continuous gradient optimization calculation is stopped, and the heuristic polygon cavity segmentation operator is called. Based on the area parameters of each medical department's functional node in the initial topology matrix, the 2D red line boundary matrix is ​​recursively bisectioned into topology. The absolute geometric quadrant of the high-efficiency emergency department is locked first. Through the scaling limitation of the local cavity partition boundary, the absolute geometric coordinate matrix of each medical department's functional room is output.

[0022] Preferably, when processing multi-dimensional spatial data sets with vertical stacking of multiple floors, the method further includes the following steps: Step S4, using the vertical traffic flow analysis module to read the vertical traffic core location data between different levels, multiplying the absolute geometric coordinate matrix of each medical department functional room on a single floor by a 3D affine transformation operator, performing topological projection alignment along the vertical axis, using the cleanliness gradient and pedestrian flow overlap probability between adjacent floors as penalty terms to participate in cross-floor data calibration in real time, and outputting the cross-floor calibrated spatial data set.

[0023] Preferably, the real-time reverse correction of the coordinate update step size of the current round based on the geometric overlap deviation includes the following steps: Step S21, calculate the overlap area between the spatial unit bounding box outline of the current round and the 2D red line boundary matrix, and determine the geometric displacement deviation of the current round based on the overlap area; Step S22, dynamically reduce the original update step size using the geometric displacement deviation to obtain the corrected coordinate update step size of the current round.

[0024] Preferably, the heuristic polygonal cell segmentation operator is invoked to recursively perform bisection topological segmentation of the 2D redline boundary matrix based on the area parameters of each medical department functional node in the initial topology matrix, including the following steps: Step S31, retrieve the high-efficiency emergency departments in the initial topology matrix and read the target planning area corresponding to the high-efficiency emergency departments; Step S32, perform orthogonal component segmentation of the 2D redline boundary matrix along the discrete boundary normal vector of the 2D redline boundary matrix to divide it into independent absolute geometric quadrants, and assign the high-efficiency emergency departments to the absolute geometric quadrants that satisfy the target planning area.

[0025] Preferably, the geometric overlap deviation is introduced as a boundary penalty term into the coordinate correction calculation, so that the planar layout data is limited by increasing gradient resistance when approaching the boundary of the 2D redline boundary matrix. This includes the following steps: Step S23, converting the geometric overlap deviation into an inverse correction vector for coordinate adjustment; Step S24, superimposing the inverse correction vector into the global connectivity convergence gradient of the current round, changing the direction of the evolution of the planar layout data, and limiting the planar layout data from exceeding the boundary interval of the 2D redline boundary matrix.

[0026] Preferably, the initial topology matrix includes medical cleanliness classification parameters. In the calculation process derived from the initial topology matrix, the method further includes the following steps: Step S11, identifying the first functional node with clean streamline attributes and the second functional node with contaminated streamline attributes in the initial topology matrix; Step S12, calculating the spatial distance between the first functional node and the second functional node in the absolute geometric coordinate matrix of each medical department functional room, and limiting the spatial distance to be greater than the set safety protection threshold.

[0027] Preferably, the initial topology matrix contains equipment load parameters. In the calculation process derived from the initial topology matrix, the method further includes the following steps: Step S13, extracting heavy equipment nodes in the initial topology matrix whose load values ​​exceed the set weight load standard; Step S14, constraining the coordinate values ​​of the heavy equipment nodes in the absolute geometric coordinate matrix of each medical department functional room to the set bottom base coordinate space.

[0028] Preferably, when the 2D redline boundary matrix has non-right-angled sharp edges, the method further includes the following steps: Step S15, extract the vertices of the boundary polygons of the 2D redline boundary matrix and identify the coordinates of sharp vertices with included angles less than 90°; Step S16, in the derivation stage of the absolute geometric coordinate matrix of each medical department's functional room, set the spatial area around the coordinates of the sharp vertices as a low-priority arrangement quadrant and restrict public department nodes from falling into the low-priority arrangement quadrant.

[0029] Preferably, the initial topology matrix includes emergency response time-related weights. In the calculation process derived from the initial topology matrix, the method further includes the following steps: Step S17, selecting emergency center nodes and operating room nodes in the initial topology matrix whose emergency response time-related weights are greater than a set time threshold; Step S18, in the absolute geometric coordinate matrix of each medical department's functional room, constraining the emergency center nodes and operating room nodes to be on the same geometric floor or to share the same set of vertical kernel projection coordinates. After outputting the absolute geometric coordinate matrix of each medical department's functional room, the method further includes the following steps: Step S5, establishing a 3D solid model for each functional room based on the absolute geometric coordinate matrix of each medical department's functional room; Step S6, performing boundary collision detection on the 3D solid model, and after confirming that the constraints of no flow line collision and absolute boundary non-intersection are met, outputting the final plan layout planning file.

[0030] A system for generating hospital building plans includes: The data acquisition module is used to acquire the initial topology matrix reflecting the functional flow line isolation rules and collaboration timeliness indicators, as well as the 2D redline boundary matrix representing the boundary conditions. The step size dynamic correction module, connected to the data acquisition module, is used to calculate the geometric overlap deviation between the planar layout data generated in the current round and the 2D redline boundary matrix in each coordinate derivation iteration cycle. Based on the geometric overlap deviation, the coordinate update step size of the current round is corrected in real time. The geometric overlap deviation is introduced into the coordinate correction calculation as a boundary penalty term, so that the planar layout data is restricted by an increasing gradient resistance when approaching the boundary of the 2D redline boundary matrix. The absolute geometric coordinate matrix of each medical department's functional room is calculated. The local cell segmentation module, connected to the step size dynamic correction module, is used to stop global continuous gradient optimization calculation when the cyclic calculation faces the external boundary of an irregular polygon, the number of iterations reaches the set upper limit, and the comprehensive energy residual of the absolute geometric coordinate matrix of each medical department's functional room is still higher than the set convergence threshold. It then calls the heuristic polygon cell segmentation operator to recursively perform bisection topological segmentation on the 2D redline boundary matrix based on the area parameters of each medical department's functional node in the initial topology matrix. It prioritizes locking the absolute geometric quadrant of high-efficiency emergency departments and outputs the absolute geometric coordinate matrix of each medical department's functional room through the scaling limitation of the local cell partition boundary.

[0031] Example 1: In the case of hospital building planning plan data processing with more than 60 department nodes and more than 200 rigid isolation boundary constraints, the central processing unit in the electronic digital data processing equipment reads the spatial constraint parameters and department association table of the target hospital planning area from the cache, writes the red line boundary vertex coordinates and their connection order into the 2D red line boundary matrix representing the boundary conditions, and extracts the legal area threshold, cleanliness level identifier, frequency of personnel flow and business collaboration, collaboration timeliness index, equipment load parameters, and emergency response timeliness association weight of each department unit. Based on the correspondence between department nodes, an initial topology matrix is ​​generated. The rows and columns of this matrix correspond to the functional nodes of each medical department, and its dimension is determined by the dimension of the discrete department topological feature vector. There is a high-frequency collaborative need between the operating room and the emergency department, forming a strong proximity constraint on the collaborative flow data. Simultaneously, the sterile barrier requirements between the two departments form a strong isolation constraint on the infection control flow. These mutually constraining features are prone to interference in a single isomorphic gradient descent control flow, causing the state transition matrix to get stuck in a local solution during iterative solving. This results in wasted computational resources, abnormal convergence of the solution matrix, and misalignment of the output planar layout coordinate matrix due to departmental physical boundary intersections or vertical traffic core cross-layer interfaces. During system operation, the data acquisition module is connected to the central processing unit and graphics processing unit via an internal data bus. A high-speed cache is used to load the initial topology matrix and the 2D redline boundary matrix; the step size dynamic correction module is mounted on the parallel computing core of the graphics processing unit to calculate the geometric overlap deviation between the bounding box contour of the spatial unit and the 2D redline boundary matrix and correct the coordinate update step size; the local cell segmentation module and the vertical traffic flow analysis module are scheduled by the central processing unit to perform local cell segmentation and multi-story cross-layer calibration when convergence anomalies occur, respectively; the plane generation process includes five consecutive stages: initial topology matrix loading and decoupling, geometric overlap deviation calculation and step size dynamic correction in iterative derivation, heuristic polygon cell segmentation when convergence anomalies occur, cross-layer topology projection alignment, and boundary collision detection output of the 3D solid model.

[0032] During the initial topology matrix calculation, the graphics processing unit invokes a dual-channel asynchronous decoupling operator to send the associated data in the matrix into two independent data channels. The first channel reads the collaboration timeliness indicators and daily business collaboration frequency between the nodes of each department, normalizes the business collaboration frequency to a dimensionless connectivity value between 0 and 1, and fills it into a symmetric two-dimensional matrix according to the node correspondence to generate a collaboration flow line connectivity matrix. Its element values ​​are used to constrain the centroid distance between department nodes with collaborative relationships; the second channel reads the cleanliness level identifier of each department node, identifies the first functional node with clean streamline attributes and the second functional node with contaminated streamline attributes, and calculates the infection control risk scalar difference according to the numerical difference of cleanliness level from high to low; when the cleanliness of the source node is lower than that of the target node and there is a personnel flow between the two, the scalar difference is multiplied by the direction identifier coefficient to generate an asymmetric penetration penalty weight and written into the corresponding section of the three-dimensional tensor to form the infection control time-varying flow direction interference tensor. Its directed blocking weight is used to represent the asymmetric penetration penalty in the streamline channel; the cooperation relationship and the isolation relationship are processed through two data channels respectively, so that the geometric proximity constraint and the sensing and control isolation constraint are kept in their respective data structures.

[0033] Within each coordinate derivation iteration cycle, the central processing unit (CPU) reads the planar layout data generated in the current round, places the bounding box outlines of the spatial units of each medical department's functional nodes and the 2D redline boundary matrix in the same coordinate system, calculates the overlap area between the bounding box outlines and the closed areas defined by the 2D redline boundary matrix, and determines the geometric displacement deviation for the current round based on the difference between the bounding box area and the overlap area. When the bounding box is within the redline boundary, the overlap area is consistent with the bounding box area; when the bounding box moves outward from the redline boundary, the area difference increases with the degree of boundary violation; the CPU then determines the geometric displacement deviation based on the difference between the bounding box area and the overlap area. The displacement deviation is dynamically reduced to decrease the original update step size, resulting in the corrected coordinate update step size for the current round. Simultaneously, the geometric overlap deviation is converted into an inverse correction vector for coordinate adjustment, following the direction from the centroid of the bounding box to the interior of the 2D redline boundary matrix. This vector is then added as a boundary penalty term to the global connectivity convergence gradient of the current round. As the geometric displacement deviation increases, the allowable coordinate update step size decreases accordingly, and the restriction of the inverse correction vector on the coordinate evolution direction is simultaneously strengthened. This causes the planar layout data to experience increasing gradient resistance as it approaches the boundary of the 2D redline boundary matrix, and restricts it from exceeding the redline boundary interval.

[0034] When handling the flow isolation constraints between departmental nodes, the central processing unit also periodically reads the virtual geometric bounding box array of each departmental node in the current iteration and calculates the three-dimensional Euclidean distance between any two bounding boxes in the geometric data space. When the spatial distance between the first functional node and the second functional node is less than the set safety protection threshold, the corresponding two departmental nodes are marked as spatial interference sources, and the sensor-controlled time-varying flow direction interference tensor is retrieved. The central processing unit (CPU) calculates the product of the reciprocal of this one-way penalty weight and the current iteration step size to obtain the local topological damping factor. The dimensionless coefficient is used to reduce the coordinate update components along the direction of the centroid connection between the two department nodes. At the same time, the asymmetric penetration penalty weight is converted into a geometric repulsion vector with the opposite direction and superimposed on the global connectivity convergence gradient, so that the motion direction of the department node deviates from the spatial interference source. Thus, while limiting the spatial distance between the first functional node and the second functional node to less than the safety protection threshold, the space for the node to continue iterating along the non-interference direction is reserved, and the coordinate update is prevented from stopping due to simply reducing the step size.

[0035] After each round of coordinate updates, the central processing unit (CPU) summarizes the global connectivity target value, boundary penalty term, and node isolation penalty term, and uses the absolute value of the difference between the summative values ​​of two adjacent rounds as the comprehensive energy residual for the current round. When the outer red line is an irregular polygon, the number of iterations reaches the preset maximum number of iterations of 1500, and the comprehensive energy residual is still higher than the set convergence threshold, the CPU stops the global continuous gradient optimization calculation and calls the heuristic polygon cell segmentation operator to perform recursive bisection topology segmentation. During segmentation, the irregular boundary represented by the 2D red line boundary matrix is ​​first discretized into a non-uniform grid, the normal vector of each discrete boundary segment is extracted, and two local cells are divided along the selected discrete boundary normal vector and its orthogonal components. Then, the energy of each medical department in the initial topology matrix is ​​read. The area parameters of each node are used to calculate the current area of ​​each local cavity. Cavities that do not meet the target planned area are recursively bisected in the same way. For local cavities that need boundary adjustment, the local normal vector of the outer boundary is used as the boundary pressure direction. The vertices of the cavity edge are scaled along this direction, and the cavity area is recalculated after each scaling until it meets the target planned area of ​​the assigned functional node or enters the next bisection. During the segmentation process, high-time-efficiency emergency departments are prioritized. Emergency center nodes and operating room nodes with high emergency timeliness association weights are assigned to the absolute geometric quadrants that meet the area requirements. The geometry of the edge auxiliary departments is adjusted by scaling the partition boundaries of the local cavities. Finally, the absolute geometric coordinate matrix of each medical department's functional room is output. The elements in this matrix represent the absolute spatial coordinates of the vertices of each functional room in a two-dimensional or three-dimensional spatial coordinate system.

[0036] When processing multi-dimensional spatial data sets with multiple floors vertically stacked, the vertical traffic flow analysis module reads the location data of vertical traffic cores between different floors through the data bus, constructs a vertical correlation feature tensor, and converts the absolute geometric coordinate matrix of a single floor into a single-layer absolute geometric coordinate matrix. Multiplying by a 3D affine transformation operator, the translation and rotation projection alignment between upper and lower floors are completed along the vertical axis; the equipment load parameters of each functional node are recorded in the initial topology matrix, and the central processing unit compares the equipment load values ​​with the set weight load standard and extracts the heavy equipment nodes whose load values ​​exceed the standard; the bottom base coordinate space is set as the chassis envelope area in the three-dimensional spatial coordinate system where the vertical height axis coordinate is equal to 0 and located within 0.5m above the foundation load-bearing structure wall; in the coordinate derivation iteration, a hard mask matrix constraint is applied to the vertical axis coordinate of the heavy equipment node, locking its vertical axis coordinate to 0 and restricting its horizontal plane coordinate to fall directly above the structural beam and column mesh, so that the heavy equipment load is transferred to the building foundation.

[0037] For emergency center nodes and operating room nodes in the initial topology matrix whose emergency response time correlation weights are greater than the set time threshold, the central processing unit (CPU) constrains them to be on the same geometric floor or to share the same set of vertical core projection coordinates during cross-floor calibration. The cleanliness gradient between adjacent floors is determined based on the cleanliness level difference of the corresponding functional nodes, and the probability of pedestrian overlap is determined based on the proportion of pedestrian flow projected to the same vertical traffic core location within the same time window to the corresponding total pedestrian flow. Both are used as penalty terms in cross-floor data calibration; the 3D affine transformation operator only operates on the absolute geometric coordinate matrix. In the three-dimensional spatial coordinates, the cleanliness gradient and the probability of overlap between pedestrian flows are not directly multiplied in the matrix. Instead, they are converted into a dimensionless penalty scalar value through a scalar calibration function. This penalty scalar value is then multiplied by the coordinate deviation unit vector to obtain the cross-layer position calibration gradient vector. The central processing unit (CPU) adds this gradient vector to the damping correction term of the three-dimensional coordinate iterative update equation and outputs the cross-layer calibrated spatial data set. After obtaining the absolute geometric coordinate matrix... Subsequently, the electronic digital data processing equipment establishes 3D solid models based on the vertex coordinates of each functional room recorded in the matrix, and checks the intersection status between each 3D solid model and between the 3D solid model and the red line boundary through the boundary collision detection loop; after confirming that the flow lines between functional rooms do not collide, the boundaries of each functional room do not intersect and do not exceed the area defined by the 2D red line boundary matrix, it outputs a complete plan layout planning document.

[0038] Example 2: When establishing the spatial planning solution test platform, a standard general medical spatial planning simulation dataset was used as input. This dataset contains 60 departmental functional nodes and 200 planar rigid isolation boundary constraints. To simulate the acquisition delay jitter and input frequency fluctuations caused by multi-source data access in an industrial deployment environment, the initial topology matrix... The data input bus is superimposed with an input delay jitter source with a variance of 15ms and frequency interference noise with a signal-to-noise ratio of 20dB. Before the test begins, the upper limit of the maximum number of iterations is determined based on the number of vertices of the outer red-line polygon and the spatial concurrent connectivity of the department nodes. As the number of vertices of the outer red-line polygon increases and the topological connection density of the department nodes increases, the non-convex constraints in the solution space also increase. To avoid prematurely ending the calculation before the comprehensive energy residual of the coordinate matrix drops to the convergence threshold, the upper limit of the maximum number of iterations is adjusted accordingly to the upper limit of the control interval. This working condition includes 60 department nodes, and the red-line boundary corresponding to the 2D red-line boundary matrix has 12 vertices. Based on this, the system sets the upper limit of the maximum number of iterations to 1500 times as the initial control benchmark of the test platform.

[0039] After the data processing device starts the calculation cycle, the graphics processing unit reads the initial topology matrix. The system invokes a dual-channel asynchronous decoupling operator to send collaborative relationship data and functional flow isolation rules into two independent data channels. In the control group without this scheme, strong spatial proximity constraints and strong isolation constraints interfere with each other in a single isomorphic gradient descent control flow. After adding 20dB of frequency interference noise, the equivalent coupling eigenvalue of the state transition matrix reaches 0.85. As iterations proceed, the multi-objective nonlinear solution gradually stops near a local minimum, continuously consuming computational resources. The energy residual in the solution space decreases by less than 0.2% every 100 iterations after 500 iterations, failing to continue converging towards the set convergence threshold. In the experimental group using this scheme, the first channel performs a moving average processing on the business collaboration frequency based on the statistical variance of the input delay jitter source. After eliminating input jitter, the average business collaboration frequency is 45.2. The business collaboration frequencies between the nodes of each department are then normalized and written into the collaborative flow connectivity matrix. The second channel reads the cleanliness level identifier, calculates the infection control risk scalar difference between the first functional node with clean streamline attributes and the second functional node with contaminated streamline attributes, and generates an infection control time-varying flow direction interference tensor containing asymmetric penetration penalty weights. .

[0040] Within each coordinate derivation iteration cycle, the central processing unit (CPU) first places the planar layout data generated in the current round and the 2D redline boundary matrix in the same coordinate system. It calculates the overlap area between the bounding box outline of each spatial unit and the region defined by the 2D redline boundary matrix, and determines the geometric displacement deviation based on the difference between the bounding box area and the overlap area. Subsequently, it reduces the original update step size according to the geometric displacement deviation and converts the geometric overlap deviation into an inverse correction vector pointing inwards towards the redline boundary. This vector is then superimposed on the global connectivity convergence gradient of the current round, causing the planar layout data to experience increasing gradient resistance as it approaches the boundary of the 2D redline boundary matrix. During the flow line isolation verification between departmental nodes, the CPU detects that the three-dimensional Euclidean distance between the operating room area and the emergency room area in the three-dimensional data space continuously decreases from the initial 5.0m to 0.8m, which is below the preset 1.2m safety protection threshold. Therefore, it retrieves the time-varying flow direction interference tensor for infection control. The one-way penalty weight for the corresponding node pair is read as 12.8. The reciprocal of this weight is multiplied by the current iteration step size of 0.05 to obtain the local topological damping factor. The value is 0.0039; this value is used to replace the coordinate update step size along the line connecting the centroids of the two department nodes in the current round, reducing the spatial advancement speed in the conflict direction by 92.2%.

[0041] While reducing the update step size of the collision direction coordinates, the central processing unit will direct the sensing-controlled time-varying flow to the interference tensor. The asymmetric penetration penalty weights are converted into geometric repulsion vectors with opposite directions and superimposed on the global connectivity convergence gradient. The coordinate update direction is thus shifted tangentially and outwards away from the spatial interference source. After entering the safety protection threshold range, nodes no longer continue to intrude along the original direction, nor do they stop due to the shrinking coordinate update step size; instead, they leave the overlapping region along the non-interference direction. To verify the operational status under different node sizes and boundary pressures, the test platform uses spatial planning data with 20, 40, and 60 nodes for comparison, and performs segmented tests on the safety protection threshold. When the safety protection threshold is set to 0.5m, the spatial unit approaches the volume boundary of the usable area under the redline constraint. After 850 iterations in the experimental group, the comprehensive energy residual decreases to 0.012, and the output absolute geometric coordinate matrix... Maintaining a 100% boundary validity rate, spatial overlap interference is zero; under the same operating conditions, the boundary overflow distortion rate of the control group is 65.5%.

[0042] When the safety protection threshold was set to 1.5m, the experimental group reached convergence after 620 iterations, and the comprehensive energy residual decreased to 0.005. Increasing the safety protection threshold to 2.5m further tightened the spatial connectivity constraints. Further increasing the threshold to 3.0m significantly reduced the coordinate update step size due to increased damping, and the rate of change of the comprehensive energy residual tended to level off, indicating that the usable building outline space was approaching geometric saturation. In tests where the outer red line was an irregular polygon, after the main loop ran for 1500 iterations, the comprehensive energy residual was still higher than the set convergence threshold. The system stopped global continuous gradient optimization calculations and called the heuristic polygon cell segmentation operator. This operator is based on the initial topology matrix... The area parameters of functional nodes in each medical department were used to recursively divide the 2D redline boundary matrix into two topological segments. The partition boundaries were adjusted according to the correspondence between the local cavity area and the target planned area, prioritizing the allocation of surgical department nodes and emergency center nodes to the absolute geometric quadrants that satisfy the target planned area. After recursive bisection topological segmentation and scaling constraints on the partition boundaries of local cavities, the experimental group achieved a 100% legality rate in generating the planar layout planning file. The control group, which did not have a corresponding local cavity segmentation process, had a legality rate of 34.5% and exhibited misalignment of the cross-layer vertical traffic core interface. The test data indicate that, while maintaining consistent input node size, boundary constraints, and interference conditions, using geometric overlap deviation as a boundary penalty term to correct the coordinate update step size, and switching to recursive bisection topological segmentation when continuous gradient optimization fails to converge, can improve the absolute geometric coordinate matrix. Continue to converge toward a coordinate state that satisfies the redline boundary and streamline isolation constraints.

[0043] Example 3: When processing planar data where there is at least one narrow included angle within 15 degrees for a non-convex polygonal redline boundary, and the sum of the statutory area thresholds of each medical department's functional nodes reaches the geometric volume limit of the outer redline boundary, the central processing unit reads the 2D redline boundary matrix and the initial topology matrix representing the boundary conditions. During the mapping from graph topology to a two-dimensional plane, continuous gradient optimization calculations are prone to step-like numerical oscillations at narrow angles, preventing the overall energy residual from falling below the set convergence threshold. Furthermore, when the number of iterations reaches the maximum limit of 1500 iterations, if global continuous gradient optimization calculations are still used, the output absolute geometric coordinate matrix... Functional room boundaries may intersect or overlap; when there are sharp vertices with an angle less than 90 degrees in the 2D redline boundary matrix, the central processing unit extracts the coordinates of the corresponding sharp vertex, forms a circular region with the sharp vertex as the center and a radius of 5m, and sets the sector space where this circular region intersects with the interior of the redline boundary polygon as a low-priority arrangement quadrant; in the absolute geometric coordinate matrix In the derivation stage, a penalty amplification factor of 10 is applied to the spatial displacement gradient that falls into the low priority layout quadrant. This prioritizes the layout of insensitive auxiliary nodes such as toilets, storage rooms, or pipe shafts in this area, while restricting public department nodes such as waiting halls and lobbies, as well as main circulation nodes, from falling into this quadrant to avoid narrow angles squeezing the main functional rooms and their passage spaces.

[0044] When the main loop reaches the maximum number of iterations (1500) and the overall energy residual is still higher than the convergence threshold, the central processing unit (CPU) stops the global continuous gradient optimization calculation, calls the heuristic polygon cell segmentation operator, and performs recursive bisection topological segmentation of the 2D redline boundary matrix in three consecutive steps: First, the geometric centroid of the polygon is calculated based on the vertex coordinates in the 2D redline boundary matrix. The discrete boundary normal vector passing through this geometric centroid is selected as the initial segmentation direction, and the initial cutting axis is established using this normal vector and its orthogonal components. Second, the irregular plane is divided into two independent discrete sub-cells along the initial cutting axis, and the current geometric area of ​​each discrete sub-cell is calculated. Third, the initial topological matrix is ​​read. The area parameters of each functional node in the medical department are used to compare the statutory area threshold of a specific functional node with the actual geometric area of ​​the corresponding discrete sub-cell, and the area adjustment factor is calculated: ;in, This is a dimensionless adjustment parameter used to correct the scaling amplitude of local cell lumen edge apex. To start from the initial topology matrix The statutory area threshold for specific medical department functional nodes extracted from [the data]. The actual geometric area of ​​the discrete sub-cells divided in the current round, in square meters; the central processing unit (CPU) uses... As the basis for adjusting the coordinates of the edge vertices, the sub-cell edge vertices are moved along the discrete boundary normal vector and its orthogonal components, and recalculated after each adjustment. When the current geometric area still cannot satisfy If necessary, continue with local scaling or perform another round of bisection on the subcell until a local cell with the target planned area is obtained.

[0045] When faced with irregular boundaries or 2D redline boundaries with irregular concave angles, the heuristic polygonal cell segmentation operator first obtains the polygonal bounding rectangle of the redline boundary, sets the geometric center of the bounding rectangle as the local coordinate origin, and establishes four orthogonal basic absolute geometric quadrants along the dominant axis and perpendicular axis of the redline boundary. When the redline boundary has non-convex or concave angles, a cutting line is extended into the redline boundary along the discrete boundary normal vector of the concave angle vertex, and then the orthogonal component of the normal vector is used to subdivide the quadrant into multiple rectangular or trapezoidal sub-cells. After completing the quadrant division, the central processing unit retrieves the initial topology matrix. The target planning areas corresponding to the emergency medical center nodes and operating room nodes are processed according to the order of target planning area from largest to smallest for high-efficiency emergency departments. Priority is given to allocating the high-efficiency emergency departments with the largest target planning area to the absolute geometric quadrant closest to the main entrance that meets the area requirements. If the area of ​​the current quadrant is insufficient, the orthogonal tangent axis is translated or rotated, and an area adjustment coefficient is used. Adjust the boundaries of the subcell cavity, recalculate the adjusted actual geometric area until the quadrant can accommodate the corresponding target planned area, and then lock the absolute geometric quadrant of the high-efficiency emergency department; use the area adjustment coefficient Adjusting the local cavity boundaries in narrow angular regions transforms global continuous gradient optimization calculations into edge vertex scaling calculations within a locally non-uniform mesh. As each edge vertex moves along the local normal vector of the outer boundary in successive cycles, its displacement gradually decreases as the difference between the actual geometric area and the target planned area diminishes. This allows the operating room node and the emergency center node to converge to mutually non-intersecting absolute positions when the available space approaches the boundary limits. After recursive bipartite topology partitioning and local cavity partition boundary scaling constraints, the system outputs an absolute geometric coordinate matrix of boundary data that conforms to the functional streamline isolation rule. The compliance rate of the generated floor plan layout documents reached 100%.

[0046] Example 4: This example combines Figures 1 to 2 This paper describes a method and system for generating hospital building planning floor plans, such as... Figure 1As shown, the system first executes step S1 to acquire basic topology and boundary data. During this stage, the system acquires the initial topology matrix reflecting functional flow isolation rules and collaboration efficiency indicators, as well as the 2D redline boundary matrix characterizing boundary conditions. After importing the basic data, the data flows to step S2 for coordinate derivation and dynamic step size correction. This process calculates the geometric overlap deviation between the current round's planar layout data and the 2D redline boundary matrix and introduces the deviation as a boundary penalty term. Real-time reverse correction of the coordinate update step size ensures that the planar layout is constrained by increasing gradient resistance as it approaches the boundary. The absolute geometric coordinate matrix of each medical department's functional room is calculated. Subsequently, the algorithm enters the global continuous gradient convergence judgment node. In this stage, the system judges the cyclic calculation facing the external boundary of an irregular polygon and the iteration round... If the set number of iterations is reached and the overall energy residual is still higher than the set convergence threshold, the computation flow is directed to the normal convergence state of global optimization when the overall energy residual drops below the set convergence threshold. There is no computational deadlock or physical interference. If the result is yes, the system triggers step S3 and calls the polygon cell segmentation operator. At this time, the system forcibly stops the global continuous gradient optimization calculation and performs recursive bisection topological segmentation on the 2D red line boundary matrix according to the area parameters of the medical departments. The absolute geometric quadrant of the high-efficiency emergency departments is locked first, and local cell boundary scaling restrictions are implemented. The above two logical calculation branches eventually converge to the output port of the planar layout planning result and finally output the absolute geometric coordinate matrix of each medical department's functional room.

[0047] like Figure 2 As shown, the data acquisition module receives functional flow isolation rules, collaborative timeliness indicators, and two-dimensional redline boundary matrices as basic input sources at the system front end. The initial topology matrix injects medical cleanliness classification parameters, equipment load parameters, and emergency response timeliness correlation weights into the core solution area. The step size dynamic correction module receives geometric overlap deviations fed back from the system and converts them into boundary penalty terms to generate inverse correction coordinate step size control instructions. The local cell cavity segmentation module loads a heuristic polygon cell cavity segmentation operator to perform recursive binary topology segmentation and prioritizes resource occupancy of high-timeliness emergency departments in spatial allocation. The vertical traffic flow analysis module loads a three-dimensional affine transformation operator to perform topological projection alignment along the vertical axis and directly incorporates the extracted cleanliness gradient and pedestrian flow overlap probability into the multi-floor interference penalty calculation. The spatial data group and model building module constructs the corresponding three-dimensional solid model based on the absolute geometric coordinate matrix transmitted from the upstream and performs forced boundary collision detection. The data streams, logical constraint parameters, and structural verification results produced by the above modules are finally aggregated at the system bottom layer and output as a planar layout planning file.

[0048] Example 5: When the addition of a special medical service unit causes a non-monotonic change in the department attribute vector, and it is necessary to determine the safety protection threshold used for coordinate correction calculation, the central processing unit retrieves the historical planning topology feature map and reads the initial topology matrix before starting the iterative solution loop. The bounding box parameters of each medical department's functional nodes are calculated, and the average equivalent radius of the bounding box of all department nodes is statistically analyzed. The graphics processing unit divides the total area of ​​the closed region defined by the 2D redline boundary matrix by the total number of department functional nodes to obtain the average allocated physical area of ​​the cell cavity corresponding to each department functional node. To ensure that the safety protection threshold accurately maps different land use intensities and boundary shape constraints in physical space and geometric topology, this embodiment constructs a physical calculation model of the safety protection threshold based on the principle of site geometric similarity scaling and the dynamic infection control buffer capacity mechanism. At the mechanism level, the total area represents the macroscopic spatial envelope base of the planned site. The logic of its positive correlation with the safety protection threshold is that when the total area of ​​the planned site is large, the physical envelope size of the site is correspondingly widened, and the system can dynamically increase the safety protection threshold to expand the infection control isolation buffer zone between the clean flow line and the contaminated flow line, maximizing the protection performance of medical infection control. Conversely, when the total area of ​​the site is shrinking, the spatial density of department nodes increases significantly. The system can synchronously and adaptively shrink the safety protection threshold in a positive proportion to avoid setting too large a fixed value in a limited space. Fixed spacing can lead to algorithmic convergence deadlock and spatial layout conflicts, thus achieving an adaptive match between high-density compact layout and low-density maximization of infection control and protection. In specific calculations, the average physical area of ​​a single department's functional cavity within the planned site is measured in square meters. Taking the square root of this average physical area transforms the two-dimensional area into a one-dimensional spatial feature linear scale, which is used to quantify the physical equivalent spatial span of a single department's cavity. Simultaneously, the curvature coefficient of the polygon vertices of the discrete redline contour is derived based on the discrete geometric angle deficit and boundary compression theory, used to characterize the geometric damping and compression attenuation effect of irregular redline boundaries on spatial arrangement. Its value is a dimensionless scaling factor less than 1. Multiplying this one-dimensional feature linear scale by the dimensionless curvature coefficient essentially maps the average linear scale of the cavities at the macroscopic site level to the physical safety protection limit distance for streamline isolation between microscopic departments after boundary topological attenuation. The central processing unit multiplies the square root of the average physical area of ​​the cavity with the curvature coefficient of the polygon vertices of the discrete redline contour, and calculates the safety protection threshold according to the following formula: ;in, This is the safety protection threshold; is the curvature coefficient of the polygon vertices of the discrete redline contour, and is a dimensionless geometric constant; The total area of ​​the enclosed region defined by the 2D redline boundary matrix; The initial topology matrix The total number of departmental functional nodes is a dimensionless pure number.

[0049] Under the condition that the total area of ​​the enclosed area is 10,000 m² and the total number of functional nodes of the departments is 100, the following will be implemented: and Substituting into the above formula, we obtain an average distribution of the cell cavity physical region as 100 m², with its square root being 10 m; when the curvature coefficient of the polygon vertices of the discrete red line contour... When the value is 0.12, the safety protection threshold is calculated. The value is 1.2m; the curvature coefficients of the polygon vertices of the discrete redline contour are determined by the vertex coordinates in the 2D redline boundary matrix; the central processing unit sequentially traverses each polygon vertex and calculates the interior angle formed by two adjacent boundary lines; for interior angles less than 180 degrees, the absolute value of the deviation from the straight angle is calculated, and all absolute values ​​of deviation are summed to obtain the total angular deflection of the closed polygon; then the total angular deflection is divided by the product of the total number of vertices and the full angle to obtain the dimensionless polygon vertex curvature coefficients; when the input redline boundary is a complex polygon with 12 vertices and multiple alternating convex and concave vertices, the above calculations and calibrations yield the following results. The value is 0.12, which is used to characterize the degree to which irregular boundaries compress the available space.

[0050] The central processing unit will set a security protection threshold. The value is written to the global register as the trigger value for spatial spacing verification between the first functional node with clean streamline attributes and the second functional node with contaminated streamline attributes. In subsequent coordinate derivation iterations, the central processing unit calculates the geometric overlap deviation between the planar layout data generated in the current round and the 2D redline boundary matrix, and uses this deviation as a boundary penalty term in the coordinate update step size correction. When the spatial spacing between the first functional node and the second functional node approaches or falls below the safety protection threshold, the sensing and control time-varying flow direction interference tensor is further retrieved. The asymmetric penetration penalty weight between corresponding nodes is used to reduce the coordinate update step size in the conflict direction, and the resulting inverse correction vector is superimposed on the global connectivity convergence gradient. Thus, planar layout data experiences increasing gradient resistance as it approaches the boundary of the 2D redline boundary matrix, and functional nodes with different clean streamline properties are isolated as they approach the safety protection threshold. (Absolute geometric coordinate matrix) It gradually converges to an absolute spatial position that satisfies boundary constraints and functional streamline isolation rules.

[0051] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0052] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for generating hospital building planning floor plans, characterized in that, Includes the following steps: Step S1: Obtain the initial topology matrix reflecting the functional streamline isolation rules and collaboration timeliness indicators, as well as the 2D redline boundary matrix characterizing the boundary conditions; Step S2: Within each coordinate derivation iteration cycle, calculate the geometric overlap deviation between the planar layout data generated in the current round and the 2D redline boundary matrix, and correct the coordinate update step size of the current round in real time based on the geometric overlap deviation. Introduce the geometric overlap deviation as a boundary penalty term into the coordinate correction calculation, so that the planar layout data is constrained by an increasing gradient resistance when approaching the boundary of the 2D redline boundary matrix, and calculate the absolute geometric coordinate matrix of each medical department's functional room. Among them, the real-time inverse correction of the coordinate update step size of the current round based on the geometric overlap deviation includes the following steps: Step S21: Calculate the overlap between the spatial unit bounding box outline of the current round and the 2D redline boundary matrix. The area is used to determine the geometric displacement deviation of the current round; step S22, the geometric displacement deviation is used to dynamically reduce the original update step size to obtain the corrected coordinate update step size of the current round; wherein, the geometric overlap deviation is introduced as a boundary penalty term into the coordinate correction calculation, so that the planar layout data is limited by increasing gradient resistance when approaching the boundary of the 2D redline boundary matrix, including the following steps: step S23, the geometric overlap deviation is converted into the inverse correction vector of coordinate adjustment; step S24, the inverse correction vector is superimposed on the global connectivity convergence gradient of the current round to change the direction of the evolution of the planar layout data and limit the planar layout data from exceeding the boundary interval of the 2D redline boundary matrix; Step S3: When the iterative calculation faces the outer boundary of an irregular polygon and the number of iterations reaches the set upper limit, and the comprehensive energy residual of the absolute geometric coordinate matrix of each medical department's functional room is still higher than the set convergence threshold, the global continuous gradient optimization calculation is stopped, and the heuristic polygon cavity segmentation operator is called. Based on the area parameters of each medical department's functional node in the initial topology matrix, the 2D redline boundary matrix is ​​recursively bisectioned topologically, prioritizing the locking of the absolute geometric quadrant of high-efficiency emergency departments. Through the scaling constraint of the local cavity partition boundary, the absolute values ​​of each medical department's functional room are output. The geometric coordinate matrix; among which, calling the heuristic polygon cell segmentation operator to recursively bisection the 2D redline boundary matrix according to the area parameters of each medical department functional node in the initial topology matrix includes the following steps: Step S31, retrieve the high-efficiency emergency departments in the initial topology matrix and read the target planning area corresponding to the high-efficiency emergency departments; Step S32, perform orthogonal component segmentation on the 2D redline boundary matrix along the discrete boundary normal vector of the 2D redline boundary matrix, divide the independent absolute geometric quadrants, and assign the high-efficiency emergency departments to the absolute geometric quadrants that satisfy the target planning area.

2. The method for generating hospital building planning floor plans according to claim 1, characterized in that, When processing multi-dimensional spatial data sets with vertical stacking of multiple floors, the method also includes the following steps: Step S4, using the vertical traffic flow analysis module to read the vertical traffic core location data between different levels, multiplying the absolute geometric coordinate matrix of each medical department functional room on a single floor by a 3D affine transformation operator, performing topological projection alignment along the vertical axis, and using the cleanliness gradient and pedestrian flow overlap probability between adjacent floors as penalty terms to participate in cross-floor data calibration in real time, and outputting the cross-floor calibrated spatial data set.

3. The method for generating hospital building planning floor plans according to claim 1, characterized in that, The initial topology matrix contains medical cleanliness classification parameters. In the calculation process derived from the initial topology matrix, the method also includes the following steps: Step S11, identify the first functional node with clean streamline attributes and the second functional node with contaminated streamline attributes in the initial topology matrix; Step S12, calculate the spatial distance between the first functional node and the second functional node in the absolute geometric coordinate matrix of each medical department functional room, and limit the spatial distance to be greater than the set safety protection threshold.

4. The method for generating hospital building planning floor plans according to claim 1, characterized in that, The initial topology matrix contains equipment load parameters. In the calculation process derived from the initial topology matrix, the method also includes the following steps: Step S13, extract heavy equipment nodes in the initial topology matrix whose load values ​​exceed the set weight load standard; Step S14, constrain the coordinate values ​​of the heavy equipment nodes in the absolute geometric coordinate matrix of each medical department functional room to the set bottom base coordinate space.

5. The method for generating hospital building planning floor plans according to claim 1, characterized in that, When the 2D redline boundary matrix has non-right-angled sharp edges, the method further includes the following steps: Step S15, extract the vertices of the boundary polygons of the 2D redline boundary matrix and identify the coordinates of sharp vertices with included angles less than 90°; Step S16, in the derivation stage of the absolute geometric coordinate matrix of each medical department's functional room, set the spatial area around the coordinates of the sharp vertices as a low-priority arrangement quadrant and restrict public department nodes from falling into the low-priority arrangement quadrant.

6. The method for generating hospital building planning floor plans according to claim 1, characterized in that, The initial topology matrix contains emergency response time-related weights. In the calculation process derived from the initial topology matrix, the method further includes the following steps: Step S17, selecting emergency center nodes and operating room nodes in the initial topology matrix whose mutual emergency response time-related weights are greater than a set time threshold; Step S18, constraining emergency center nodes and operating room nodes to be on the same geometric floor or sharing the same set of vertical kernel projection coordinates in the absolute geometric coordinate matrix of each medical department's functional rooms. After outputting the absolute geometric coordinate matrix of each medical department's functional rooms, the method further includes the following steps: Step S5, establishing 3D solid models for each functional room based on the absolute geometric coordinate matrix of each medical department's functional rooms; Step S6, performing boundary collision detection on the 3D solid models, confirming that the constraints of no flow line collision and absolute boundary non-intersection are met, and then outputting the final plan layout planning file.

7. A system for generating hospital building planning floor plans, used to implement the method for generating hospital building planning floor plans as described in claim 1, characterized in that, include: The data acquisition module is used to acquire the initial topology matrix reflecting the functional flow line isolation rules and collaboration timeliness indicators, as well as the 2D redline boundary matrix representing the boundary conditions. The step size dynamic correction module, connected to the data acquisition module, is used to calculate the geometric overlap deviation between the planar layout data generated in the current round and the 2D redline boundary matrix in each coordinate derivation iteration cycle. Based on the geometric overlap deviation, the coordinate update step size of the current round is corrected in real time. The geometric overlap deviation is introduced into the coordinate correction calculation as a boundary penalty term, so that the planar layout data is restricted by an increasing gradient resistance when approaching the boundary of the 2D redline boundary matrix. The absolute geometric coordinate matrix of each medical department's functional room is calculated. The local cell segmentation module, connected to the step size dynamic correction module, is used to stop global continuous gradient optimization calculation when the cyclic calculation faces the external boundary of an irregular polygon, the number of iterations reaches the set upper limit, and the comprehensive energy residual of the absolute geometric coordinate matrix of each medical department's functional room is still higher than the set convergence threshold. It then calls the heuristic polygon cell segmentation operator to recursively perform bisection topological segmentation on the 2D redline boundary matrix based on the area parameters of each medical department's functional node in the initial topology matrix. It prioritizes locking the absolute geometric quadrant of high-efficiency emergency departments and outputs the absolute geometric coordinate matrix of each medical department's functional room through the scaling limitation of the local cell partition boundary.

Images