Network point site selection method and device, electronic equipment and storage medium

Abstract

Embodiments of the present application disclose a network point site selection method and device, electronic equipment and storage medium, which belong to the technical field of computers, and the method comprises: obtaining a prediction data set, the prediction data set comprising pre-site selection feature data; inputting the pre-site selection feature data into at least two target agent models to obtain at least two prediction output results, the target agent model being obtained by training site selection feature data and target data in a training data set, the target data being obtained based on the site selection feature data; iteratively processing the at least two prediction output results based on a preset evolutionary algorithm to obtain an iteration result, and obtaining an optimal site selection of a network point when the iteration result meets a preset convergence condition. The embodiments of the present application improve the decision accuracy, reduce the calculation cost and simplify the calculation process by training the agent model and using the evolutionary algorithm to determine the optimal site selection address of the network point.

Description

Technical Field

[0001] The embodiments of the present invention relate to computer technology, and more particularly to a site selection method, apparatus, electronic device and storage medium. Background Technology

[0002] Offline branches play an indispensable role in various business sectors, such as bank branches and electronic equipment sales outlets. Determining the optimal location for a branch within a given area to serve users is a key consideration.

[0003] One existing solution provides a convolution-based self-service bank location selection method, which is a single-objective optimization method. The goal is to minimize the layout cost of self-service banks. In terms of technical implementation, the convolution function is used to simulate the superposition effect of self-service bank outlets in meeting demand. The model uses a heuristic algorithm to solve for the minimum cost value required for the layout of self-service bank outlets and the corresponding number of outlets, number of self-service machines and their locations while satisfying the constraints.

[0004] The above scheme is implemented using only a simple genetic algorithm optimization framework. In complex real-world application scenarios, it may suffer from poor decision-making accuracy and low precision. Furthermore, the scheme uses a two-dimensional convolution function to simulate and quantify the target value, which is cumbersome and complex, and has a high computational cost. Summary of the Invention

[0005] This invention provides a site selection method, apparatus, electronic device, and storage medium that can improve existing site selection schemes.

[0006] In a first aspect, embodiments of the present invention provide a site selection method, including:

[0007] Obtain a prediction dataset, which includes pre-located feature data;

[0008] The pre-location feature data is input into at least two target proxy models to obtain at least two prediction output results. The target proxy models are obtained by training on the location feature data and target data in the training dataset. The target data is obtained based on the location feature data.

[0009] Based on a preset evolution algorithm, at least two of the predicted output results are iterated to obtain the iterative results. When the iterative results satisfy the preset convergence condition, the optimal location of the network point is obtained.

[0010] Optionally, the target proxy model is obtained by training the location feature data and target data in the training dataset in the following manner:

[0011] Using the location feature data as input and the target data as output, a preset proxy model is trained to obtain the target proxy model.

[0012] Optionally, the location feature data includes at least one subset of location features, and the target data includes at least one subset of targets;

[0013] The step of using the location feature data as input and the target data as output to train a preset proxy model to obtain the target proxy model includes:

[0014] At least one of the location feature subsets and at least one of the target subsets are cross-matched to obtain multiple training data pairs;

[0015] The preset proxy model is trained based on the location feature subset and the target subset in each training data pair to obtain multiple target proxy models.

[0016] Optionally, the step of iterating over at least two predicted outputs based on a preset evolutionary algorithm to obtain iterative results includes:

[0017] Based on a preset evolution algorithm, at least two of the predicted output results are mapped to obtain a mapping result;

[0018] The mapping result is then input again into at least two of the target agent models for iteration to obtain an iteration result.

[0019] Optionally, obtaining the optimal address of the network point when the iteration result satisfies a preset convergence condition includes:

[0020] The gradient change value is obtained by comparing the current iteration result with the previous iteration result;

[0021] When the gradient change value is less than a first preset value, the iteration result satisfies the preset convergence condition, and the optimal address of the network point is obtained based on the current iteration result.

[0022] Optionally, obtaining the optimal address of the network point when the iteration result satisfies a preset convergence condition includes:

[0023] Obtain the number of iterations required to input the mapping result into at least two of the target agent models for iteration;

[0024] When the number of iterations meets the second preset value, the iteration result meets the preset convergence condition, and then the optimal address of the network point is obtained based on the current iteration result.

[0025] Optionally, the current iteration result is the optimal prediction target for the prediction output result;

[0026] The step of obtaining the optimal address of the network point based on the current iteration result includes:

[0027] Obtain the target input data for the optimal predicted target, wherein the target input data is the optimal location of the network point.

[0028] Secondly, embodiments of the present invention provide a site selection device, the device comprising:

[0029] The data acquisition module is used to acquire the prediction dataset, which includes pre-located feature data;

[0030] The result output module is used to input the pre-location feature data into at least two target proxy models to obtain at least two prediction output results. The target proxy models are obtained by training on the location feature data and target data in the training dataset. The target data is obtained based on the location feature data.

[0031] The location acquisition module is used to iterate on at least two of the predicted output results based on a preset evolution algorithm to obtain the iteration result, and when the iteration result satisfies the preset convergence condition, obtain the optimal location of the network point.

[0032] Thirdly, embodiments of the present invention also provide an electronic device, the electronic device comprising:

[0033] At least one processor; and

[0034] A memory communicatively connected to the at least one processor; wherein,

[0035] The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the site selection method according to any embodiment of the present invention.

[0036] Fourthly, embodiments of the present invention also provide a computer-readable storage medium storing computer instructions, which are used to cause a processor to execute the site selection method described in any embodiment of the present invention.

[0037] The site selection scheme of this invention first inputs pre-selection feature data from the prediction dataset into at least two target surrogate models to obtain at least two prediction outputs. The target surrogate models are obtained by training on the site selection feature data and target data from the training dataset. The target data is obtained based on the site selection feature data; specifically, the target data consists of operational benefit data related to site selection, ensuring that the obtained target surrogate models better meet the scenario requirements. Finally, the at least two prediction outputs are iterated using a preset evolutionary algorithm to obtain iterative results. When the iterative results satisfy a preset convergence condition, the optimal site location is obtained. This embodiment improves decision accuracy by using a preset evolutionary algorithm for multiple iterations. By training the surrogate models and using an evolutionary algorithm to determine the optimal site location, it achieves the beneficial effects of improving decision accuracy, reducing computational costs, and simplifying the calculation process.

[0038] It should be understood that the description in this section is not intended to identify key or important features of the embodiments of the present invention, nor is it intended to limit the scope of the present invention. Other features of the embodiments of the present invention will become readily apparent from the following description. Attached Figure Description

[0039] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0040] Figure 1 This is a flowchart illustrating the site selection method provided in an embodiment of the present invention;

[0041] Figure 2 This is another flowchart illustrating the site selection method provided in this embodiment of the invention;

[0042] Figure 3 This is a schematic diagram of a site selection device provided in an embodiment of the present invention;

[0043] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0044] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0045] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, the accompanying drawings show only the parts relevant to the present invention, and not all of the structures.

[0046] Figure 1 This is a flowchart illustrating a site selection method provided in an embodiment of the present invention. This embodiment is applicable to the selection of site locations during the layout of offline sites. The method can be executed by a site selection device, which can be implemented in hardware and / or software and can be configured in computer equipment such as a server. (Reference) Figure 1 The method may specifically include the following steps:

[0047] S110. Obtain the prediction dataset.

[0048] The data in the prediction dataset is used as input to the trained model to obtain test results, which are then used to evaluate the model's performance. In this embodiment, the site selection scheme is used to lay out offline sites for various business areas, and the prediction dataset includes pre-selection feature data.

[0049] The pre-selection site feature data can be characteristic data about multiple leasable shop locations selected within a large area where outlets need to be built. For example, it can include: outlet coordinates, area traffic, rental costs, and customer ratings for each shop location. The customer rating can be a rating given by the company based on the frequency of user visits to the outlets and the business transactions conducted, reflecting customer traffic in the current area. The specific content of the pre-selection site feature data is not limited to this example and should be based on the actual needs of the business domain.

[0050] S120. Input the pre-selected location feature data into at least two target proxy models to obtain at least two prediction output results.

[0051] The aforementioned target proxy model was obtained by training on the location feature data and target data in the training dataset.

[0052] Before training the model, the operational objectives for site selection and the key characteristic factors influencing these objectives are first determined based on application requirements. The key characteristic factors influencing the operational objectives are the site selection characteristic data mentioned above, and the operational objectives are the target data mentioned above.

[0053] In the branch location selection scheme provided in this embodiment of the invention, the location feature data in the training dataset and the pre-location feature data in the prediction dataset are of the same type, only their content differs. Taking the banking sector as an example, all feature data of a bank's existing branches are collected from legal and compliant channels. After preprocessing the feature data, a total dataset is obtained. 60% of the total dataset can be used as the training dataset for model training, and 40% can be used as the prediction dataset for evaluating model performance, etc. The specific method of dividing the training and prediction datasets is not limited here.

[0054] The aforementioned preprocessing operations for feature data may include data deduplication, data validation, and data cleaning, so that all data in the obtained total dataset can reflect the corresponding data characteristics. Specific preprocessing methods are not limited here.

[0055] The target data is obtained based on location characteristic data, meaning that maximizing the revenue of the current branch is the objective, taking into account the current location characteristic data. For example, the target data may include operating revenue and digital transformation evaluation indicators. The digital transformation evaluation indicators can be understood as the number of depositors in the current branch and the status of depositor transactions. The purpose of including branch digital transformation evaluation indicators in the target data is to actively adapt to the current wave of digitalization and intelligentization in banking, closely align with real-world application scenarios, and thus ensure higher reliability of the final model output. The specific content of the target data is not limited here.

[0056] When training a target surrogate model using location feature data and target data, the location feature data is used as input and the target data is used as output to perform the model training process.

[0057] The number of target proxy models includes at least two. For example, multiple target proxy models can be obtained by arbitrarily combining the determined location feature data with the target data. Correspondingly, when the pre-location feature data is input into multiple target proxy models, multiple output results will be obtained. The purpose of this is to improve the accuracy of the output results through transfer optimization in subsequent steps. The specific number of target proxy models is not limited here.

[0058] S130. Iterate over at least two predicted outputs based on a preset evolution algorithm to obtain the iteration results. When the iteration results meet the preset convergence conditions, obtain the optimal location of the network points.

[0059] In the current step, the method of iterating over at least two predicted outputs based on a preset evolutionary algorithm can be as follows: First, iterate over at least two predicted outputs to obtain an iterative result for at least two predicted outputs. The current iterative result is the optimization result obtained based on the preset evolutionary algorithm. If the current iterative result does not meet the convergence condition, the current iterative result can be repeatedly input into the above at least two target proxy models to iterate repeatedly based on the preset evolutionary algorithm until the above convergence condition is met. The final iterative result obtained is the optimal location of the network point.

[0060] The aforementioned preset evolutionary algorithm can be a multifactorial evolutionary algorithm (MFEA), a multi-objective evolutionary algorithm (MOEA), a neural evolutionary algorithm, or a differential evolution algorithm (DE), etc. The specific evolutionary algorithm is not limited here.

[0061] The above-mentioned methods for determining whether the iteration result meets the preset convergence conditions can be as follows: when the number of iterations meets the preset value, the final output iteration result can be used as the optimal address of the network point; or when the gradient value between the current iteration result and the previous iteration result meets the preset requirements, it indicates that the current iteration result meets the convergence conditions, and the current iteration result can be used as the optimal address of the network point. The specific methods for determining whether the iteration result meets the preset convergence conditions are not limited here.

[0062] The site selection method provided in this embodiment of the invention first inputs pre-selection feature data from the prediction dataset into at least two target surrogate models to obtain at least two prediction outputs. The target surrogate models are obtained by training on the site selection feature data and target data from the training dataset. The target data is obtained based on the site selection feature data; specifically, the target data consists of operational benefit data related to site selection, ensuring that the obtained target surrogate models better meet the scenario requirements. Finally, the at least two prediction outputs are iterated using a preset evolutionary algorithm to obtain iterative results. When the iterative results satisfy a preset convergence condition, the optimal site location is obtained. This embodiment improves decision accuracy by using a preset evolutionary algorithm for multiple iterations. By training the surrogate models and using an evolutionary algorithm to determine the optimal site location, the method achieves the beneficial effects of improving decision accuracy, reducing computational costs, and simplifying the computation process.

[0063] Figure 2 This is another schematic flowchart of the site selection method provided in this embodiment of the invention. The relationship between this embodiment and the above embodiments further refines the corresponding features of the above embodiments. Figure 2 As shown, the method may include the following steps:

[0064] S210. Obtain the prediction dataset.

[0065] The prediction dataset includes pre-located feature data.

[0066] S220. Input the pre-selected location feature data into at least two target proxy models to obtain at least two prediction output results.

[0067] The target surrogate model is obtained by training on the location feature data and target data in the training dataset. The target data is obtained based on the location feature data.

[0068] The site selection scheme provided in this embodiment can obtain a target proxy model by using site selection feature data as input and target data as output to train a preset proxy model.

[0069] The aforementioned pre-defined proxy model can be implemented using a Kriging model. By using a Kriging model as the proxy model, the location feature data and the target data are used as the input and output of the model, respectively. After multiple rounds of training on the training dataset, the target proxy model is obtained.

[0070] In a preferred embodiment, the site selection scheme provided in this embodiment includes site selection feature data comprising at least one subset of site selection features, and target data comprising at least one subset of target features. For example, taking site selection feature data including: site coordinates L, rental cost C, regional traffic T, and regional customer comprehensive rating R; and target data including: operating revenue PR and digital transformation evaluation index DG, each data point can represent a subset. Therefore, in this example, the site selection feature data can contain 4 subsets of site selection features, and the target data can contain 2 subsets of target features. The specific number of subsets of site selection features included in the site selection feature data and the number of subsets of target features included in the target data are not limited here, and are determined according to actual business needs.

[0071] Correspondingly, using location feature data as input and target data as output, training a preset proxy model and obtaining a target proxy model can be achieved through the following process: cross-matching at least one subset of location features and at least one subset of targets to obtain multiple training data pairs; training a preset proxy model based on the subset of location features and the subset of targets in each training data pair to obtain multiple target proxy models.

[0072] If the feature subset includes: site coordinates L, rental cost C, regional traffic T, and regional customer comprehensive rating R; and the target subset includes: operating revenue PR and digital transformation evaluation index DG, then the way to obtain multiple training data pairs can be: each feature subset and each target subset can be cross-matched to obtain 8 training data pairs. The location feature subset and target subset contained in each training data pair are used to train a Kriging model, and 8 target proxy models can be obtained accordingly.

[0073] It should be noted that when training multiple target proxy models, only a preset number of training data pairs can be selected for training to obtain a corresponding number of target proxy models. The purpose of doing so is that if the number of models is small, the output results may not be representative; if all feature subsets and target subsets are cross-matched, the computational load is large, the model calculation method is complex, and a large number of similar models are obtained.

[0074] For example, taking the acquisition of four target agent models as an example, training data pair P1 includes network location coordinates L and operating revenue PR; training data pair P2 includes rental costs C and operating revenue PR; training data pair P3 includes regional traffic T and digital transformation evaluation index DG; and training data pair P4 includes regional customer comprehensive rating R and digital transformation evaluation index DG, etc. The specific number of target agent models and the combination of training data pairs are not limited here, and should be based on the actual needs of the developers.

[0075] S230. Based on a preset evolution algorithm, at least two predicted output results are mapped to obtain a mapping result.

[0076] The pre-defined evolutionary algorithm performs mapping operations such as unified representation, homogeneous crossover, and vertical cultural transmission on at least two predicted outputs to obtain a mapping result. The current process can be understood as extracting data with significant similarity from the at least two predicted outputs obtained in step S220 to obtain a mapping result. This mapping result can then be input into the target proxy model in subsequent steps. The advantages of this approach are that it reduces the amount of data input in subsequent iterations, allows for parameter tuning based on simple machine learning algorithms, reduces computational cost, and improves computational efficiency.

[0077] S231. Input a mapping result into at least two target agent models for iteration to obtain an iteration result.

[0078] In the current step, the obtained mapping result is input into at least two target agent models to obtain at least two output results for the first mapping result. The operation of iterating over at least two output results of the first mapping result based on the preset evolution algorithm is repeated to obtain an iterative result.

[0079] In one implementation, step S231, when the iteration result meets the preset convergence condition, the optimal address of the network point can be obtained by comparing the current iteration result with the previous iteration result to obtain the gradient change value; when the gradient change value is less than the first preset value, the iteration result meets the preset convergence condition, and then the optimal address of the network point is obtained based on the current iteration result.

[0080] Taking the output of each iteration result as a matrix as an example, the gradient change value can be obtained by comparing the current iteration result with the previous iteration result. The gradient change value can be obtained by calculating the mean squared difference of the values ​​in the current matrix and the previous matrix. If the value of the gradient change value is less than the first preset value, it indicates that the iteration result meets the first convergence condition.

[0081] The first preset value can be 0.2, 0.1 or 0.01, etc. The specific selection of the first preset value is not limited here, but is based on the accuracy requirements of the developers.

[0082] Another implementation method is as follows: In step S231, when the iteration result meets the preset convergence condition, the optimal address of the network point can also be obtained by: obtaining the number of iterations to input the mapping result into at least two target proxy models for iteration; when the number of iterations meets the second preset value, the iteration result meets the preset convergence condition, and then the optimal address of the network point is obtained according to the current iteration result.

[0083] In the current method, whether the iteration result meets the preset convergence condition can be determined based on the number of iterations. The current second preset value can be obtained by the developers based on their experience and needs. For example, it can be 50 times, 50 times, or 100 times, etc. The specific selection of the second value is not limited here and is determined by the developers' experience.

[0084] S240. Obtain the target input data for the optimal prediction target. The target input data is the optimal location of the network point.

[0085] When the iteration result satisfies the preset convergence condition, the optimal address of the network points is further obtained based on the current iteration result. The current iteration result is then the optimal prediction target for the predicted output. Here, the current optimal prediction target can be understood as the output result of the trained target proxy model.

[0086] Since the output results are target data, indicating evaluation indicators such as operating income and digital transformation assessment indicators, and do not include the location of the outlets, it is necessary to obtain the target input data for the current prediction target. The current target input data is the optimal location of the outlets.

[0087] The branch location selection method provided in this embodiment of the invention is implemented by training a proxy model and using an evolutionary algorithm to determine the optimal location of the branch, which can reduce the computational difficulty. At the same time, this embodiment takes into account the operational efficiency of the branch and incorporates digital indicators into the scope of consideration, actively adapting to the current wave of digitalization and intelligentization in banking, closely aligning with real-world application scenarios, and obtaining highly reliable branch location selection results. The solution provided in this embodiment achieves the beneficial effects of simplifying the calculation process and improving the accuracy of decision-making.

[0088] Figure 3 This is a schematic diagram of a site selection device provided in an embodiment of the present invention. This device is suitable for executing the site selection method provided in an embodiment of the present invention. Figure 3 As shown, the device may specifically include: a data acquisition module 310, a result output module 320, and an address selection module 330, wherein:

[0089] Data acquisition module 310 is used to acquire a prediction dataset, the prediction dataset including pre-location feature data;

[0090] The result output module 320 is used to input the pre-location feature data into at least two target proxy models to obtain at least two prediction output results. The target proxy models are obtained by training on the location feature data and target data in the training dataset. The target data is obtained based on the location feature data.

[0091] The location acquisition module 330 is used to iterate on at least two of the predicted output results based on a preset evolution algorithm to obtain the iteration result, and when the iteration result satisfies the preset convergence condition, obtain the optimal location of the network point.

[0092] The site selection device of this invention first inputs pre-selection feature data from the prediction dataset into a target surrogate model to obtain a prediction output result. The target surrogate model is obtained by training on the site selection feature data and target data from the training dataset. The target data is obtained based on the site selection feature data; specifically, the target data consists of operational benefit data related to site selection, ensuring the obtained target surrogate model better meets scenario requirements. Finally, the prediction output result is iterated based on a preset evolutionary algorithm to obtain an iteration result. When the iteration result satisfies a preset convergence condition, the optimal site location is obtained. This embodiment improves decision accuracy by using a preset evolutionary algorithm for multiple iterations. By training the surrogate model and using an evolutionary algorithm to determine the optimal site location, it achieves the beneficial effects of improving decision accuracy, reducing computational costs, and simplifying the calculation process.

[0093] In one embodiment, the target proxy model is obtained by training the location feature data and target data in the training dataset in the following manner: using the location feature data as input and the target data as output, a preset proxy model is trained to obtain the target proxy model.

[0094] In one embodiment, the result output module 320 is specifically used to perform cross-matching of at least one of the location feature subsets and at least one of the target subsets to obtain multiple training data pairs; and to train the preset proxy model based on the location feature subset and the target subset in each of the training data pairs to obtain multiple target proxy models.

[0095] In one embodiment, the address selection acquisition module 330 includes: a result mapping unit and a model iteration unit, wherein:

[0096] The result mapping unit is used to map at least two of the predicted output results based on a preset evolution algorithm to obtain a mapping result;

[0097] A model iteration unit is used to input one of the mapping results back into at least two of the target agent models for iteration to obtain one of the iteration results.

[0098] In one embodiment, the address selection acquisition module 330 further includes: a numerical comparison unit and an address selection acquisition unit, wherein:

[0099] The numerical comparison unit is used to compare the current iteration result with the previous iteration result to obtain the gradient change value;

[0100] The location acquisition unit is used to obtain the optimal location of the network point based on the current iteration result when the gradient change value is less than a first preset value and the iteration result satisfies the preset convergence condition.

[0101] In one embodiment, the address selection module 330 further includes: a count acquisition unit, wherein:

[0102] The iteration acquisition unit is used to acquire the number of iterations in which the mapping result is input into at least two target agent models for iteration;

[0103] The location acquisition unit is further configured to obtain the optimal location of the network point based on the current iteration result when the number of iterations satisfies the second preset value and the iteration result satisfies the preset convergence condition.

[0104] In one embodiment, the current iteration result is the optimal prediction target with respect to the prediction output result;

[0105] The location acquisition unit is specifically used to acquire target input data for obtaining the optimal predicted target, wherein the target input data is the optimal location of the network point.

[0106] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional modules is merely an example. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. The specific working process of the functional modules described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0107] This invention also provides an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to execute the site selection method according to any embodiment of this invention.

[0108] This invention also provides a computer-readable medium storing computer instructions that, when executed by a processor, implement the site selection method described in any embodiment of this invention.

[0109] The following is for reference. Figure 4 It shows a schematic diagram of the structure of a computer system 500 suitable for implementing an electronic device according to embodiments of the present invention. Figure 4The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of the present invention.

[0110] like Figure 4 As shown, the computer system 500 includes a central processing unit (CPU) 501, which can perform various appropriate actions and processes based on programs stored in read-only memory (ROM) 502 or programs loaded from storage section 508 into random access memory (RAM) 503. The RAM 503 also stores various programs and data required for the operation of the system 500. The CPU 501, ROM 502, and RAM 503 are interconnected via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0111] The following components are connected to I / O interface 505: an input section 506 including a keyboard, mouse, etc.; an output section 507 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 508 including a hard disk, etc.; and a communication section 509 including a network interface card such as a LAN card, modem, etc. The communication section 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to I / O interface 505 as needed. A removable medium 511, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on drive 510 as needed so that computer programs read from it can be installed into storage section 508 as needed.

[0112] In particular, according to the embodiments disclosed in this invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 509, and / or installed from removable medium 511. When the computer program is executed by central processing unit (CPU) 501, it performs the functions defined above in the system of this invention.

[0113] It should be noted that the computer-readable medium shown in this invention can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this invention, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0114] 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 various embodiments of the present invention. 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. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0115] The modules and / or units described in the embodiments of the present invention can be implemented in software or hardware. The described modules and / or units can also be located in a processor; for example, a processor can be described as including a data acquisition module, a result output module, and an address selection module. The names of these modules do not necessarily limit the module itself.

[0116] In another aspect, the present invention also provides a computer-readable medium, which may be included in the device described in the above embodiments; or it may exist independently and not assembled into the device. The computer-readable medium carries one or more programs, which, when executed by the device, cause the device to include: acquiring a prediction dataset, the prediction dataset including pre-location feature data; inputting the pre-location feature data into at least two target proxy models to obtain at least two prediction output results, the target proxy models being obtained by training on location feature data and target data in a training dataset, the target data being obtained based on the location feature data; iterating on the at least two prediction output results based on a preset evolutionary algorithm to obtain an iteration result, and obtaining the optimal location of the network point when the iteration result satisfies a preset convergence condition.

[0117] According to the technical solution of this invention, decision accuracy can be improved by using a preset evolutionary algorithm for multiple iterations. By training the surrogate model and using the evolutionary algorithm to determine the optimal address of the network point, the beneficial effects of improving decision accuracy, reducing computational cost, and simplifying the calculation process are achieved.

[0118] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can occur depending on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method of site selection for a mesh point, characterized by, The method comprises the following steps: obtaining a prediction data set, wherein the prediction data set comprises pre-site selection feature data, and the pre-site selection feature data comprises the coordinates of each shop location, the regional traffic, the rental cost and the customer rating; inputting the pre-site selection feature data into at least two target agent models to obtain at least two prediction output results, wherein the target agent models are obtained by training site selection feature data and target data in a training data set, the target data is obtained based on the site selection feature data, the data types of the pre-site selection feature data and the site selection feature data are the same, the target data comprises business income and a digital transformation evaluation index, and the digital transformation evaluation index comprises the number of depositors in the current site and the depositors' business handling situation; mapping the at least two prediction output results based on a preset evolution algorithm to obtain a mapping result; inputting the mapping result into the at least two target agent models again for iteration to obtain an iteration result; when the iteration result meets a preset convergence condition, obtaining target input data of an optimal prediction target, and the target input data is the optimal site selection of the site; wherein when the iteration result meets the preset convergence condition, the implementation of obtaining the optimal site selection of the site is that: comparing the current iteration result with the last iteration result to obtain a gradient change value; when the gradient change value is less than a first preset value, the iteration result meets the preset convergence condition, and then the optimal site selection of the site is obtained according to the current iteration result; or, obtaining the number of iterations of inputting the mapping result into the at least two target agent models for iteration; when the number of iterations meets a second preset value, the iteration result meets the preset convergence condition, and then the optimal site selection of the site is obtained according to the current iteration result; wherein the site selection feature data comprises at least one site selection feature subset, and the target data comprises at least one target subset; the target agent models are obtained by training the site selection feature data and the target data in the training data set in the following manner: cross-matching the at least one site selection feature subset and the at least one target subset to obtain a plurality of training data pairs; training a preset agent model based on the site selection feature subset and the target subset in each training data pair to obtain a plurality of target agent models.

2. A site selection apparatus for a mesh point, characterized by, The method comprises the following steps: a data acquisition module is configured to obtain a prediction data set, wherein the prediction data set comprises pre-site selection feature data, and the pre-site selection feature data comprises the coordinates of each shop location, the regional traffic, the rental cost and the customer rating; An result output module is configured to input the pre-site feature data into at least two target agent models to obtain at least two prediction output results, wherein the target agent models are obtained by training site feature data and target data in a training data set, the target data is obtained based on the site feature data, the data types of the pre-site feature data and the site feature data are the same, the target data includes business income and a digital transformation evaluation index, and the digital transformation evaluation index includes the number of depositors in a current site and a depositor business handling situation. A site obtaining module is configured to map the at least two prediction output results based on a preset evolution algorithm to obtain a mapping result, input the mapping result into the at least two target agent models again to perform iteration to obtain an iteration result, and obtain target input data of an optimal prediction target when the iteration result meets a preset convergence condition, wherein the target input data is an optimal site of the site. When the iteration result meets the preset convergence condition, the implementation of obtaining the optimal site of the site includes: comparing a current iteration result and a last iteration result to obtain a gradient change value, the iteration result meets the preset convergence condition when the gradient change value is less than a first preset value, and the optimal site of the site is obtained based on the current iteration result; or, an iteration number of inputting the mapping result into the at least two target agent models to perform iteration is obtained, the iteration result meets the preset convergence condition when the iteration number meets a second preset value, and the optimal site of the site is obtained based on the current iteration result. The site feature data includes at least one site feature subset, and the target data includes at least one target subset. The result output module is specifically configured to train the site feature data and the target data in the training data set to obtain the target agent models in the following manner: cross-matching the at least one site feature subset and the at least one target subset to obtain a plurality of training data pairs, training a preset agent model based on the site feature subset and the target subset in each training data pair to obtain a plurality of target agent models.

3. An electronic device, comprising: The electronic device includes: at least one processor; and a memory connected with the at least one processor in communication; wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to perform the site selection method of claim 1.

4. A computer-readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to implement the site selection method of claim 1.

Images