Post-earthquake community building group recovery scheduling and toughness evaluation multi-objective optimization method

Abstract

The invention discloses a post-earthquake community building group recovery scheduling and toughness evaluation multi-objective optimization method, and belongs to the technical field of post-earthquake community building group recovery scheduling, and the method comprises the steps: obtaining a seismic source coordinate point, constructing a post-earthquake building group network based on the seismic source coordinate point, and obtaining the image data and building structure information of the post-earthquake building group network; building structure feature data is obtained based on the image data of the building group network and the building structure information, the building structure feature data is calculated through a recurrent neural network, and a building repair strategy is determined after a building toughness prediction score is obtained; schedulable resources are obtained in real time, the balance between the building repair strategy and the schedulable resources is solved based on a multi-target particle swarm optimization algorithm, and an optimal solution set of building repair scheduling is generated; and repairing the building group based on the optimal solution set of the building repair scheduling. Through the multi-target particle swarm optimization algorithm, the balance between the building repair strategy and the schedulable resources is realized.

Description

Technical Field

[0001] The present invention belongs to the technical field of post-earthquake community building complex recovery scheduling, and in particular relates to a multi-objective optimization method for post-earthquake community building complex recovery scheduling and resilience evaluation. Background Art

[0002] Post-earthquake repair is the process of restoring damaged infrastructure, buildings and services after an earthquake. Post-earthquake repair can be divided into two categories: perfect repair and imperfect repair. Perfect repair refers to restoring the damaged structure or facility to its original state before the earthquake, that is, the performance and safety of the repaired structure are the same as those of the newly built state. Imperfect repair refers to repairing the damaged structure or facility to a state where only some functions are restored, rather than a complete restoration, and only the basic stability of the building structure needs to be ensured.

[0003] However, after the earthquake, a large number of buildings were destroyed. How to evaluate and quantify the resilience of community buildings and reasonably arrange the restoration process of a large number of buildings with limited community resources is a major challenge. At the same time, it is also necessary to decide which facilities to repair first and what repair method to use, which is also a hot topic of current research. Summary of the invention

[0004] In order to solve the above technical problems, the present invention proposes a multi-objective optimization method for post-earthquake community building complex recovery scheduling and resilience evaluation to solve the problems existing in the above-mentioned prior art.

[0005] To achieve the above objectives, the present invention provides a multi-objective optimization method for post-earthquake community building complex recovery scheduling and resilience evaluation, comprising:

[0006] Acquire earthquake source coordinate points, construct a post-earthquake building complex network based on the earthquake source coordinate points, and acquire image data and building structure information of the post-earthquake building complex network;

[0007] Acquiring building structure characteristic data based on the image data and building structure information of the building complex network, and calculating the building structure characteristic data through a recursive neural network to obtain a building resilience prediction score;

[0008] determining a building repair strategy based on the building resilience prediction score;

[0009] Acquire schedulable resources in real time, solve the balance between the building repair strategy and the schedulable resources based on a multi-objective particle swarm optimization algorithm, and generate an optimal solution set for building repair scheduling;

[0010] The community buildings are repaired based on the optimal solution set of the building repair scheduling.

[0011] Preferably, the process of obtaining earthquake source coordinate points and constructing a post-earthquake building complex network based on the earthquake source coordinate points includes:

[0012] Obtain the earthquake source coordinate point, and construct a radius range with the earthquake source coordinate point as the center and the distribution density of the building to obtain the post-earthquake building range;

[0013] Determine the spatial distance between buildings in the post-earthquake building range based on the Manhattan distance method, and simultaneously obtain the functional connection between the buildings;

[0014] The post-earthquake building range is used as a spatial constraint, the buildings are used as nodes, and the spatial distances between the buildings and the functions between the buildings are used as edges in the network to construct the post-earthquake building complex network.

[0015] Preferably, the process of acquiring the image data and building structure information of the post-earthquake building complex network includes:

[0016] Constructing a flight route for the drone based on the spatial distance between buildings in the post-earthquake building range, parameter information of a depth camera on the drone, weather information, flight belt overlap, and flight belt width;

[0017] The drone takes aerial photos of the target area through the drone flight route to obtain image data of the post-earthquake building complex network;

[0018] The building geometry information, structure type and material, component information and building system information are acquired and integrated based on the historical building data to generate the building structure information.

[0019] Preferably, the process of acquiring building structure characteristic data based on the image data and building structure information of the building complex network includes:

[0020] Determine the initial functional status of the building after the earthquake based on historical building information;

[0021] Fusing the image data of the building complex network and the features of the building structure information to obtain a fused data set;

[0022] The features of the fused data set are extracted based on a multi-layer perceptron, and the extracted features are classified according to the initial functional state of the building after the earthquake to obtain the building structure feature data.

[0023] Preferably, the process of fusing the image data of the building complex network and the features of the building structure information to obtain a fused data set includes:

[0024] Extracting features of the building structure information based on an autoencoder to obtain key feature data;

[0025] The image data is cropped into small-scale blocks, and then the pixel values ​​of the small-scale blocks are normalized to the interval [0, 1] to generate normalized blocks;

[0026] Obtaining the correlation between the key feature data and the normalized image block through the Spearman rank correlation coefficient;

[0027] The statistical indicators of the key feature data are converted into multiple data views and mapped onto the spatial distribution of the normalized image blocks to obtain the fused data set.

[0028] Preferably, the process of calculating the building structure characteristic data by a recursive neural network to obtain a building resilience prediction score comprises:

[0029] Obtain historical building feature data and build a spatial scaling attention mechanism and recurrent neural network;

[0030] Connecting the hidden layer of the recursive neural network to the spatial scaling attention mechanism to generate an optimized recursive neural network;

[0031] Training the optimized recursive neural network based on the historical building characteristic data to obtain a resilience prediction network;

[0032] The building structure characteristic data is input into the resilience prediction network for evaluation to generate the building resilience prediction score.

[0033] Preferably, the expression of the spatial scaling attention mechanism is:

[0034]

[0035]

[0036] Among them, Concat() is a concatenation operation. is the weight parameter, N i is the number of attention heads in the i-th stage, so the dimension of each head is LSR() performs a linear dimensionality reduction operation on the spatial dimension of the input sequence.

[0037] Preferably, the process of determining a building repair strategy based on the building resilience prediction score comprises:

[0038] Assessing the current structural condition and damage of the building based on the building resilience prediction score;

[0039] Determining a repair strategy for the building based on the current structural status and damage of the building;

[0040] The building restoration strategy includes perfect building restoration and imperfect building restoration.

[0041] Preferably, the process of generating an optimal solution set for building repair scheduling includes:

[0042] Construct an objective function based on minimizing repair cost and minimizing repair time;

[0043] Acquire schedulable resources in real time, and initialize a group of particles based on the schedulable resources and the building repair strategy, wherein each particle represents a repair strategy;

[0044] Iteratively update the position and velocity of the particle based on the objective function to generate an optimal particle swarm in the search space;

[0045] Converting the optimal particle swarm into a Pareto front, and evaluating each solution of the Pareto front to generate an evaluation result;

[0046] Construct a priority whitelist based on historic building attributes;

[0047] An optimal solution set for building repair scheduling is constructed based on the priority whitelist and the evaluation results.

[0048] Compared with the prior art, the present invention has the following advantages and technical effects:

[0049] The present invention can construct a complete post-earthquake building complex network by acquiring the earthquake source coordinate points and building structure information. This comprehensive building complex network not only includes geographic spatial relationships, but also covers the structural information of the buildings, thereby providing comprehensive basic data. By using a recursive neural network to process the building structure feature data, the resilience score of the building can be accurately predicted. This method can take into account the complex building structure characteristics and provide more accurate and reliable resilience assessment results. According to the resilience score of the building, a data-driven repair strategy is formulated. In this way, those buildings with lower resilience and higher risks can be given priority, improving the efficiency and effectiveness of repair.

[0050] In addition, the present invention achieves a balance between building repair strategies and schedulable resources through a multi-objective particle swarm optimization algorithm. This method can simultaneously consider multiple conflicting objectives, such as minimizing repair costs, maximizing repair efficiency, etc., to generate an optimal repair scheduling plan. Through the generated optimal repair scheduling set, resources can be effectively managed and allocated to achieve rapid and effective repair of the entire community building complex. This method can minimize the recovery time and cost of the community after the disaster and improve the community's disaster resistance and sustainable development. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0052] Figure 1 This is a flow chart of a multi-objective optimization method for post-earthquake community building complex recovery scheduling and resilience evaluation according to an embodiment of the present invention. DETAILED DESCRIPTION

[0053] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0054] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0055] Embodiment 1

[0056] like Figure 1 As shown, this embodiment provides a multi-objective optimization method for post-earthquake community building complex recovery scheduling and resilience evaluation, including:

[0057] This embodiment uses a multi-objective optimization analysis method to quickly predict the recovery scheduling and resilience of community building groups after an earthquake, taking into account the functional interdependence between buildings under the constraint of limited recovery resources. This method can provide a useful reference for community construction group resilience evaluation and recovery plan formulation, and provide a scientific basis for earthquake disaster prevention and reduction decision-making and planning of urban management agencies. The specific process includes:

[0058] Acquire earthquake source coordinate points, construct a post-earthquake building complex network based on the earthquake source coordinate points, and acquire image data and building structure information of the post-earthquake building complex network;

[0059] Acquiring building structure characteristic data based on the image data and building structure information of the building complex network, and calculating the building structure characteristic data through a recursive neural network to obtain a building resilience prediction score;

[0060] determining a building repair strategy based on the building resilience prediction score;

[0061] Acquire schedulable resources in real time, solve the balance between the building repair strategy and the schedulable resources based on a multi-objective particle swarm optimization algorithm, and generate an optimal solution set for building repair scheduling;

[0062] The community buildings are repaired based on the optimal solution set of the building repair scheduling.

[0063] Further optimizing the scheme, the process of obtaining the earthquake source coordinate points and constructing the post-earthquake building complex network based on the earthquake source coordinate points includes:

[0064] Obtain the earthquake source coordinate point, and construct a radius range with the earthquake source coordinate point as the center and the distribution density of the building to obtain the post-earthquake building range;

[0065] Determine the spatial distance between buildings in the post-earthquake building range based on the Manhattan distance method, and simultaneously obtain the functional connection between the buildings;

[0066] The post-earthquake building range is used as a spatial constraint, the buildings are used as nodes, and the spatial distances between the buildings and the functions between the buildings are used as edges in the network to construct the post-earthquake building complex network.

[0067] Based on the spatial distance between buildings, the proximity relationship is determined. Classical spatial analysis methods such as Euclidean distance or Manhattan distance can be used to measure the distance between buildings.

[0068] Considering the functional connections between buildings, for example, closer connections can be established between buildings with the same purpose or buildings that share the same infrastructure.

[0069] Convert buildings and their connections into a network model: Each building is a node in the network, and the node contains the building's attribute information. The connection between buildings is the edge in the network, and the edge can have a weight to reflect the strength of the association or spatial distance between buildings.

[0070] Analyze and optimize the constructed post-earthquake building complex network: Use network analysis methods (such as graph theory methods) to explore the structural characteristics of the network, such as node degree, clustering coefficient, centrality, etc.

[0071] Further optimizing the scheme, the process of obtaining the image data and building structure information of the post-earthquake building complex network includes:

[0072] Constructing a flight route for the drone based on the spatial distance between buildings in the post-earthquake building range, parameter information of a depth camera on the drone, weather information, flight belt overlap, and flight belt width;

[0073] The drone takes aerial photos of the target area through the drone flight route to obtain image data of the post-earthquake building complex network;

[0074] The building geometry information, structure type and material, component information and building system information are acquired and integrated based on the historical building data to generate the building structure information.

[0075] Structural features include but are not limited to:

[0076] Basic information of the building: such as the age, height, purpose, etc. of the building.

[0077] Structural type and material: such as reinforced concrete structure, steel structure, etc.

[0078] Building component information: such as the size, material and connection method of columns, beams, slabs, etc.

[0079] Geographic location information: geographical coordinates of the building, frequency of earthquake activity and other geographical environment information.

[0080] Further optimizing the scheme, the process of obtaining building structure characteristic data based on the image data and building structure information of the building complex network includes:

[0081] Determine the initial functional status of the building after the earthquake based on historical building information;

[0082] Fusing the image data of the building complex network and the features of the building structure information to obtain a fused data set;

[0083] The features of the fused data set are extracted based on a multi-layer perceptron, and the extracted features are classified according to the initial functional state of the building after the earthquake to obtain the building structure feature data.

[0084] The scheme is further optimized to fuse the image data of the building complex network with the features of the building structure information. The process of obtaining the fused data set includes:

[0085] Extracting features of the building structure information based on an autoencoder to obtain key feature data;

[0086] The image data is cropped into small-scale blocks, and then the pixel values ​​of the small-scale blocks are normalized to the interval [0, 1] to generate normalized blocks;

[0087] Obtaining the correlation between the key feature data and the normalized image block through the Spearman rank correlation coefficient;

[0088] Set the key feature data to (X = {x_1, x_2, \ldots, x_n}), and the data device of the normalized block to (Y = {y_1, y_2, \ldots, y_n}). First, we need to convert the observations of each set of data into their ranks (arranged from small to large):

[0089] Calculate the rank (\text{rank}(x_i)) of (X), denoted as (R_X).

[0090] Calculate the rank (\text{rank}(y_i)) of (Y), denoted as (R_Y).

[0091] Then, the Spearman rank correlation coefficient (\rho) can be calculated by the following formula:

[0092] [\rho=1-\frac{6\sum_{i=1}^{n}(R_{X,i}-R_{Y,i})^2}{n(n^2-1)}]

[0093] in,

[0094] (R_{X,i}) is the rank of (x_i) in (X),

[0095] (R_{Y,i}) is the rank of (y_i) in (Y),

[0096] (n) is the number of sample data.

[0097] The numerator in this formula (\sum_{i=1}^{n}(R_{X,i}-R_{Y,i})^2) represents the sum of the squares of the differences in rank between the two sets of data, and the denominator (n(n^2-1)) is a factor used to normalize the coefficient so that the Spearman rank correlation coefficient ranges from -1 to 1.

[0098] The statistical indicators of the key feature data are converted into multiple data views and mapped onto the spatial distribution of the normalized image blocks to obtain the fused data set.

[0099] Further optimizing the scheme, the process of calculating the building structure characteristic data by a recursive neural network to obtain a building resilience prediction score includes:

[0100] Obtain historical building feature data and build a spatial scaling attention mechanism and recurrent neural network;

[0101] Connecting the hidden layer of the recursive neural network to the spatial scaling attention mechanism to generate an optimized recursive neural network;

[0102] Training the optimized recursive neural network based on the historical building characteristic data to obtain a resilience prediction network;

[0103] Specifically, the sample set contains two parts, one is the training features, and the other is the label; the training features are used as the input of the model, and the labels are used to verify the output of the pre-trained sorting model: when constructing sample pairs, a pairing method will be adopted. Assuming that there are 100 training technology data sets and image sets from 1 to 100, then the training features of No. 1 will form 99 groups of sample pairs with the training features of No. 2-100, and the training features of No. 2 will form 99 groups of sample pairs with the training features of No. 1 and No. 3 to 100. Therefore, in principle, for the training samples of n training plots, n*(n-1) sample pairs can be formed, that is, a large number of samples are not required to construct the sample set.

[0104] Precision, recall, average precision (AP) and mean average precision (mAP) are used as evaluation indicators.

[0105] Precision is the proportion of true positive samples among all predicted positive samples, and is calculated as:

[0106]

[0107] Among them, TP is the number of positive samples predicted by the model as positive samples, and FP is the number of negative samples predicted by the model as positive samples.

[0108] The recall rate is the proportion of all true positive samples that are predicted to be positive samples. The formula is as follows:

[0109]

[0110] Among them, TP is the number of positive samples predicted by the model as positive samples, and FN is the number of positive samples predicted by the model as negative samples.

[0111] The horizontal axis of the PR curve is the recall rate, and the vertical axis is the precision rate, which is also called the precision-recall curve. The average precision is the surface area enclosed under the PR curve. The formula is as follows:

[0112]

[0113] In the field of target detection, data sets usually contain multiple categories. AP is often used to measure the detection effect of a model for a certain category of targets. If the detection effect of multiple categories of targets is evaluated, the AP value of each category needs to be summed and averaged to obtain the mean average precision. mAP is a more comprehensive evaluation indicator that can reflect the overall performance of the model in multi-category target detection tasks. The formula is as follows:

[0114]

[0115] The IoU threshold affects the mAP value of the model. The commonly used mAP indicators are mAP@0.5 and mAP@0.5:0.95. mAP@0.5 represents the average precision of all categories when the IoU threshold is 0.5.

[0116] The building structure characteristic data is input into the resilience prediction network for evaluation to generate the building resilience prediction score.

[0117] The spatial scaling attention mechanism is a mechanism for processing multi-scale data and relationships, which is applicable to multi-level features in building structures. It can adjust the weights according to the feature relationships at different scales, thereby more effectively capturing the importance of different parts of the building structure.

[0118] Collect and prepare structural characteristic data for building resilience prediction, including but not limited to information such as the building's structural type, material, component size, and geographical location.

[0119] The prepared building structure feature data is used as the input data of RNN. These data can be organized into a sequence or graph structure, depending on how the structural features of the building are represented.

[0120] In the design of RNN, the spatial scaling attention mechanism is incorporated to enhance the model's understanding and utilization of the relationship between different features. The usual steps include:

[0121] Map the original feature data into a high-dimensional space to facilitate neural network processing.

[0122] Design a spatially scaled attention mechanism to adjust the weights based on the spatial relationship between different features. This can be achieved in many ways, such as weighting the attention based on position information, or considering the relative distance between features.

[0123] Ensure that the model can simultaneously process and integrate architectural structural feature data from different scales. This can be achieved through a multi-level attention mechanism or a multi-layer RNN structure.

[0124] Using a recurrent neural network with spatially scaled attention to calculate building resilience prediction scores on building structural feature data:

[0125] Further optimizing the scheme, the expression of the spatial scaling attention mechanism is:

[0126]

[0127] Among them, Concat() is a concatenation operation. is the weight parameter, N i is the number of attention heads in the i-th stage, so the dimension of each head is LSR() performs a linear dimensionality reduction operation on the spatial dimension of the input sequence.

[0128] LSR(x)=Norm(Pooling(x,P)W S );

[0129] in, W represents the length of the input sequence, and P represents the size of the attention layer reduction at each stage, which defaults to 7. S is to reduce the input sequence to C i dimensional linear projection operation.

[0130] Further optimizing the scheme, the process of determining a building repair strategy based on the building resilience prediction score includes:

[0131] Assessing the current structural condition and damage of the building based on the building resilience prediction score;

[0132] Determining a repair strategy for the building based on the current structural status and damage of the building;

[0133] The building restoration strategy includes perfect building restoration and imperfect building restoration.

[0134] For critical infrastructure such as hospitals, water plants, power plants and other lifeline projects, perfect repairs should be carried out immediately to restore their functions as soon as possible; while for temporary resettlement buildings or material storage facilities, a non-perfect repair strategy with basic stability guarantees can be adopted, only ensuring the safety and stability of the structure in the short term.

[0135] Further optimizing the scheme, the process of generating an optimal solution set for building repair scheduling includes:

[0136] Construct an objective function based on minimizing repair cost and minimizing repair time;

[0137] Acquire schedulable resources in real time, and initialize a group of particles based on the schedulable resources and the building repair strategy, wherein each particle represents a repair strategy;

[0138] Iteratively update the position and velocity of the particle based on the objective function to generate an optimal particle swarm in the search space;

[0139] Converting the optimal particle swarm into a Pareto front, and evaluating each solution of the Pareto front to generate an evaluation result;

[0140] Construct a priority whitelist based on historic building attributes;

[0141] An optimal solution set for building repair scheduling is constructed based on the priority whitelist and the evaluation results.

[0142] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A multi-objective optimization method for post-earthquake community building complex recovery scheduling and resilience evaluation, characterized in that: The following steps are involved: Acquire earthquake source coordinate points, construct a post-earthquake building complex network based on the earthquake source coordinate points, and acquire image data and building structure information of the post-earthquake building complex network; Acquiring building structure characteristic data based on the image data and building structure information of the building complex network, and calculating the building structure characteristic data through a recursive neural network to obtain a building resilience prediction score; determining a building repair strategy based on the building resilience prediction score; Acquire schedulable resources in real time, solve the balance between the building repair strategy and the schedulable resources based on a multi-objective particle swarm optimization algorithm, and generate an optimal solution set for building repair scheduling; The community buildings are repaired based on the optimal solution set of the building repair scheduling.

2. The multi-objective optimization method for post-earthquake community building complex recovery scheduling and resilience evaluation according to claim 1 is characterized by: The process of obtaining earthquake source coordinate points and constructing a post-earthquake building complex network based on the earthquake source coordinate points includes: Obtain the earthquake source coordinate point, and construct a radius range with the earthquake source coordinate point as the center and the distribution density of the building to obtain the post-earthquake building range; Determine the spatial distance between buildings in the post-earthquake building range based on the Manhattan distance method, and simultaneously obtain the functional connection between the buildings; The post-earthquake building range is used as a spatial constraint, the buildings are used as nodes, and the spatial distances between the buildings and the functions between the buildings are used as edges in the network to construct the post-earthquake building complex network.

3. The multi-objective optimization method for post-earthquake community building complex recovery scheduling and resilience evaluation according to claim 2 is characterized in that: The process of obtaining the image data and building structure information of the post-earthquake building complex network includes: Constructing a flight route for the drone based on the spatial distance between buildings in the post-earthquake building range, parameter information of a depth camera on the drone, weather information, flight belt overlap, and flight belt width; The drone takes aerial photos of the target area through the drone flight route to obtain image data of the post-earthquake building complex network; The building geometry information, structure type and material, component information and building system information are acquired and integrated based on the historical building data to generate the building structure information.

4. The multi-objective optimization method for post-earthquake community building complex recovery scheduling and resilience evaluation according to claim 1 is characterized in that: The process of acquiring building structure characteristic data based on the image data and building structure information of the building complex network includes: Determine the initial functional status of the building after the earthquake based on historical building information; Fusing the image data of the building complex network and the features of the building structure information to obtain a fused data set; The features of the fused data set are extracted based on a multi-layer perceptron, and the extracted features are classified according to the initial functional state of the building after the earthquake to obtain the building structure feature data.

5. The multi-objective optimization method for post-earthquake community building complex recovery scheduling and resilience evaluation according to claim 4 is characterized in that: The process of fusing the image data of the building complex network and the features of the building structure information to obtain a fused data set includes: Extracting features of the building structure information based on an autoencoder to obtain key feature data; The image data is cropped into small-scale blocks, and then the pixel values ​​of the small-scale blocks are normalized to the interval [0, 1] to generate normalized blocks; Obtaining the correlation between the key feature data and the normalized image block through the Spearman rank correlation coefficient; The statistical indicators of the key feature data are converted into multiple data views and mapped onto the spatial distribution of the normalized image blocks to obtain the fused data set.

6. The multi-objective optimization method for post-earthquake community building complex recovery scheduling and resilience evaluation according to claim 1 is characterized in that: The process of calculating the building structure characteristic data by a recursive neural network to obtain a building resilience prediction score includes: Obtain historical building feature data and build a spatial scaling attention mechanism and recurrent neural network; Connecting the hidden layer of the recursive neural network to the spatial scaling attention mechanism to generate an optimized recursive neural network; Training the optimized recursive neural network based on the historical building characteristic data to obtain a resilience prediction network; The building structure characteristic data is input into the resilience prediction network for evaluation to generate the building resilience prediction score.

7. The multi-objective optimization method for post-earthquake community building complex recovery scheduling and resilience evaluation according to claim 6 is characterized in that: The expression of the spatial scaling attention mechanism is: Among them, Concat() is a concatenation operation. is the weight parameter, N i is the number of attention heads in the i-th stage, so the dimension of each head is LSR() performs a linear dimensionality reduction operation on the spatial dimension of the input sequence.

8. The multi-objective optimization method for post-earthquake community building complex recovery scheduling and resilience evaluation according to claim 1 is characterized in that: The process of determining a building restoration strategy based on the building resilience prediction score includes: Assessing the current structural condition and damage of the building based on the building resilience prediction score; Determining a repair strategy for the building based on the current structural status and damage of the building; The building restoration strategy includes perfect building restoration and imperfect building restoration.

9. The multi-objective optimization method for post-earthquake community building complex recovery scheduling and resilience evaluation according to claim 1 is characterized in that: The process of generating an optimal solution set for building repair scheduling includes: Construct an objective function based on minimizing repair cost and minimizing repair time; Acquire schedulable resources in real time, and initialize a group of particles based on the schedulable resources and the building repair strategy, wherein each particle represents a repair strategy; Iteratively update the position and velocity of the particle based on the objective function to generate an optimal particle swarm in the search space; Converting the optimal particle swarm into a Pareto front, and evaluating each solution of the Pareto front to generate an evaluation result; Construct a priority whitelist based on historic building attributes; An optimal solution set for building repair scheduling is constructed based on the priority whitelist and the evaluation results.

Images