Layout generation optimization method and system for digital twin of space facility

Abstract

The invention relates to the technical field of digital twinning, in particular to a layout generation optimization method and system for a digital twinning body of a space facility, and the method comprises the steps: constructing a two-dimensional data matrix for each layout scheme of the space facility through employing the evaluation indexes of a preset number of channels in a discretization grid; constructing a social force model based on the two-dimensional data matrix, and determining a layout cost index; analyzing evaluation indexes about efficiency safety, energy consumption and a flexible channel in the two-dimensional data matrix to obtain a multi-dimensional evaluation vector, obtaining a dynamic weight vector by adopting a Bayesian network, obtaining a close degree score by utilizing a decision analysis method, obtaining scene adaptation robustness in combination with an element numerical value dispersion degree of the dynamic weight vector, and determining a fitness function, so as to obtain the scene adaptation robustness. And adopting a particle swarm optimization algorithm to obtain an optimal layout scheme. The invention aims to enhance the comprehensive performance of the space facility layout scheme.

Description

Technical Field

[0001] This application relates to the field of digital twin technology, specifically to a method and system for generating and optimizing the layout of a digital twin of a spatial facility. Background Technology

[0002] With the deepening of digital transformation, the planning and operation of complex spatial facilities face higher requirements. Traditional layout design relies on experience and static diagrams, which cannot effectively evaluate multiple objectives such as overall efficiency, safety, and energy consumption. However, the digital twin layout simulation optimization platform, built using modeling technology, the Internet of Things, and intelligent algorithms, integrates physical rules and real-time data, realizing automatic generation and iterative evaluation based on computers. It transforms the previous two-dimensional spatial schemes into automatically generated and evolving spatial layout schemes, greatly improving space utilization and the efficiency of human-machine collaboration.

[0003] In the design and optimization of spatial facility layout, existing technologies often use linear methods such as weighted summation to couple multiple objectives such as efficiency and safety. The weight setting relies on subjective experience, which makes it difficult to objectively reflect the nonlinear conflicts and mutual constraints between objectives. The layout evaluation weights are usually fixed and cannot be adaptively adjusted according to different scenarios such as "daily", "peak" and "emergency". This results in poor adaptability and insufficient robustness of the scheme in actual switching, reducing decision-making efficiency and practicality. Summary of the Invention

[0004] In view of the above, it is necessary to provide a method and system for optimizing the layout of digital twins of spatial facilities to solve the above problems.

[0005] The first aspect of this application provides a method for generating and optimizing the layout of a digital twin of a space facility, the method comprising: For space facilities, a two-dimensional data matrix for each layout scheme is constructed by using evaluation indicators of a preset number of channels in a discretized grid; the evaluation indicators are used to represent the layout information of the space facilities and the grid. For each layout scheme, the fluctuation characteristics of the evaluation index of the channel in the two-dimensional data matrix are analyzed. Based on the evaluation index of the two-dimensional data matrix, a social force model is constructed to analyze the orderliness of the operation process of the spatial facility layout under static conditions and the safety issues in dealing with the emergency evacuation of the crowd, and to determine the layout cost index of each layout scheme. The distribution characteristics of evaluation indicators related to efficiency and safety, energy consumption, and flexible channels in the two-dimensional data matrix of each layout scheme are analyzed to form a multi-dimensional evaluation vector for each layout scheme. A Bayesian network is used to obtain a dynamic weight vector, and further decision analysis methods are used to obtain the proximity score corresponding to each layout scheme, thereby obtaining the scene adaptation robustness of each layout scheme. Based on the scene adaptability robustness and proximity score of each layout scheme, the fitness function is determined, and the optimal layout scheme is obtained by using an optimization algorithm.

[0006] Specifically, the construction of the two-dimensional data matrix for each layout scheme is as follows: each row represents a discretized grid, and each column is an evaluation index for a preset channel. There are 5 preset channels. The evaluation index of channel 1 is used to reflect the spatial geometric information of the facilities within the grid; the evaluation index of channel 2 is used to reflect the usage frequency of the facilities within the grid; the evaluation index of channel 3 is used to reflect the per capita flow rate within the grid; the evaluation index of channel 4 is used to reflect the material transfer frequency within the grid; and the evaluation index of channel 5 is used to reflect the distance information of the grid from the entrance and exit.

[0007] The process of determining the layout cost index for each layout scheme is as follows: Based on the aforementioned fluctuation characteristics, the entropy weight method is used to obtain the comprehensive evaluation value of each evaluation indicator; Based on the output of the social force model, the evacuation risk index of each layout scheme is calculated. The result of positively fusing the negative correlation mapping of the comprehensive evaluation value of each layout scheme with the evacuation risk index is used as the layout cost index for each layout scheme.

[0008] The step of obtaining the comprehensive evaluation value of each evaluation index using the entropy weight method is as follows: All channels except channel 1 in the two-dimensional matrix corresponding to each layout scheme are used as input to the entropy weight method. The weight vector of each evaluation index is output, and the weighted sum of all evaluation indices for each grid is used to obtain the comprehensive evaluation value.

[0009] Specifically, the calculation of the evacuation risk index for each layout scheme is as follows: The social force model outputs the overall evacuation completion time and the maximum instantaneous congestion density during the evacuation process for each layout scheme. The overall evacuation completion time and the maximum instantaneous congestion density are normalized respectively, and the product between the two after normalization is calculated as the evacuation risk index for each layout scheme.

[0010] The process of obtaining the multidimensional evaluation vector for each layout scheme is as follows: For each layout scheme, the layout cost index is used as the efficiency and safety evaluation data; the sum of all data of all grids in channel 2 is used as the energy consumption evaluation data; based on the proportion and distribution of mobile equipment grids in channel 1 of each layout scheme, combined with the concentration of high material transfer frequency areas in channel 4, the equipment movement time required for layout adjustment is simulated through digital twin simulation to obtain flexibility evaluation data. After normalizing all the evaluation data for each layout scheme, a multi-dimensional evaluation vector is formed.

[0011] Specifically, obtaining the proximity score for each layout scheme is as follows: The normalized data and dynamic weight vector of each element in the multidimensional evaluation vector of each layout scheme are used as input to the TOPSOS algorithm, and the output is the closeness score of each layout scheme to the positive ideal solution.

[0012] Specifically, the scenario adaptability robustness of each layout scheme is obtained as follows: Calculate the dispersion of the stitching score for each layout scheme in all preset typical scenarios, and use the negative correlation mapping result of the dispersion as the scenario adaptation robustness of each layout scheme.

[0013] The fitness function is specifically defined by the following formula: In the formula, F represents fitness, α and β both represent preset weights, C represents the proximity score of each layout scheme, and B represents the scene adaptation robustness of each layout scheme.

[0014] Secondly, embodiments of this application also provide a layout generation and optimization system for a digital twin of a space facility, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of any of the methods described above.

[0015] This application has at least the following beneficial effects: This application addresses the problem of the difficulty in objectively coupling efficiency and safety goals by using the entropy weight method and the social force model to reflect the multiplicative amplification effect between efficiency orderliness and evacuation risk. This solves the problem that the traditional weighted summation method is highly subjective and cannot effectively express the nonlinear conflict between goals.

[0016] Secondly, to address the issues of fixed weight settings and poor adaptability of solutions under dynamic scenario switching, Bayesian networks and TOPSIS are used to reflect the overall performance stability of the solution when weights change, eliminating the limitations of subjective preferences for manually preset weights and static assumptions about the scenario.

[0017] Finally, in the optimization algorithm, the optimal and suboptimal solutions in the layout solution space are considered simultaneously during the optimization process. It can automatically perform dynamic trade-offs and compare and select layout solutions, thereby selecting a set of Pareto non-dominated solutions that are close to the optimal value of the current scenario and have a certain degree of cross-scenario adaptability. This enhances the overall performance of the layout scheme and the stability of the system over a long period of time, reduces computational redundancy and decision complexity, and provides a guarantee for the intelligent planning and dynamic optimization of facilities throughout their entire life cycle. Attached Figure Description

[0018] Figure 1 A flowchart illustrating the steps of a layout generation and optimization method for a digital twin of a spatial facility, provided in one embodiment of this application; Figure 2 A flowchart illustrating the process of obtaining scenario adaptation robustness in one embodiment of this application. Detailed Implementation

[0019] In the description of the embodiments in this application, the words "exemplary," "or," and "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design scheme described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of the words "exemplary," "or," and "for example" is intended to present the relevant concepts in a specific manner.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in this application's specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the application.

[0021] It should also be noted that the terms "first" and "second" in this application and its accompanying drawings are used to distinguish similar objects, rather than to describe a specific order or sequence. The methods disclosed in the embodiments of this application or the methods shown in the flowcharts include one or more steps for implementing the method. Without departing from the scope of protection of this application, the execution order of multiple steps can be interchanged, and some steps can also be deleted.

[0022] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0023] The following description, in conjunction with the accompanying drawings, details a specific scheme for a layout generation and optimization method and system for a digital twin of a space facility provided in this application.

[0024] Please see Figure 1 The diagram illustrates a flowchart of a layout generation and optimization method for a digital twin of a space facility according to an embodiment of this application. The method includes the following steps: Step 1: For spatial facilities, construct a two-dimensional data matrix for each layout scheme by using evaluation indicators of a preset number of channels in the discretized grid.

[0025] Because the layout design of spatial facilities involves a trade-off between efficiency, safety, and energy consumption, an overly dense arrangement may hinder the smooth operation of tasks, while an overly sparse layout may cause delays in emergency rescue and create potential risks. Therefore, for each spatial facility layout scheme, a two-dimensional data matrix of size W×C is constructed to represent its layout, where W represents the number of discretized grid units, and C represents the number of feature channels in the discretized grid units. The specific feature channels are shown below: Channel 1: Numerical encoding represents the state of the grid; Channel 2: Average number of times devices are used per hour in each grid (0 is recorded if there are no devices); Channel 3: Average traffic per person per grid (number of people / hour); Channel 4: Frequency of material transfer in this grid; Channel 5: The shortest Euclidean distance from this grid to the entrance / exit; The data collected from channels 2, 3, 4, and 5 are linearly scaled to the range of 0 to 1 using the Min-Max normalization method to eliminate the influence of dimensions.

[0026] Specifically, based on the spatial geometric information of the facility (CAD drawings or 3D point clouds, etc.), the length and width directions are used as the basis for grid division, and each grid is used as a separate state description space. Simultaneously, the facility asset database is retrieved to obtain the location and layout type of equipment on each grid and mapped to the corresponding preset state code. Specifically, a multi-hot encoding method is used to characterize the grid state, using a six-bit binary vector to correspond to six attributes: idle area, fixed equipment, movable equipment, workbench, entrance / exit, and storage area. Each bit has a value of 1 indicating the presence of the attribute and 0 indicating its absence. For example, an area that simultaneously serves as a main passage and stores movable equipment can be encoded as [0,0,1,0,1,0], and this six-bit binary vector is used as the data for channel 1. The average number of times equipment is used within the grid over various time periods is extracted from the equipment operation logs and used as the data for channel 2. The average number of people passing through each grid per unit time is obtained based on historical monitoring data and used as the data for channel 3. The material transfer frequency within each grid is obtained through simulated material transfer process statistics and used as the data for channel 4. Based on the state of the grid in channel 1, the shortest Euclidean distance between the grid location coordinates and the entrance / exit coordinates is used as the data for channel 5.

[0027] Thus, a two-dimensional data matrix is ​​obtained during the space facility optimization process. The rows represent the discretized grid cells, with a total of W rows, and the columns represent the feature channels of each grid, with a total of C columns. In this embodiment, C=5. In the two-dimensional matrix, the element in the i-th row and j-th column represents the data value of the i-th grid at the j-th channel.

[0028] Step 2: For each layout scheme, analyze the fluctuation characteristics of the evaluation indicators of the channels in the two-dimensional data matrix, and construct a social force model based on the evaluation indicators in the two-dimensional data matrix to analyze the orderliness of the operation process of the spatial facility layout under static conditions and the safety issues in dealing with emergency evacuation of people, and determine the layout cost index of each layout scheme.

[0029] Because the spatial facility layout to be coordinated involves inherent contradictions among multiple objectives such as efficiency, safety, and energy consumption (for example, while a compact layout facilitates smooth movement between people, it occupies more emergency evacuation routes, increasing safety hazards), it is not conducive to comprehensively judging the selection of various factors and the ideal layout scheme. When the layout state of the grid is changed using an optimization algorithm (Channel 1), the equipment usage frequency (Channel 2) and per capita traffic flow (Channel 3) of that grid are updated synchronously based on a preset historical data mapping table or simple rules. Therefore, channels 2-5 of the two-dimensional matrix corresponding to each layout scheme are used as input to the entropy weight method, outputting the weight vector of each evaluation index. Furthermore, a weighted sum of all evaluation indices for each grid is performed to obtain the comprehensive evaluation value. Specifically, the comprehensive evaluation value E is obtained by reconstructing all data from channels 2-5 of the extracted two-dimensional matrix. The initial evaluation matrix is ​​defined as follows: each row represents a grid, and each column corresponds to an evaluation index. The higher the overall evaluation value, the stronger the orderliness of the work process and the less redundancy and resource waste under the layout scheme. From the perspective of information theory, it objectively reflects the smoothness of the process and the time economy achieved by the layout scheme in static operation scenarios.

[0030] However, efficiency assessments cannot capture personnel safety risks during emergencies. For example, dense equipment layouts may lead to evacuation congestion and stampede risks when passageways are narrow or exits are unevenly distributed. To address this issue, channels 1, 3, and 5 of the two-dimensional matrix in each layout scheme are used as inputs to the social force model. During the social force model simulation, a fixed random seed is set based on the data from channel 3 in each grid as a probability distribution. Initial personnel positions are generated across all grids, with the initial total number of personnel set to 30% of the total number of grids. Specific positions are assigned based on the normalized data from channel 3 of each grid as a probability distribution. An obstacle map is constructed using the binary codes of all grids, where fixed equipment, movable equipment, and workbenches are considered obstacles; entrances and exits are considered exits; storage areas are considered obstacles because they often contain materials in real-world scenarios; and vacant areas are considered passable areas. The average walking speed is set to a range of 0.8–1.5 m / s, with the lower limit considering the mobility of the elderly and those with mobility impairments, and the upper limit considering the jogging speed in emergencies. In this embodiment, the average walking speed is specifically set to 1.2 m / s. Based on the randomly generated initial personnel positions, the overall evacuation completion time T and the maximum instantaneous congestion density during the evacuation process for each layout scheme are output. The output of the social force model is post-processed: firstly, the overall evacuation completion time T and the maximum instantaneous congestion density are statistically analyzed. Both were standardized using Z-Score standardization, and the standardized overall evacuation completion time T and maximum instantaneous congestion density were then evaluated using the sigmoid function. After normalization, the product of the normalized values ​​and the calculated values ​​is used as the evacuation risk index S for each layout scheme, where the maximum instantaneous congestion density is... The peak value reflects the local risk in the space, while the overall evacuation completion time T reflects the duration of risk exposure. The product of the two describes the duration of the evacuated population in the dangerous state. The larger the value, the more likely the layout is to form high-density congestion in an emergency, and the longer the congestion lasts, resulting in a higher overall safety risk.

[0031] Furthermore, the result of positively fusing the negative correlation mapping result of the comprehensive evaluation value of each layout scheme with the evacuation risk index is used as the layout cost index of each layout scheme. In this embodiment, the negative correlation mapping result of the variable is determined by the reciprocal of the variable. It should be noted that, in order to prevent the denominator from being 0, a non-zero parameter adjustment factor needs to be added to the denominator. In this embodiment, the value is 0.01. In other embodiments, the negative correlation mapping result of the variable can also be obtained by using the opposite number of the variable as the independent variable of the exp() function, where exp() represents an exponential function with the natural constant as the base. In this embodiment, the positive fusion of multiple variables is performed by multiplication.

[0032] It should be understood that the entropy weight method objectively assigns weights to data from multiple indicators and calculates comprehensive evaluation values. In existing technologies, social force models simulate the evacuation dynamics of people, and the output evacuation time multiplied by the crowd density is the risk measure of the space; the larger the value, the more dangerous the space. However, in the actual multi-objective optimization process of layout, layout efficiency and layout safety are mutually restrictive. An efficient layout can easily lead to fewer evacuation channels, thereby increasing the overall safety risk of the layout. The layout cost index is the product of the comprehensive evaluation value and the evacuation risk index. When the comprehensive evaluation value decreases (efficiency decreases) or the evacuation risk index increases (safety risk increases), the layout cost index increases accordingly, further reflecting the nonlinear conflict and amplification effect between efficiency and safety.

[0033] As can be seen from the above, this application combines the comprehensive evaluation value obtained by the entropy weight method with the safety risk index derived from the social force model. In this process, not only are efficiency loss and safety risk amplified by the same order of magnitude, but it can also reflect the total amplification effect caused by the simultaneous occurrence of efficiency decline and risk increase. That is, when the layout is both inefficient and has a large safety hazard, the layout cost index value increases significantly. Therefore, the layout cost index is a global comprehensive cost measure under the dual indicators of layout efficiency and safety.

[0034] Step 3: Analyze the distribution characteristics of evaluation indicators related to efficiency, safety, energy consumption, and flexible channels in the two-dimensional data matrix of each layout scheme, form a multi-dimensional evaluation vector for each layout scheme, obtain the dynamic weight vector using a Bayesian network, and further use decision analysis methods to obtain the proximity score corresponding to each layout scheme. Combine the degree of dispersion of the element values ​​of the dynamic weight vector to obtain the scene adaptation robustness of each layout scheme.

[0035] Because spatial facilities operate in different states, the evaluation dimensions for spatial layout schemes vary depending on the scenario under each state. Therefore, traditional multi-objective optimization methods, which use fixed weights to solve layout problems, may lack adaptability to unforeseen circumstances during actual operation, thus affecting the efficiency and safety of the layout scheme.

[0036] The layout cost index of each layout scheme is used as the efficiency and safety evaluation data for that layout scheme; the sum of all data of all grids in channel 2 for each layout scheme is used as the energy consumption evaluation data for that layout scheme; based on the proportion and distribution of mobile equipment grids identified in channel 1 of each layout scheme, combined with the concentration of high material transfer frequency areas in channel 4, the equipment movement time required for layout adjustment is simulated through digital twin simulation as the flexibility evaluation data for that layout scheme. Specifically, the flexibility evaluation data is obtained by importing the current layout 3D model and equipment attribute data into the digital twin environment, simulating equipment movement paths, and calculating disassembly, handling, and installation times, thereby quantifying the scheme's adaptability to changes in the work process. All the above evaluation data, including the efficiency and safety evaluation data, energy consumption evaluation data, and flexibility evaluation data of each layout scheme, are reverse-processed and then normalized using the sigmoid function to form a multi-dimensional evaluation vector for each layout scheme; in this embodiment, the reverse-processing is performed by taking the reciprocal, and if the denominator is 0, a very small preset parameter is added to the denominator, which is 0.01 in this embodiment.

[0037] Using the multidimensional evaluation vector of each layout scheme as input to the Bayesian network, four typical scenarios are defined: "Daily," "Peak," "Emergency," and "Energy Saving." Specifically, if the probability of a certain scenario is too low, the system's response to that scenario will be insufficient due to its extremely small weight during Bayesian network inference. Similarly, if the probability of a certain scenario is too high, its excessive dominance will suppress the weight generation of other scenarios, leading to an imbalance in the system's scenario response. Therefore, the prior probabilities of each node are given to be in the range of 0 to 0.6. In this embodiment, the probabilities of the four scenarios are specifically determined as follows: Daily mode 0.5, Peak mode 0.2, Emergency mode 0.2, and Energy Saving mode 0.1. The dynamic weight vector W of the three evaluation data defined for each layout scheme is output. ;in, , , These represent the dynamic weights of the efficiency and safety evaluation data, energy consumption evaluation data, and flexibility evaluation data for each layout scheme.

[0038] The dynamic weight vector uses numerical values ​​to determine the weight ratio of the three feature indicators "efficiency and safety", "energy consumption" and "flexibility" in a Bayesian network in a certain operational scenario (such as daily, peak, emergency, energy saving, etc.). This allows the Bayesian network to flexibly adjust the weight ratio between various indicators according to different scenarios, making the layout optimization results more in line with the actual situation.

[0039] However, even with dynamic weights, a quantitative method is still needed to comprehensively compare the merits of multiple candidate layout schemes. Normalized data for each layout scheme across all evaluation dimensions and the dynamic weight vector W are used as inputs to the TOPSOS algorithm. The weighted Euclidean distance method is used to calculate the distance of each layout scheme from the positive / negative ideal solution. The specific weights are determined by the evaluation data for different layout schemes; the positive ideal solution is the optimal value in each dimension, and the negative ideal solution is the worst value in each dimension. Finally, the proximity score C of each layout scheme to the positive ideal solution is output to express the degree of closeness of the current scheme to the target scheme. The normalization method used is the maximum-minimum normalization method.

[0040] Therefore, the scenario adaptation robustness of each layout scheme is calculated. Specifically, the dispersion of the tracking score for each layout scheme under all preset typical scenarios (daily, peak, emergency, energy saving) is calculated, and the negative correlation mapping result of the dispersion is used as the scenario adaptation robustness of each layout scheme. The dispersion can be calculated using the variance, standard deviation, or coefficient of variation between the weights of each evaluation data point. In this embodiment, the standard deviation between the weights of each evaluation data point is used. The negative correlation mapping result of the variable is calculated using the reciprocal of the variable. It should be noted that to prevent the reciprocal calculation from being meaningless when the variable is 0, a preset parameter needs to be added to the denominator; in this embodiment, the value is 0.01. The flowchart for obtaining the scenario adaptation robustness is as follows: Figure 2 As shown.

[0041] The TOPSIS method is a ranking method for solving multi-criteria decision-making. Its output proximity score indicates how close the solution is to the optimal solution. The larger the proximity score, the better the overall performance of the solution under the given weights. In the prior art, the standard deviation of the weights is a statistic that measures the dispersion of a set of weights. The larger the value, the more unbalanced the weight distribution. In dynamic multi-objective optimization, it may cause some indicators to focus excessively on a few indicators, resulting in unreasonable weight distribution and thus reducing the overall robustness of the solution.

[0042] It should be understood that when the weight distribution is relatively balanced, i.e., the dispersion is small, the normalized value approaches 0.5, indicating that the current weights can relatively balance the various dimensions. At this time, the robustness of scene adaptation is mainly dominated by the closeness score, reflecting the true comprehensive level of the solution under balanced weights. When the weight distribution is highly unbalanced, i.e., the dispersion is large, the normalized value approaches 1. At this time, even if the closeness score is high, the robustness of scene adaptation will be significantly suppressed, because the comprehensive advantage of the solution is based on a serious weight bias, and its robustness of scene adaptation is poor.

[0043] Step 4: Based on the scene adaptation robustness and proximity score of each layout scheme, determine the fitness function and use the particle swarm optimization algorithm to obtain the optimal layout scheme.

[0044] After obtaining the multidimensional evaluation vector and scene adaptation robustness of each layout scheme, in order to efficiently search the layout solution space and simultaneously obtain a non-dominated solution set with excellent overall performance and good scene robustness, this application adopts an improved particle swarm optimization algorithm to iteratively optimize the layout scheme. The population size in the particle swarm algorithm is set to half the number of grid cells W to prevent local optima due to an excessively small population size and increased computational burden due to an excessively large population size. The maximum number of iterations is set to 200, the inertia weight is set to 0.5, and both the cognitive learning factor and the social learning factor are set to 2.0 to control the learning intensity of particles towards their own historical best position and the group's best position. The maximum particle velocity is set to 2 to limit particle velocity updates and prevent particle oscillation. The specific particle swarm algorithm content is as follows: Each particle represents a candidate layout scheme, and its position vector is composed of the state encoding sequence of all grids. In this application, the entire spatial facility divided into W grids is regarded as a particle, and each particle is a vector of length W. Each element of the vector is a binary code of length 6, and each element includes the binary code of six attributes: free area, fixed equipment, movable equipment, workbench, entrance / exit, and storage area. This encoding method directly calls the information of channel 1 in the two-dimensional data matrix of each layout scheme to match its specific site layout form.

[0045] The initial population is obtained through two methods: (1) random perturbation based on historical layout or baseline scheme, for example, swapping the device type codes of 2 grids; (2) completely random generation; using the above two methods, the diversity and effectiveness of the solution set are balanced under the premise of ensuring that the population has a certain feasibility; in this embodiment, (2) is adopted.

[0046] To overcome the problems of subjective weight setting and large solution set size in traditional multi-objective optimization, this application incorporates the scenario adaptability robustness of each layout scheme as a core component of the fitness function. Specifically, the fitness function is defined as: Where F represents fitness, α and β are both preset weights, and both take values ​​in the range of 0 to 1, satisfying α + β = 1 to achieve a normalized tradeoff between immediate performance and long-term robustness. The specific values ​​are determined based on the difference between the instantaneous and long-term operation of the facility; in this embodiment, α = 0.7 and β = 0.3 are specifically chosen. In the fitness function, C represents the proximity score of each layout scheme, reflecting the overall performance of the layout scheme in the current scenario; B represents the scenario adaptability robustness of each layout scheme, reflecting the stability of the layout scheme with changes in scenario weights. Therefore, variable B is explicitly added to the optimization objective. During the algorithm's search process, it is necessary to obtain the layout scheme with the highest score for each scenario, while also considering the applicability of the scheme when the weights of other scenarios change, thereby reducing the contradiction between optimization objectives and preventing the obtained layout scheme from being only the optimal layout scheme in a certain scenario.

[0047] The algorithm iteratively generates a particle population and continuously adjusts the particle positions based on the historical best position and the global best position of each individual (layout encoding). It uses a fitness function to measure the quality of newly generated layouts until convergence is achieved. Finally, a set of Pareto non-dominated solutions is obtained, each solution meeting both proximity (C) and robustness (B) requirements. This approach achieves the advantage of high overall computational speed, overcoming the disadvantages of traditional multi-objective evolutionary algorithms, such as excessively large solution sets and the need for extensive post-processing. It can generate optimal layout schemes for specific production and operation scenarios and respond to changes in these scenarios, ultimately ensuring that various production and operation facilities can fully utilize their performance and maintain a high level of reliability and stability during actual use.

[0048] Thus, by introducing scenario-adaptive robustness to make targeted improvements to the particle swarm optimization algorithm, a multi-objective intelligent optimization method for spatial facility layout problems is obtained, which is conducive to achieving more efficient operation and management of facilities throughout their entire life cycle.

[0049] Based on the same inventive concept as the above method, this application embodiment also provides a layout generation and optimization system for a space facility digital twin, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any one of the above-described layout generation and optimization methods for a space facility digital twin.

[0050] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than that shown in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. In the descriptions corresponding to the flowcharts and block diagrams in the accompanying drawings, the operations or steps corresponding to different blocks may also occur in a different order than disclosed in the description; sometimes there is no specific order between different operations or steps. For example, two consecutive operations or steps may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. Each block in a block diagram and / or flowchart, and combinations of blocks in a block diagram and / or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0051] It will be apparent to those skilled in the art that this application is not limited to the details of the exemplary embodiments described above, and that this application can be implemented in other specific forms without departing from its essential characteristics. Therefore, the embodiments described above should be considered exemplary and non-limiting in all respects; modifications to the technical solutions described in the foregoing embodiments, or equivalent substitutions of some technical features, without causing the essence of the corresponding technical solutions to deviate from the scope of the technical solutions in the embodiments of this application, should all be included within the protection scope of this application.

Claims

1. A method for optimizing the layout of a digital twin of a spatial facility, characterized in that, The method includes the following steps: For space facilities, a two-dimensional data matrix for each layout scheme is constructed by using evaluation indicators of a preset number of channels in a discretized grid; the evaluation indicators are used to represent the layout information of the space facilities and the grid. For each layout scheme, the fluctuation characteristics of the evaluation index of the channel in the two-dimensional data matrix are analyzed. Based on the evaluation index in the two-dimensional data matrix, a social force model is constructed to analyze the orderliness of the operation process of the spatial facility layout under static conditions and the safety issues in dealing with the emergency evacuation of the crowd, and to determine the layout cost index of each layout scheme. The distribution characteristics of evaluation indicators related to efficiency and safety, energy consumption, and flexible channels in the two-dimensional data matrix of each layout scheme are analyzed to form a multi-dimensional evaluation vector for each layout scheme. A Bayesian network is used to obtain a dynamic weight vector, and further decision analysis methods are used to obtain the proximity score corresponding to each layout scheme, thereby obtaining the scene adaptation robustness of each layout scheme. Based on the scene adaptability robustness and proximity score of each layout scheme, the fitness function is determined, and the optimal layout scheme is obtained by using an optimization algorithm.

2. The layout generation and optimization method for a digital twin of a space facility as described in claim 1, characterized in that, The construction of the two-dimensional data matrix for each layout scheme is as follows: each row represents a discretized grid, and each column is an evaluation index for a preset channel. There are 5 preset channels. The evaluation index of channel 1 is used to reflect the spatial geometric information of the facilities within the grid; the evaluation index of channel 2 is used to reflect the usage frequency of the facilities within the grid; the evaluation index of channel 3 is used to reflect the per capita flow rate within the grid; the evaluation index of channel 4 is used to reflect the material transfer frequency within the grid; and the evaluation index of channel 5 is used to reflect the distance information of the grid from the entrance and exit.

3. The layout generation and optimization method for a digital twin of a space facility as described in claim 2, characterized in that, The process of determining the layout cost index for each layout scheme is as follows: Based on the aforementioned fluctuation characteristics, the entropy weight method is used to obtain the comprehensive evaluation value of each evaluation indicator; Based on the output of the social force model, the evacuation risk index of each layout scheme is calculated; The result of positively fusing the negative correlation mapping of the comprehensive evaluation value of each layout scheme with the evacuation risk index is used as the layout cost index for each layout scheme.

4. The layout generation and optimization method for a digital twin of a space facility as described in claim 3, characterized in that, The steps for obtaining the comprehensive evaluation value of each evaluation indicator using the entropy weight method are as follows: All channels except channel 1 in the two-dimensional matrix corresponding to each layout scheme are used as input to the entropy weight method. The weight vector of each evaluation index is output, and the weighted sum of all evaluation indices for each grid is used to obtain the comprehensive evaluation value.

5. The layout generation and optimization method for a digital twin of a space facility as described in claim 3, characterized in that, The calculation of the evacuation risk index for each layout scheme is specifically as follows: The social force model outputs the overall evacuation completion time and the maximum instantaneous congestion density during the evacuation process for each layout scheme. The overall evacuation completion time and the maximum instantaneous congestion density are normalized respectively, and the product between the two after normalization is calculated as the evacuation risk index for each layout scheme.

6. The layout generation and optimization method for a digital twin of a space facility as described in claim 2, characterized in that, The process of obtaining the multidimensional evaluation vector for each layout scheme is as follows: For each layout scheme, the layout cost index is used as the efficiency and safety evaluation data; the sum of all data of all grids in channel 2 is used as the energy consumption evaluation data; based on the proportion and distribution of mobile equipment grids in channel 1 of each layout scheme, combined with the concentration of high material transfer frequency areas in channel 4, the equipment movement time required for layout adjustment is simulated through digital twin simulation to obtain flexibility evaluation data. After normalizing all the evaluation data for each layout scheme, a multi-dimensional evaluation vector is formed.

7. The layout generation and optimization method for a digital twin of a space facility as described in claim 1, characterized in that, The proximity score for each layout scheme is obtained as follows: The normalized data and dynamic weight vector of each element in the multidimensional evaluation vector of each layout scheme are used as input to the TOPSOS algorithm, and the output is the closeness score of each layout scheme to the positive ideal solution.

8. The layout generation and optimization method for a digital twin of a space facility as described in claim 1, characterized in that, The scenario adaptability robustness of each layout scheme is obtained as follows: Calculate the dispersion of the stitching score for each layout scheme in all preset typical scenarios, and use the negative correlation mapping result of the dispersion as the scenario adaptation robustness of each layout scheme.

9. The layout generation and optimization method for a digital twin of a space facility as described in claim 1, characterized in that, The fitness function is determined by the following formula: In the formula, F represents fitness, α and β both represent preset weights, C represents the proximity score of each layout scheme, and B represents the scene adaptation robustness of each layout scheme.

10. A layout generation and optimization system for a digital twin of a space facility, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1-9.

Images