A multi-source data fusion intelligent city construction analysis method

Abstract

The application discloses a kind of multi-source data fusion wisdom city construction analysis method, it is related to data processing technical field, first, the multi-source heterogeneous data of city construction project is acquired, the multi-source heterogeneous data covers multiple data that influence construction progress;Second, construction progress prediction model is constructed, and innovatively multi-strategy particle swarm optimization algorithm is used to carry out global optimization to the hyperparameter of model, to improve the generalization ability and prediction accuracy of model, obtain the optimized construction progress prediction model;Then, the multi-source heterogeneous data is analyzed and predicted using the optimized construction progress prediction model, and the accurate construction progress prediction result is output;Finally, the prediction result is used as the core basis of wisdom city construction analysis, and is pushed to the terminal equipment designated by staff in real time, provides data support for project management decision, significantly improves the scientificity and reliability of construction progress prediction, and has important significance to improve the management efficiency and fine level of wisdom city construction.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to a smart city construction analysis method based on multi-source data fusion. BACKGROUND

[0002] With the acceleration of urbanization, urban construction projects are increasingly characterized by large scale, long cycle, multiple participants, and strong uncertainty. Effective control of construction progress is the core link to ensure timely delivery of projects, control costs, and ensure quality. Traditional progress management mainly relies on the experience of project managers and regular site meetings. Progress information is mainly collected manually and through regular inspections, with low data update frequency, which cannot reflect the real situation on site in real time. Moreover, the data source is single, making it difficult to fully consider the dynamic influence of multiple complex factors such as weather, supply chain, manpower, and equipment. Because of the lack of quantitative analysis tools, decision-makers have difficulty assessing the impact of different adjustment schemes on the final duration, resulting in a strong subjective decision-making process and high risk. SUMMARY

[0003] The present application provides a smart city construction analysis method based on multi-source data fusion, aiming to solve the problem of single data source in the prior art, which makes it difficult to fully consider the dynamic influence of multiple complex factors such as weather, supply chain, manpower, and equipment. Moreover, because of the lack of quantitative analysis tools, decision-makers have difficulty assessing the impact of different adjustment schemes on the final duration, resulting in a strong subjective decision-making process and high risk.

[0004] The present application provides a smart city construction analysis method based on multi-source data fusion, comprising:

[0005] Obtaining multi-source heterogeneous data of urban construction projects; wherein the multi-source heterogeneous data includes multiple construction progress influencing data;

[0006] Constructing a construction progress prediction model and optimizing the construction progress prediction model using a multi-strategy particle swarm optimization algorithm to obtain an optimized construction progress prediction model;

[0007] Using the optimized construction progress prediction model to predict the multi-source heterogeneous data of the urban construction project to determine the construction progress prediction result;

[0008] Taking the construction progress prediction result as the smart city construction analysis result and transmitting the smart city construction analysis result to the device designated by the staff.

[0009] In some possible implementations, after taking the construction progress prediction result as the smart city construction analysis result, the method further comprises:

[0010] Obtaining expected progress requirements input by a staff, and deploying construction resources of the urban construction project according to the expected progress requirements and the construction progress prediction result.

[0011] In some possible implementation manners, the plurality of construction progress influence data includes urban construction project basis data, environmental influence data, construction technology influence data, and construction resource influence data.

[0012] In some possible implementation manners, the construction progress prediction model is constructed, and a multi-strategy particle swarm optimization algorithm is used to optimize the construction progress prediction model to obtain the optimized construction progress prediction model, including:

[0013] A convolutional neural network is used to construct the construction progress prediction model, and historical multi-source heterogeneous data input by the staff or pre-stored and real progress corresponding to the historical multi-source heterogeneous data are obtained;

[0014] Sample data are constructed by using the historical multi-source heterogeneous data, and sample labels are constructed by using the real progress corresponding to the historical multi-source heterogeneous data, to obtain training data;

[0015] Based on the hyperparameters of the construction progress prediction model, a particle swarm is initialized, wherein the particle swarm includes a plurality of different particles, and each particle includes all to-be-trained hyperparameters of the construction progress prediction model;

[0016] Based on the training data, an error function value corresponding to each particle is obtained, and a particle with the minimum error function value is determined as an optimal particle;

[0017] Based on the optimal particle, a golden sine chain search strategy is used to perform initial search on the particles in the particle swarm, to obtain particles after initial search;

[0018] Based on the optimal particle, a particle swarm information fusion search strategy is used to perform information fusion on the particles after initial search, to obtain particles after information fusion;

[0019] An adaptive global mutation strategy is used to perform adaptive global search on the particles after information fusion, to obtain particles after adaptive global search;

[0020] The golden sine chain search strategy, the particle swarm information fusion search strategy, and the adaptive global mutation strategy are repeatedly executed until an optimization end condition is met, and an optimal particle is reacquired;

[0021] Hyperparameters in the reacquired optimal particle are used as final hyperparameters of the construction progress prediction model, to obtain the optimized construction progress prediction model.

[0022] In some possible implementation manners, the particle swarm is initialized based on the hyperparameters of the construction progress prediction model, and the initialization includes:

[0023] The hyperparameters are randomly initialized between upper and lower limits corresponding to the hyperparameters of the construction progress prediction model, and a vector composed of the hyperparameters after random initialization is obtained to obtain a particle. After multiple repeated initializations, a particle swarm is obtained.

[0024] In some possible implementation manners, the error function value corresponding to each particle is obtained based on the training data, and the obtaining includes:

[0025] The sample data in the training data is taken as input data, the sample label in the training data is taken as expected output data, and the root mean square error function is used to obtain the error function value corresponding to each particle.

[0026] In some possible implementation manners, the initial search is performed on the particles in the particle swarm based on the optimal particle by using a golden sine chain search strategy to obtain particles after the initial search, and the initial search includes:

[0027] The particles in the particle swarm are arranged in descending order of error function values to obtain particles after the arrangement;

[0028] For any one of the particles after the arrangement, another particle is matched to the particle according to the arrangement number of the particle to obtain a paired particle corresponding to each particle;

[0029] The initial search is performed on the particles in the particle swarm according to the optimal particle and the paired particle to obtain the particles after the initial search.

[0030] In some possible implementation manners, the information fusion is performed on the particles after the initial search based on the optimal particle by using a particle swarm information fusion search strategy to obtain particles after the information fusion, and the information fusion includes:

[0031] For any one of the particles after the initial search, the Euclidean distance between the particle after the initial search and other particles is determined;

[0032] The particle information learning scale parameter is obtained according to the Euclidean distance between the particle after the initial search and other particles;

[0033] For any one of the particles after the initial search, the particle swarm fusion information corresponding to the particle after the initial search is obtained according to the particle information learning scale parameter;

[0034] The information fusion is performed on the particles after the initial search according to the optimal particle and the particle swarm fusion information corresponding to the particles after the initial search to obtain the particles after the information fusion.

[0035] In some possible implementation manners, the particle after the information fusion is subjected to adaptive global search by adopting an adaptive global mutation strategy, to obtain a particle after adaptive global search, including:

[0036] obtaining a global mutation factor;

[0037] obtaining an adaptive inertia weight, a first adaptive learning factor and a second adaptive learning factor based on a current optimization number;

[0038] obtaining a search speed of the particle after the information fusion in a current optimization process according to the global mutation factor, the adaptive inertia weight, the first adaptive learning factor and the second adaptive learning factor;

[0039] subjecting the particle after the information fusion to adaptive global search according to the search speed of the particle after the information fusion in the current optimization process, to obtain a particle after adaptive global search.

[0040] In some possible implementation manners, the construction progress prediction result is taken as a smart city construction analysis result, and the smart city construction analysis result is transmitted to a device designated by a staff member, including:

[0041] taking the construction progress prediction result as a smart city construction analysis result;

[0042] encrypting the smart city construction analysis result and the corresponding multi-source heterogeneous data by adopting a symmetric encryption algorithm, to obtain target transmission data carrying the smart city construction analysis result;

[0043] transmitting the target transmission data carrying the smart city construction analysis result to the device designated by the staff member.

[0044] Beneficial effects:

[0045] The application provides a smart city construction analysis method based on multi-source data fusion. First, multi-source heterogeneous data of a city construction project is obtained, which covers various data affecting construction progress. Second, an initial construction progress prediction model is constructed, and a multi-strategy particle swarm optimization algorithm is innovatively adopted to globally optimize hyperparameters of the model, so as to improve generalization ability and prediction accuracy of the model, and obtain an optimized construction progress prediction model. Third, the optimized construction progress prediction model is used to analyze and predict the multi-source heterogeneous data, and an accurate construction progress prediction result is output. Finally, the prediction result is taken as a core basis for smart city construction analysis, and is pushed to a terminal device designated by a staff member in real time, to provide data support for project management decision-making, and significantly improve scientificity and reliability of construction progress prediction, which is of great significance to improving management efficiency and fine level of smart city construction. Attached Figure Description

[0046] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments of this application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0047] Figure 1 This is a flowchart of a smart city construction analysis method based on multi-source data fusion proposed in one embodiment of this application. Detailed Implementation

[0048] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0049] Please refer to Figure 1 The flowchart below illustrates a smart city construction analysis method based on multi-source data fusion, as provided in this application embodiment. The method includes:

[0050] S11. Obtain multi-source heterogeneous data of urban construction projects; wherein, the multi-source heterogeneous data includes various construction progress impact data;

[0051] In some possible implementations, the various construction progress impact data include basic data of urban construction projects, environmental impact data, construction technology impact data, and construction resource impact data.

[0052] The urban construction project can be an urban building construction project, and the urban construction project basis data can be related building information such as the floor area, volume, size, floor height, and foundation depth of the urban construction project. The building information can be obtained from design drawings, and should be arranged in a fixed order for easy identification. The environmental impact data can be the soil quality of the location of the urban construction project (coded as a number), weather information before the time point to be predicted (for example, if the actual time is January 1st and the time to be predicted is January 15th, the weather information from January 1st to January 15th can be included, coded as a number), and working day information (for example, whether the dates between January 1st and January 15th are working days or non-working days, coded as a number). The construction technology impact data can be the technology code of the construction process (for example, the core tube construction of a super high-rise building requires a special climbing formwork system, and the construction of a large-span structure requires a complex support system or a prefabricated assembly technology for the bottom layer. A construction technology library can be pre-set, and each technology in the construction technology library is configured with a unique code, which can be selected by the staff to select the specific construction technology). The construction resource impact data can be the equipment type (coded as a number), the number of equipment under each equipment type, the construction personnel type (coded as a number), and the number of construction personnel under each construction personnel type.

[0053] The urban construction project basis data, environmental impact data, construction technology impact data, and construction resource impact data can then be arranged into a data matrix according to a fixed row code. For example, the urban construction project basis data is pre-set to occupy two rows of data (the specific number of rows is set by the staff), and the urban construction project basis data is placed in the two rows. However, it should be noted that the urban construction project basis data, environmental impact data, construction technology impact data, and construction resource impact data should all be able to be placed in the corresponding rows under the condition that the number of fixed data in each row is guaranteed. The urban construction project basis data, environmental impact data, construction technology impact data, and construction resource impact data can also be arranged as one row of data each to form a data matrix. If a row of data is not enough, zero padding can be performed before or after the data to form an identifiable data matrix.

[0054] S12, constructing a construction progress prediction model, and optimizing the construction progress prediction model by using a multi-strategy particle swarm optimization algorithm to obtain an optimized construction progress prediction model;

[0055] The construction progress prediction model can be constructed based on a deep learning model, such as a convolutional neural network, to facilitate the identification of multi-source heterogeneous data of the urban construction project.

[0056] The simple regression model or the standard neural network has limited generalization ability and prediction accuracy when facing highly nonlinear and dynamically changing construction environments, and the optimization of the hyperparameters depends on the Adam optimizer for adjustment, which is inefficient and difficult to find a global optimal solution. Therefore, the construction progress prediction model is optimized by the multi-strategy particle swarm optimization algorithm, the optimization ability of the construction progress prediction model is improved, so that the optimized construction progress prediction model can improve the prediction accuracy of the construction progress under the consideration of multi-source heterogeneous data.

[0057] S13, using the optimized construction progress prediction model to predict the multi-source heterogeneous data of the urban construction project to determine the construction progress prediction result;

[0058] The optimized construction progress prediction model learns a large amount of historical urban construction project related data, and these historical urban construction project related data usually reflect the correlation with the construction progress, so the optimized construction progress prediction model has the ability to predict the multi-source heterogeneous data of the urban construction project. Inputting the multi-source heterogeneous data into the optimized construction progress prediction model can obtain the construction progress prediction result.

[0059] Optionally, before the multi-source heterogeneous data of the urban construction project is input into the optimized construction progress prediction model, the multi-source heterogeneous data can also be normalized to reduce the complexity of the data and improve the data recognition efficiency of the optimized construction progress prediction model. However, it is worth noting that if the multi-source heterogeneous data of the urban construction project is normalized, the sample data needs to be normalized before being input into the construction progress prediction model during the process of obtaining the optimized construction progress prediction model, so as to ensure that the data can be effectively identified.

[0060] S14, the construction progress prediction result is used as the analysis result of the smart city construction, and the analysis result of the smart city construction is transmitted to the device designated by the staff.

[0061] The embodiment of the application provides a kind of multi-source data fusion wisdom city construction analysis method, first, the multi-source heterogeneous data of city construction project is acquired, the multi-source heterogeneous data covers multiple data affecting construction progress;Second, the initial construction progress prediction model is constructed, and the hyperparameters of the model are globally optimized using multi-strategy particle swarm optimization algorithm innovatively, to improve the generalization ability and prediction accuracy of the model, to obtain the optimized construction progress prediction model;Then, the multi-source heterogeneous data is analyzed and predicted using the optimized construction progress prediction model, and the accurate construction progress prediction result is output;Finally, the prediction result is used as the core basis of wisdom city construction analysis, and is pushed to the terminal equipment designated by staff in real time, to provide data support for project management decision, significantly improve the scientificity and reliability of construction progress prediction, and it has important significance to improve the management efficiency and fine level of wisdom city construction.

[0062] In some possible implementation manners, after the construction progress prediction result is used as the wisdom city construction analysis result, the method further includes:

[0063] The expected progress requirement input by the staff is acquired, and the construction resources of the city construction project are allocated according to the expected progress requirement and the construction progress prediction result.

[0064] A minimum resource threshold and a maximum resource threshold can be set for each construction resource, and the construction resources are adjusted between the minimum resource threshold and the maximum resource threshold of each construction resource, so that the construction progress prediction result meets the expected progress requirement input by the staff, thereby realizing the allocation of the construction resources of the city construction project. In the adjustment process, the multi-strategy particle swarm optimization algorithm described in the embodiment of the application can be used to encode and optimize the construction resources, to find the construction resource combination that minimizes the difference between the expected progress requirement and the construction progress prediction result, thereby realizing the allocation of the construction resources of the city construction project. The expected progress requirement input by the staff can be a specific value or a range. In the case of a range, the construction resource combination that minimizes the difference between the median of the expected progress requirement and the construction progress prediction result can be found. However, it should be noted that the construction resources are usually integers, so in the process of encoding and optimizing the construction resources using the multi-strategy particle swarm optimization algorithm, the particle needs to be rounded up in each optimization.

[0065] In some possible implementation manners, the construction progress prediction model is constructed, and the multi-strategy particle swarm optimization algorithm is used to optimize the construction progress prediction model, to obtain the optimized construction progress prediction model, including:

[0066] S12.1, adopt a convolutional neural network to construct a construction progress prediction model, and obtain historical multi-source heterogeneous data input by a worker or pre-stored and real progress corresponding to the historical multi-source heterogeneous data.

[0067] Optionally, in addition to adopting a convolutional neural network to construct a construction progress prediction model, other neural networks can also be adopted to construct a construction progress prediction model.

[0068] S12.2, sample data is constructed with the historical multi-source heterogeneous data, and sample labels are constructed with real progress corresponding to the historical multi-source heterogeneous data, to obtain training data.

[0069] S12.3, initialize a particle swarm based on hyperparameters of the construction progress prediction model, wherein the particle swarm includes a plurality of different particles, each particle includes all training hyperparameters of the construction progress prediction model, and the training hyperparameters can be all or part of the hyperparameters of the construction progress prediction model and can be defined by the worker.

[0070] Illustratively, initializing the particle swarm based on the hyperparameters of the construction progress prediction model includes: randomly initializing between upper and lower limits corresponding to the hyperparameters of the construction progress prediction model, and forming a vector with the hyperparameters after random initialization to obtain a particle, and obtaining the particle swarm after repeating the initialization multiple times.

[0071] S12.4, obtain an error function value corresponding to each particle based on the training data, and determine a particle with the minimum error function value as an optimal particle.

[0072] Illustratively, obtaining the error function value corresponding to each particle based on the training data includes: taking sample data in the training data as input data, taking sample labels in the training data as expected output data, and adopting a root mean square error function to obtain the error function value corresponding to each particle.

[0073] S12.5, adopt a golden sine chain search strategy to perform initial search on particles in the particle swarm based on the optimal particle, to obtain particles after initial search.

[0074] Illustratively, adopting the golden sine chain search strategy to perform initial search on the particles in the particle swarm based on the optimal particle to obtain the particles after initial search includes:

[0075] arranging the particles in the particle swarm in descending order of error function values to obtain arranged particles;

[0076] for any arranged particle, matching the particle with another particle according to the arrangement number of the particle to obtain a paired particle corresponding to each particle;

[0077] Assuming the total number of particles in the particle swarm is N, then for the th particle after permutation... j The nth particle, which is the nth particle after matching and arranging. j +1 particle as a pairing particle j =1,2,…,N, and because j When =N, there is no first N. j +1 particle, therefore for the th after the arrangement N Each particle is randomly matched with another particle as its pairing particle.

[0078] Based on the optimal particle and the paired particle, an initial search is performed on the particles in the particle swarm, resulting in the following particles after the initial search:

[0079]

[0080]

[0081] in, For the first t During the second optimization process, the first j The particles after the arrangement The d dimensional hyperparameters, d =1,2,…,D, where D is the total number of hyperparameters in the particle. For particles The d The hyperparameters in the first dimension t The optimization speed of the next optimization process For the first j The particle after the initial search d dimensional hyperparameters, The first random number between (0,1) The second random number between (0,1) It is a sine function. For particles The first pair of particles d dimensional hyperparameters, The third random number between (0,1) = -π+(1-τ)*2τ, = -π+τ*2τ, As the first search coefficient, Let be the second search coefficient, π be pi, and τ be a constant term. , The first fixed learning factor can be set to 0.7, 1, or 1.5; The optimal particle's first d Dimensional hyperparameters.

[0082] The Golden Sine Chain Search strategy leverages the idea of ​​the Golden Sine Search, enabling the algorithm to search along a Golden Sine path as it searches for a better position. Simultaneously, all particles engage in a chain-like joint search, which can improve the ability to find better solutions while ensuring local search accuracy.

[0083] S12.6. Based on the optimal particle, a particle swarm information fusion search strategy is used to fuse the information of the particles after the initial search to obtain the fused particles.

[0084] For example, based on the optimal particle, a particle swarm information fusion search strategy is used to fuse the information of the particles after the initial search to obtain the fused particles, including:

[0085] For any particle after the initial search, determine the Euclidean distance between the particle after the initial search and other particles.

[0086] Based on the Euclidean distance between the particle and other particles after the initial search, the particle information learning scale parameter is obtained as follows:

[0087]

[0088] in, For the first m The particle after the initial search and the first n The particle information learning scale parameter between particles after the initial search. This parameter controls the learning range and can be set to 0.5. The learning intensity control parameter can be set to 1.5; e Represents the natural constant. For the first m The particle after the initial search and the first n The Euclidean distance between particles after the initial search.

[0089] For any particle after the initial search, based on the particle information learning scale parameter, the particle swarm fusion information corresponding to the particle after the initial search is obtained as follows:

[0090]

[0091] in, For the first t +1 optimization process, the particle's first... d Particle swarm fusion information corresponding to the 3D hyperparameters For information fusion parameters, and , This is the maximum value for the information fusion parameter, and can be set to 1. The minimum value of the information fusion parameter can be set to 0.0001; T is the maximum number of optimization iterations. For the first d The upper limit of the hyperparameters, For the first d The lower bound of the hyperparameter. For the first t During the second optimization process, the first m The particle after the initial search d dimensional hyperparameters, For the first t During the second optimization process, the first n The particle after the initial search d Dimensional hyperparameters.

[0092] Based on the particle swarm fusion information corresponding to the optimal particle and the particles after the initial search, information fusion is performed on the particles after the initial search to obtain the fused particles:

[0093]

[0094] in, For the first n The first particle after information fusion d Dimensional hyperparameters.

[0095] The particle swarm information fusion search strategy, under the premise of the algorithm, performs a large-scale solution space search by learning information from different particles, and then performs a local search based on the optimal particle. This can effectively avoid getting trapped in local optima and improve the search capability in the unknown domain of the solution space.

[0096] S12.7. Adaptive global mutation strategy is used to perform adaptive global search on the particles after information fusion to obtain particles after adaptive global search.

[0097] For example, an adaptive global mutation strategy is used to perform an adaptive global search on the particles after information fusion to obtain particles after the adaptive global search, including:

[0098] The global mutation factor is obtained as follows:

[0099]

[0100] in, For the first t Global variation factor in the sub-optimization process. For the first ta global mutation factor in the first optimization process, and in the first optimization process, the global mutation factor is randomly generated between (0, 1), is a remainder function, is a circular constant;

[0101] On the basis of the current optimization number, the adaptive inertia weight, the first adaptive learning factor and the second adaptive learning factor are obtained as:

[0102]

[0103]

[0104]

[0105] wherein, is an adaptive inertia weight, is a maximum value of the inertia weight, which can be set to 1; is a minimum value of the inertia weight, which can be set to 0.001; is a first adaptive learning factor, is a maximum value of the first adaptive learning factor, which can be set to 1; is a minimum value of the first adaptive learning factor, which can be set to 0.001; is a second adaptive learning factor, is a maximum value of the second adaptive learning factor, is a minimum value of the second adaptive learning factor;

[0106] According to the global mutation factor, the adaptive inertia weight, the first adaptive learning factor and the second adaptive learning factor, the search speed of the particle after information fusion in the current optimization process is obtained as:

[0107]

[0108] wherein, is the first optimization process, t is the first optimization process, k is the first optimization process, is the first optimization process, d is the first optimization process, is the first optimization process, is the first optimization process, d is the first optimization process, t is the first optimization process, is the first optimization process, is the first optimization process, d is the first optimization process, t is the first optimization process, is a fourth random number between (0, 1), the fifth random number between (0, 1), the first d dimensional hyperparameter of the fifth particle.

[0109] According to the search speed of the particle after information fusion in the current optimization process, the adaptive global search is performed on the particle after information fusion, and the particle after adaptive global search is obtained.

[0110]

[0111] wherein, the first k dimensional hyperparameter of the fifth particle. d dimensional hyperparameter of the fifth particle.

[0112] The adaptive inertia weight and the second adaptive learning factor of the adaptive global mutation strategy gradually decrease from 1 to 0.001 with the increase of the optimization number; the first adaptive learning factor gradually increases from 0.001 to 1 with the increase of the optimization number; so that the particle can have stronger global search ability at the beginning of the algorithm, and the search process is in a chaotic state, which is more conducive to jumping out of the local optimal solution, and the algorithm gradually learns only the information of the optimal particle in the later stage, thereby improving the convergence accuracy of the algorithm.

[0113] Optionally, the search process of the adaptive global mutation strategy can be controlled by using the simulated annealing algorithm, so that the algorithm can be guaranteed to a certain extent that the search stagnation phenomenon will not occur in the global search process, thereby improving the robustness of the algorithm.

[0114] S12.8, the golden sine chain search strategy, the particle swarm information fusion search strategy and the adaptive global mutation strategy are repeatedly executed until the optimization end condition is met, and the optimal particle is reacquired.

[0115] The optimization end condition can be set as: the optimization ends when the current optimization number is greater than or equal to the preset maximum optimization number.

[0116] S12.9, the hyperparameters in the reacquired optimal particle are taken as the final hyperparameters of the construction progress prediction model, and the optimized construction progress prediction model is obtained.

[0117] The embodiments of the present application provide a multi-strategy particle swarm optimization algorithm, which optimizes the construction progress prediction model by using the golden sine chain search strategy, the particle swarm information fusion search strategy and the adaptive global mutation strategy, thereby improving the optimization ability of the construction progress prediction model, so that the optimized construction progress prediction model can improve the prediction accuracy of the construction progress under the consideration of multi-source heterogeneous data.

[0118] In some possible implementation manners, the construction progress prediction result is taken as a smart city construction analysis result, and the smart city construction analysis result is transmitted to a device designated by a staff member, including:

[0119] The construction progress prediction result is taken as a smart city construction analysis result.

[0120] The smart city construction analysis result and corresponding multi-source heterogeneous data are encrypted by using a symmetric encryption algorithm to obtain target transmission data carrying the smart city construction analysis result.

[0121] The target transmission data carrying the smart city construction analysis result is transmitted to the device designated by the staff member.

[0122] The smart city construction analysis result and corresponding multi-source heterogeneous data are encrypted by using a symmetric encryption algorithm to obtain target transmission data carrying the smart city construction analysis result, so that more comprehensive decision-making data is provided for the staff member. In the encryption process, the public key corresponding to the device designated by the staff member can be used for encryption, so that the data of the city construction project cannot be leaked.

[0123] Each of the embodiments in the specification is described in a progressive manner, and each embodiment focuses on the difference from other embodiments. The same or similar parts of each embodiment can be referred to each other.

[0124] The embodiments of the present application are described with reference to the flowcharts and / or block diagrams of the methods, devices, electronic devices and computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of the flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing terminal device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal device produce a means for implementing the functions specified in the flowcharts and / or block diagrams. Figure 1 The functions specified in one flow or multiple flows and / or blocks Figure 1 The device for implementing the functions specified in one block or multiple blocks.

[0125] These computer program instructions can also be stored in a computer readable memory that can guide the computer or other programmable data processing terminal device to work in a specific way, so that the instructions stored in the computer readable memory produce a product including instruction devices, which implement the functions specified in the flowcharts and / or block diagrams. Figure 1 The functions specified in one flow or multiple flows and / or blocks Figure 1 The device for implementing the functions specified in one block or multiple blocks.

[0126] These computer program instructions can also be loaded into a computer or other programmable data processing terminal device, so that a series of operational steps are performed on the computer or other programmable terminal device to generate a computer implemented process, so that the instructions executed on the computer or other programmable terminal device provide a process for implementing the functions specified in the flowchart block or blocks. Figure 1 These computer program instructions can also be loaded into a computer or other programmable data processing terminal device, so that a series of operational steps are performed on the computer or other programmable terminal device to generate a computer implemented process, so that the instructions executed on the computer or other programmable terminal device provide a process for implementing the functions specified in the flowchart block or blocks. Figure 1 These computer program instructions can also be loaded into a computer or other programmable data processing terminal device, so that a series of operational steps are performed on the computer or other programmable terminal device to generate a computer implemented process, so that the instructions executed on the computer or other programmable terminal device provide a process for implementing the functions specified in the flowchart block or blocks.

[0127] Although the preferred embodiments of the application have been described, those skilled in the art will be able to make additional changes and modifications to these embodiments once they grasp the basic inventive concepts. Therefore, the appended claims are intended to cover all such changes and modifications that fall within the scope of the embodiments of the application.

[0128] Finally, it should be noted that, in the present document, the terms such as first and second, and the like, are used merely to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between such entities or operations. Moreover, the terms "comprising", "including", or any other variant thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements does not include only those elements in the list, but can also include other elements not expressly listed or inherent to such process, method, article, or terminal device. Without further limitation, an element preceded by "comprises a" does not, without more constraints, foreclose the existence of additional identical elements in the process, method, article, or terminal device that includes the element.

[0129] The principles and implementation modes of the present application are described herein by using specific examples, and the above descriptions of the examples are only used to help understand the method of the present application and its core idea; meanwhile, for those skilled in the art, according to the idea of the present application, the specific implementation modes and application ranges will have changes, and in view of the above, the content of the present description should not be understood as a limitation of the present application.

Claims

1.A smart city construction analysis method of multi-source data fusion, characterized in that, The method comprises the following steps: acquiring multi-source heterogeneous data of a city construction project; wherein the multi-source heterogeneous data comprises multiple construction progress influence data; constructing a construction progress prediction model and optimizing the construction progress prediction model by using a multi-strategy particle swarm optimization algorithm to obtain an optimized construction progress prediction model, comprising: constructing a construction progress prediction model by using a convolutional neural network and acquiring historical multi-source heterogeneous data input by a staff member or pre-stored and real progress corresponding to the historical multi-source heterogeneous data; constructing sample data by using the historical multi-source heterogeneous data and sample labels by using the real progress corresponding to the historical multi-source heterogeneous data to obtain training data; initializing a particle swarm on the basis of hyperparameters of the construction progress prediction model; wherein the particle swarm comprises multiple different particles, and each particle comprises all to-be-trained hyperparameters of the construction progress prediction model; acquiring an error function value corresponding to each particle on the basis of the training data and determining a particle with the minimum error function value as an optimal particle; initially searching the particles in the particle swarm on the basis of the optimal particle by using a golden sine chain search strategy, comprising: arranging the particles in the particle swarm in descending order of error function values to obtain arranged particles; matching each arranged particle with another particle according to the arrangement number of the arranged particle to obtain a paired particle corresponding to each particle; and initially searching the particles in the particle swarm on the basis of the optimal particle and the paired particle to obtain particles after initial searching; fusing information of the particles after initial searching on the basis of the optimal particle by using a particle swarm information fusion search strategy to obtain particles after information fusion; performing adaptive global search on the particles after information fusion by using an adaptive global mutation strategy to obtain particles after adaptive global search; repeating the golden sine chain search strategy, the particle swarm information fusion search strategy, and the adaptive global mutation strategy until an optimization end condition is met, and reacquiring an optimal particle; taking the hyperparameters in the reacquired optimal particle as final hyperparameters of the construction progress prediction model to obtain the optimized construction progress prediction model; predicting the multi-source heterogeneous data of the city construction project by using the optimized construction progress prediction model to determine a construction progress prediction result; taking the construction progress prediction result as a smart city construction analysis result and transmitting the smart city construction analysis result to a device designated by a staff member. 2.The method of claim 1, wherein, After taking the construction progress prediction result as the smart city construction analysis result, the method further comprises the following steps: acquiring an expected progress requirement input by a staff member and allocating construction resources of the city construction project according to the expected progress requirement and the construction progress prediction result. 3.The method of claim 1, wherein, The multiple construction progress influence data comprises city construction project basic data, environmental influence data, construction technology influence data, and construction resource influence data. 4.The method of claim 1, wherein, Initializing the particle swarm on the basis of the hyperparameters of the construction progress prediction model comprises the following steps: Randomly initialize between upper and lower limits corresponding to the hyperparameters of the construction progress prediction model, and form a vector of the hyperparameters after random initialization to obtain a particle. After multiple repeated initializations, a particle swarm is obtained. 5.The method of claim 1, wherein, On the basis of the training data, an error function value corresponding to each particle is obtained, including: Taking the sample data in the training data as input data and the sample label in the training data as expected output data, the root mean square error function is used to obtain the error function value corresponding to each particle. 6.The smart city construction analysis method of multi-source data fusion according to claim 1, characterized in that, On the basis of the optimal particle, a particle swarm information fusion search strategy is used to fuse the particles after the initial search to obtain particles after information fusion, including: For any one particle after the initial search, the Euclidean distance between the particle after the initial search and other particles is determined. According to the Euclidean distance between the particle after the initial search and other particles, a particle information learning scale parameter is obtained. For any one particle after the initial search, according to the particle information learning scale parameter, the particle swarm fusion information corresponding to the particle after the initial search is obtained. According to the optimal particle and the particle swarm fusion information corresponding to the particle after the initial search, the particle after the initial search is fused to obtain the particle after information fusion. 7.The method of claim 6, wherein, An adaptive global search strategy is used to perform adaptive global search on the particle after information fusion to obtain a particle after adaptive global search, including: A global mutation factor is obtained. On the basis of the current optimization times, an adaptive inertia weight, a first adaptive learning factor and a second adaptive learning factor are obtained. According to the global mutation factor, adaptive inertia weight, first adaptive learning factor and second adaptive learning factor, the search speed of the particle after information fusion in the current optimization process is obtained. According to the search speed of the particle after information fusion in the current optimization process, the adaptive global search is performed on the particle after information fusion to obtain the particle after adaptive global search. 8.The method of claim 1, wherein, The construction progress prediction result is used as the smart city construction analysis result, and the smart city construction analysis result is transmitted to the device designated by the staff, including: The construction progress prediction result is used as the smart city construction analysis result. The symmetric encryption algorithm is used to encrypt the smart city construction analysis result and the corresponding multi-source heterogeneous data to obtain target transmission data carrying the smart city construction analysis result. The target transmission data carrying the smart city construction analysis result is transmitted to the device designated by the staff.

Images