CN113283680B - An address selection method, device, equipment and storage medium thereof - Google Patents

Abstract

The disclosure provides an addressing method, device, equipment and storage medium thereof. The method comprises the steps of determining an initial model containing decision variables, an objective function and constraint conditions; carrying out approximate optimization on the objective function and the constraint condition to generate an optimized objective model; acquiring basic data, wherein the basic data comprises candidate addresses taking the decision variables as characteristics; and generating a recommended address sequence containing a plurality of recommended addresses according to the values of the decision variables of the candidate addresses and the optimized target model, so that the method can be widely applied to various industries and site selection of different regions, and has strong universality.

Description

An address selection method, device, equipment and storage medium thereof

technical field

The present invention relates to the field of computer technology, and in particular, to an address selection method, apparatus, device and storage medium thereof.

Background technique

With the development of the industry, various industries (such as offline education, restaurant franchise, etc.) may face large-scale site selection problems. The existing site selection algorithms generally use site selection based on their own business experience. For example, in different industries, the variables that the site selection depends on may be different; even in the same education industry, when selecting campus sites in different regions, it is often necessary to rely on the local business experience of each city for site selection. The lack of a unified mathematical optimization model is not universal.

Based on this, a site selection scheme that adapts to a wider range of azimuths is required.

SUMMARY OF THE INVENTION

In view of this, the embodiments of the present disclosure provide a more universal site selection solution to at least partially solve the above problems.

According to an aspect of the present disclosure, a method for site selection is provided, including: determining an initial model including decision variables, an objective function and constraints; performing approximate optimization on the objective function and constraints to generate an optimized target model; Obtain basic data, which includes candidate addresses characterized by the decision variable; generate a recommended address sequence including multiple recommended addresses according to the value of the decision variable of the candidate address and the optimized target model

According to a second aspect of the present disclosure, there is provided an apparatus for location selection, comprising: a determination module for determining an initial model including decision variables, an objective function and constraints; an optimization module for performing approximate optimization on the objective function and constraints, Generate an optimized target model; an acquisition module acquires basic data, the basic data includes candidate addresses characterized by the decision variable; a recommendation module, according to the value of the decision variable of the candidate address and the optimized The target model of generates a recommended address sequence containing multiple recommended addresses.

According to a third aspect of the present disclosure, there is provided an electronic device comprising: a processor; and a memory storing a program, wherein the program includes instructions that, when executed by the processor, cause the processor to The method as described in the first aspect is performed.

According to a fourth aspect of the present disclosure, there is provided a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause the computer to perform the method of the first aspect.

According to a fifth aspect of the present disclosure, there is provided a computer program product comprising a computer program, wherein the computer program is executed by a processor according to the method of the first aspect.

One or more technical solutions provided in the embodiments of the present disclosure determine an initial model including decision variables, objective functions and constraints; perform approximate optimization on the objective functions and constraints to generate an optimized target model; acquire basic data , the basic data includes candidate addresses characterized by the decision variable; a recommended address sequence including multiple recommended addresses is generated according to the value of the decision variable of the candidate address and the optimized target model. The present disclosure first comprehensively considers the influence of decision variables, determines an objective function according to business objectives, thereby abstracts a unified initial model, and approximately optimizes the objective function and constraints in the initial model based on the actual situation, which can be widely applied to various types of The location selection of industries and different regions has strong universality.

Description of drawings

Further details, features and advantages of the present disclosure are disclosed in the following description of exemplary embodiments in conjunction with the accompanying drawings, in which:

FIG. 1 is a schematic flowchart of a site selection method provided by an embodiment of the present disclosure;

2 is a schematic diagram of a site selection engineering architecture provided by an embodiment of the present disclosure;

3 is a schematic diagram of selecting candidate addresses from basic data according to an embodiment of the present disclosure;

FIG. 4 is a schematic structural diagram of an address selection device according to an embodiment of the present disclosure;

5 shows a block diagram of an exemplary electronic device that can be used to implement embodiments of the present disclosure.

Detailed ways

Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While certain embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be construed as limited to the embodiments set forth herein, but rather are provided for the purpose of A more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes, and are not intended to limit the protection scope of the present disclosure.

It should be understood that the various steps described in the method embodiments of the present disclosure may be performed in different orders and/or in parallel. Furthermore, method embodiments may include additional steps and/or omit performing the illustrated steps. The scope of the present disclosure is not limited in this regard.

As used herein, the term "including" and variations thereof are open-ended inclusions, ie, "including but not limited to". The term "based on" is "based at least in part on." The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". Relevant definitions of other terms will be given in the description below. It should be noted that concepts such as "first" and "second" mentioned in the present disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or interdependence.

It should be noted that the modifications of "a" and "a plurality" mentioned in the present disclosure are illustrative rather than restrictive, and those skilled in the art should understand that unless the context clearly indicates otherwise, they should be understood as "one or a plurality of". multiple". The names of messages or information exchanged between multiple devices in the embodiments of the present disclosure are only for illustrative purposes, and are not intended to limit the scope of these messages or information.

The solution of the present disclosure is described below with reference to the accompanying drawings. As shown in FIG. 1 , FIG. 1 is a schematic flowchart of an address selection method provided by an embodiment of the present disclosure, which specifically includes:

S101, determine an initial model including decision variables, objective functions and constraints.

In different business contexts, decision variables can be selected based on needs. Decision variables can be a specific address or various features contained in a Point of Interest (POI). For example, a point of interest can be a residential area, an office building, or a shopping mall, etc., and the decision variables can include the location coordinates, traffic, sanitation, flow of people, and noise level of the point of interest.

The constraint condition is a constraint on the interest point generated by the relative relationship of the interest point, and the constraint condition may depend on the value of each decision variable in the interest point. When selecting candidate addresses in some business scenarios, there may already be points of interest associated with the business that needs to be located in reality. Therefore, the selection of points of interest needs to be constrained based on constraints.

For example, in the scenario where product A needs to join a business or open a product chain store, a chain store of product A already exists at address 1, or, product A's competitor product B (or can become a competitor product) already exists at address 2 ), the constraints can be based on the existing business in address 1 or address 2 to constrain the selection of other nearby points of interest, which may strengthen the selection probability of nearby addresses (for example, The business opened by the competing product in address 2 does not conflict with its own business at all, even complementary), or it may be weakened (for example, the chain store opened by product A in address 1 is exactly the same type of business to be opened). That is, the influence of competing products may be positive or negative, and the constraints may generate positive or negative constraints on the selection of candidate addresses based on business requirements and the values of decision variables in POIs.

The objective function may be an index selected based on actual needs, or an empirical function obtained by fitting based on actual experience. For example, in the offline education scenario, the ultimate goal can be to satisfy the final school, parents, students, teachers and other parties on the candidate address (this may require an empirical function to be used as the objective function to describe), or The ultimate goal can be to maximize the school's income (and the school's income can be comprehensively evaluated through the corresponding indicators, that is, the function fitted by the corresponding indicators can be used as the objective function), or you can choose Indicators that are highly related to performance are used as the objective function (for example, calculated through actual data, the correlation between the number of students and school performance is more than 0.99, and the students predicted by the values of the decision variables based on the candidate address can be used directly. person times as the objective function).

The determined initial model can be expressed as follows:

；……（1）

;……(1)

；……（2）in,

;……(2)

即为目标函数，公式（1）表征了取前N个最小的

所对应的

作为推荐地址。in

That is, the objective function, formula (1) represents the first N smallest

corresponding to

as a recommended address.

为候选地址，

为另一候选地址。

为本品的地址，

为竞品的地址，

为指定类型（例如对于在线教育选址的场景中，指定类型可以是小区或者商场等等，在消费品加盟选址的场景中，指定类型可以是购物广场或者百货大楼等等）的候选地址的集合。Formula (2) is the constraint condition.

is the candidate address,

is another candidate address.

the address of this product,

is the address of the competitor,

A collection of candidate addresses for a specified type (for example, in the scenario of online education site selection, the specified type can be a community or shopping mall, etc., in the scenario of consumer product franchise site selection, the specified type can be a shopping plaza or a department store, etc.) .

、

和

分别为各自的约束阈值，可以基于经验给定。i、j、k和N、M、P均为自然数。The four inequalities in formula (2) represent its own constraints, the constraints of this product, the constraints of competing products, and the constraints of business. one of them

,

and

are their respective constraint thresholds, which can be given based on experience. i, j, k and N, M, and P are all natural numbers.

S103, perform approximate optimization on the objective function and the constraint conditions, and generate an optimized objective model.

The general form of the optimization modeling of the location selection is given above in the aforementioned methods. Based on the actual needs, the objective function and constraints can also be approximately optimized. Specifically, it may include using an approximately optimized fitting function and an empirical function to replace the aforementioned objective function.

For example, the objective function uses business experience to select core features in a linearly weighted form to form the objective function, or the output function of training a prediction model is determined as the objective function.

Constraints can also be constrained in an approximate way, so as to obtain the constraints of this product, the constraints of competing products, or business constraints that are more in line with the business type. For example, for its own constraints, it can be approximated by a sequence greedy algorithm to obtain a recommended address sequence of candidate addresses. For competing product constraints, it can be approximated into a feature constraint, for example, competing within a few kilometers of the candidate address. The quantity of the product is less than the specified quantity, etc.

S105: Acquire basic data, where the basic data includes candidate addresses characterized by the decision variable.

The basic data may be data obtained from a geographic information system (geographic information system, GIS). Geographic information system is based on geospatial data, adopts geographic model analysis method, provides various spatial and dynamic geographic information in a timely manner, and collects, stores, analyzes and visualizes various geospatial information. For example, a GIS system can give the distance to the nearest school at an address, the distance to the nearest cell, the traffic situation at the address, and the geographic environment attributes of the address (eg, urban, suburban, populated areas, mountains, etc.) , river) and so on.

Therefore, according to the latitude and longitude coordinates of each candidate address, the value of the corresponding decision variable can be given based on the basic data. For example, for address 3, the value of its decision variable may be (school district score 8, traffic score 8, people flow score 10) and so on.

S107: Generate a recommended address sequence including multiple recommended addresses according to the value of the decision variable of the candidate address and the optimized target model.

所对应的

作为推荐地址，并且可以基于

的取值从小到大进行排序，从而得到包含多个推荐地址的推荐地址序列。在推荐地址序列中，

越小，推荐地址

排序越靠前，也表征该地址

可能是在目标函数下更优异的地址。After multiple candidate addresses containing decision variables are obtained, they can be input into the aforementioned target model, and the target function corresponding to each candidate address is calculated based on the constraints of the target model and the value of the decision variables of the candidate addresses. The value of , and then select the first N smallest

corresponding to

as a recommended address and can be based on

The value of is sorted from small to large, so as to obtain a recommended address sequence containing multiple recommended addresses. In the recommended address sequence,

The smaller the recommended address

The higher the order is, the more the address is represented

Possibly a better address under the objective function.

One or more technical solutions provided in the embodiments of the present disclosure determine an initial model including decision variables, objective functions and constraints; perform approximate optimization on the objective functions and constraints to generate an optimized target model; acquire basic data , the basic data includes candidate addresses characterized by the decision variable; a recommended address sequence including multiple recommended addresses is generated according to the value of the decision variable of the candidate address and the optimized target model. The present disclosure first comprehensively considers the influence of decision variables, determines an objective function according to business objectives, thereby abstracts a unified initial model, and approximately optimizes the objective function and constraints in the initial model based on the actual situation, which can be widely applied to various types of The location selection of industries and different regions has strong universality.

In one embodiment, the optimized target model can be implemented using an actual engineering architecture. As shown in FIG. 2 , FIG. 2 is a schematic diagram of a site selection engineering architecture provided by an embodiment of the present disclosure. In this architecture, it is mainly necessary to build an application platform with the target model as the online strategy layer. At the same time, the supporting base and data layer are connected at the back end, and the application layer (ie, the business and user layers) is connected at the front end.

Specifically, the application platform can obtain basic data including POIs of schools, communities or points of interest from the docked geographic information system, and can directly use the obtained POIs of schools, communities or points of interest as candidate points. The value of the decision variable of the corresponding candidate address is directly obtained in the geographic information system to form the base layer, and this part can be pre-completed on an offline basis.

Further, on the basis of the base, the obtained basic data can be further preprocessed, including the formation of entity data (coordinates of each entity in the geographic system), recall data (possible candidate addresses), and blacklist data. (that is, a location has been listed as an unsuitable location) and characteristic data (optimized values of decision variables at each location), and so on.

Therefore, the application platform can directly generate the recommended address sequence according to the basic data and the target model in the online policy layer, and display the recommended address sequence to the user at the application layer. Site selection through the application platform has high reusability and expansibility, and can be directly applied to various types of industry site selection.

In one embodiment, for the constructed application platform, an evaluation index may also be established to evaluate the performance of the created application platform. Specifically, when the application platform displays the recommended addresses, the user may not always select from the recommended address sequence. For example, the recommended address sequence gives address 1, address 2 and address 10, but the user himself selects address 3 and address 10 in the map as the final opening addresses. Therefore, the total amount of addresses in the recommended address sequence can be determined; and an evaluation index is determined according to the total amount of addresses and the number of addresses selected by the user in the recommended address sequence, for evaluating the performance of the application platform . The evaluation index is usually negatively correlated with the total number of addresses and positively correlated with the number of selected addresses.

用户选取的地址集合）/num（用户选取的地址）。例如，推荐地址集合给出了地址1至地址10的10个地址，而用户选取的地址集合为地址1、地址2和地址11，则用户采纳率=2/3=66.6%。In one embodiment, the ratio of the number of addresses selected by the user in the recommended address sequence to the total amount of addresses may be determined as the user site adoption rate, and the user site acceptance rate may be determined as evaluation indicators. That is, user adoption rate = num (recommended address set

User-selected address set)/num (user-selected addresses). For example, the recommended address set gives 10 addresses from address 1 to address 10, and the address set selected by the user is address 1, address 2 and address 11, then the user adoption rate=2/3=66.6%.

The evaluation index = max user adoption rate & min (num recommended address sets), that is, when the number of recommended address sets is less and the user adoption rate is greater, the evaluation index is higher. Assuming that in another example, if the recommended address set gives 20 addresses from address 1 to address 20, and the address set selected by the user is address 1, address 2, address 11, address 12 and address 21, address 22, then The user adoption rate at this time=4/6=66.6%, but since the number of recommended address sets in the previous example is smaller, the performance of the application platform corresponding to the previous example will be higher.

In one embodiment, the optimization of the objective model may be the optimization of an objective function. For example, some decision variables are selected, and an objective function that linearly weights some decision variables is constructed.

。The features here can be school type features included in the candidate address (whether there is a school nearby, school quality assessment), community type features (whether there is a nearby community, the type of community, the population of the community, etc.), traffic characteristics (the degree of transportation convenience) , traffic noise level, etc.). Thus, a linear weighting function composed of partial features is formed, that is,

.

和

即为候选地址的经度和纬度坐标，

表征了在经纬坐标为

和

的情形下的第r个特征，l为特征的维度。in,

and

are the longitude and latitude coordinates of the candidate address,

Characterized in latitude and longitude coordinates as

and

The rth feature in the case of , l is the dimension of the feature.

的预测模型，其中，G即为预测模型的目标函数，W*为预测模型的目标函数的各参数的取值。具体而言，Alternatively, a prediction model can be pre-trained to predict the value of the objective function corresponding to a certain candidate address. For example, given a set X of training samples characterized by the school's income or the number of students and the decision variables corresponding to the school's address, the model is trained, so that the training can obtain a satisfactory

The prediction model of , where G is the objective function of the prediction model, and W* is the value of each parameter of the objective function of the prediction model. in particular,

。

.

为预设的经验项，用于防止预测模型的过拟合。当最终计算得到的W*能够满足

时，即可以认为预测模型训练完成，进而可以基于候选地址的各决策变量的取值而预测一个候选地址在目标函数下的取值，并通过其与预设值的差异判断其是否适合作为推荐地址。Among them, h(Xi, W) is the predicted value of the training sample Xi (that is, the candidate address Xi) when the parameter value is W, yi is the label value of the training sample Xi, L is the loss function,

It is a preset empirical term used to prevent overfitting of the prediction model. When the final calculated W* can satisfy

, it can be considered that the training of the prediction model is completed, and then the value of a candidate address under the objective function can be predicted based on the value of each decision variable of the candidate address, and whether it is suitable as a recommendation based on the difference from the preset value is judged. address.

In one embodiment, the constraints may specifically include:

Self-constraint conditions, used to constrain the geographic distribution between the recommended addresses obtained in the recommended address sequence, for example, constrain the distance between the recommended addresses to be no less than a preset value;

And/or, this product constraint, when there is a first type of address related to its own business in the basic data, the self constraint is used to constrain each recommended address in the recommended address sequence and the first type of address. The geographic distribution of a class of addresses. For example, it is used to constrain the distance between each recommended address and the first type of address not to be less than a preset value;

And/or, competitive product constraints, when there is a second type of address related to a competitor in the basic data, the competitive product constraints are used for each recommended address in the recommended address sequence. Geographical distribution of the second type of address, for example, if the competing product and the business type to be opened are the same, the distance between each recommended address and the second type of address shall not be lower than the preset value, and if the competing product and the to-be-opened address are of the same type If the service types are complementary, the distance between each recommended address and the second type of address shall not be greater than the preset value;

And/or, the point of interest constraint is used to constrain the selection method of determining the candidate address according to the point of interest in the acquired basic data, for example, directly selecting the point of interest from the basic data as the candidate address, or facing the basic map data. For further grid division, candidate addresses are selected from the divided grids;

And/or, the business constraints are used to constrain the geographic environment attributes of the recommended addresses in the recommended address sequence, for example, the geographic environment where the recommended addresses are located must not be mountains, rivers, and the like.

In an embodiment, the approximate optimization of the constraints may include approximate optimization of the self-constraints. Specifically, a sequence greedy algorithm may be used to approximate a sequence algorithm for jointly generating recommendation points. That is, after determining the i-th recommended address in the recommended address sequence, determine the remaining candidate addresses that are greater than the preset distance from the i-th recommended address, where i is a natural number greater than 1; The candidate address with the smallest distance from the i-th recommended address is determined as the i-1-th recommended address in the recommended address sequence. In this way, the efficiency of the generated recommended address sequence can be improved.

In one embodiment, the approximate optimization of the constraints may include approximate optimization of the constraints of the product. For example, for offline education scenarios, the geographical distribution of the teaching points of this product needs to meet a certain network structure, so that the average distance for users to go to school cannot exceed the preset value. Therefore, the constraints of this product can be approximately optimized as follows: within a preset range, the number of the first type of addresses and the recommended addresses exceeds the preset number range (that is, the density of this product must reach a certain value); or, The average distance between a specified type of POI (including a residential area or a school) and the nearest first-type addresses and recommended addresses does not exceed the preset value, so that it can be closer to the actual scene requirements.

Of course, in other scenarios, for example, when a chain food store joins, in order to avoid vicious competition among the products, it may be necessary to set the number of the first type of addresses and the recommended addresses within a preset range. If it exceeds a certain preset value, it is necessary to control the density of this product.

In one embodiment, the approximate optimization of the constraints may include approximate optimization of the constraints of competing products. For example, the constraints of competing products can be directly converted into feature constraints, that is, within the preset distance range of the second type of addresses, the number of recommended addresses in the recommended address sequence does not exceed the preset number threshold, so that the pre-determined number of recommended addresses Within the set range, the number of competing products is less than the specified threshold.

In an embodiment, when the basic data is geospatial data (ie, map data) obtained from a geographic information system, approximately optimizing the constraints of the point of interest includes: performing an optimization on the geospatial data Perform grid division, and if there is a school, community or POI in the grid, determine the school, community or POI as a candidate address.

If there is no school, community or POI in a grid, further evaluation can be performed based on the values of decision variables of other entity data in the grid, that is, other entity addresses can be selected from the grid As a candidate address, for example, select the vicinity of a bus stop in the grid as a candidate address.

Further, if multiple candidate addresses can be selected in one grid, only a specified number (eg, 1) of addresses can be selected from one grid as candidate addresses, so as to avoid too many candidate addresses. As shown in FIG. 3 , FIG. 3 is a schematic diagram of obtaining candidate addresses from basic data according to an embodiment of the present disclosure. In this illustration, the black dots represent schools, communities or POIs, and the white dots represent other physical addresses (including traffic stops, etc.) There may be multiple black dots in some grids, so only one is selected in the end. As a candidate address, for a grid without a school, community or POI, the location of the white point or the location of the grid center point can be selected as the candidate address.

In an embodiment, the application platform that has been constructed can also be rearranged based on business requirements. For example, the recommended address sequence that has been obtained can be rearranged based on the relative distance to a certain cell, thereby Can meet actual business needs. Further, recommendation information may be associated with the recommended addresses in the recommended address sequence, and the recommended addresses and the recommended information may be displayed. For example, the address is 300m away from the xx community, and it is predicted that the number of students may reach 600 within 3 months of opening, thereby improving the user's site selection experience.

In a second aspect of the embodiment of the present disclosure, an address selection device is also provided. As shown in FIG. 4 , FIG. 4 is a schematic structural diagram of the address selection device provided by the embodiment of the present disclosure, including:

Determining module 401, determining an initial model including decision variables, objective functions and constraints;

The optimization module 403 performs approximate optimization on the objective function and constraints, and generates an optimized objective model;

Obtaining module 405, obtaining basic data, where the basic data includes candidate addresses characterized by the decision variable;

The recommendation module 407 generates a recommended address sequence including multiple recommended addresses according to the value of the decision variable of the candidate address and the optimized target model.

In a third aspect of the embodiments of the present disclosure, exemplary embodiments of the present disclosure further provide an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor. The memory stores a computer program executable by the at least one processor for causing the electronic device to perform a method according to an embodiment of the present disclosure when executed by the at least one processor.

Exemplary embodiments of the present disclosure also provide a non-transitory computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor of a computer, is used to cause the computer to execute a computer program according to an embodiment of the present disclosure. method.

Exemplary embodiments of the present disclosure also provide a computer program product comprising a computer program, wherein the computer program, when executed by a processor of a computer, is used to cause the computer to perform a method according to an embodiment of the present disclosure.

Referring to FIG. 5 , a structural block diagram of an electronic device 800 that can serve as a server or client of the present disclosure will now be described, which is an example of a hardware device that can be applied to various aspects of the present disclosure. Electronic devices are intended to represent various forms of digital electronic computer devices, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. Electronic devices may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are by way of example only, and are not intended to limit implementations of the disclosure described and/or claimed herein.

As shown in FIG. 5 , the electronic device 800 includes a computing unit 801 , which can be programmed according to a computer program stored in a read only memory (ROM) 802 or loaded into a random access memory (RAM) 803 from a storage unit 808 . Various appropriate actions and processes are performed. In the RAM 803, various programs and data necessary for the operation of the device 800 can also be stored. The computing unit 801 , the ROM 802 , and the RAM 803 are connected to each other through a bus 804 . An input/output (I/O) interface 805 is also connected to bus 804 .

Various components in the electronic device 800 are connected to the I/O interface 805 , including: an input unit 806 , an output unit 807 , a storage unit 808 , and a communication unit 809 . The input unit 806 may be any type of device capable of inputting information to the electronic device 800, and the input unit 806 may receive input numerical or character information and generate key signal input related to user settings and/or function control of the electronic device. The output unit 807 may be any type of device capable of presenting information, and may include, but is not limited to, a display, speakers, video/audio output terminals, vibrators, and/or printers. The storage unit 804 may include, but is not limited to, magnetic disks and optical disks. Communication unit 809 allows electronic device 800 to exchange information/data with other devices through computer networks such as the Internet and/or various telecommunication networks, and may include, but is not limited to, modems, network cards, infrared communication devices, wireless communication transceivers, and/or chips Groups such as Bluetooth™ devices, WiFi devices, WiMax devices, cellular communication devices and/or the like.

Computing unit 801 may be various general-purpose and/or special-purpose processing components with processing and computing capabilities. Some examples of computing units 801 include, but are not limited to, central processing units (CPUs), graphics processing units (GPUs), various specialized artificial intelligence (AI) computing chips, various computing units that run machine learning model algorithms, digital signal processing processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 801 performs the various methods and processes described above. For example, in some embodiments, the addressing method according to the first aspect may be implemented as a computer software program tangibly embodied on a machine-readable medium, such as storage unit 808 . In some embodiments, part or all of the computer program may be loaded and/or installed on the electronic device 800 via the ROM 802 and/or the communication unit 809 . In some embodiments, the computing unit 801 may be configured by any other suitable means (eg, by means of firmware) to perform the addressing method as in the first aspect.

Program code for implementing the methods of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general purpose computer, special purpose computer or other programmable data processing apparatus, such that the program code, when executed by the processor or controller, performs the functions/functions specified in the flowcharts and/or block diagrams. Action is implemented. The program code may execute entirely on the machine, partly on the machine, partly on the machine and partly on a remote machine as a stand-alone software package or entirely on the remote machine or server.

In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in connection with the instruction execution system, apparatus or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media may include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media would include one or more wire-based electrical connections, portable computer disks, hard disks, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (EPROM or flash memory), fiber optics, compact disk read only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

As used in this disclosure, the terms "machine-readable medium" and "computer-readable medium" refer to any computer program product, apparatus, and/or apparatus for providing machine instructions and/or data to a programmable processor (eg, magnetic disk, optical disk, memory, programmable logic device (PLD)), including a machine-readable medium that receives machine instructions as machine-readable signals. The term "machine-readable signal" refers to any signal used to provide machine instructions and/or data to a programmable processor.

To provide interaction with a user, the systems and techniques described herein may be implemented on a computer having: a display device (eg, a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user ); and a keyboard and pointing device (eg, a mouse or trackball) through which a user can provide input to the computer. Other kinds of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (eg, visual feedback, auditory feedback, or tactile feedback); and can be in any form (including acoustic input, voice input, or tactile input) to receive input from the user.

The systems and techniques described herein can be implemented on a computing system that includes back-end components (eg, as a data server), or a computing system that includes middleware components (eg, an application server), or a computing system that includes front-end components (eg, a user computer having a graphical user interface or web browser through which a user can interact with implementations of the systems and techniques described herein), or including such backend components, middleware components, Or any combination of front-end components in a computing system. The components of the system may be interconnected by any form or medium of digital data communication (eg, a communication network). Examples of communication networks include: Local Area Networks (LANs), Wide Area Networks (WANs), and the Internet.

A computer system can include clients and servers. Clients and servers are generally remote from each other and usually interact through a communication network. The relationship of client and server arises by computer programs running on the respective computers and having a client-server relationship to each other.

Claims (13)

1. A site selection method, comprising: Determine the initial model including decision variables, objective function and constraints; Perform approximate optimization on the objective function and constraints to generate an optimized objective model; Obtain basic data, the basic data includes candidate addresses characterized by the decision variable; Generate a recommended address sequence including multiple recommended addresses according to the value of the decision variable of the candidate address and the optimized target model; The performing approximate optimization on the objective function includes: training a prediction model using the training samples characterized by the decision variable, and determining an output function of the prediction model as the objective function; Before obtaining the basic data, the method further includes: constructing an application platform with the target model as an online strategy layer; the obtaining the basic data includes: or basic data of POIs; Described adopting the training samples characterized by the decision variable to obtain a prediction model, and determining the output function of the prediction model as the objective function, including: Given a set X of training samples characterized by the school's income or the number of students, and the decision variables corresponding to the school's address, perform model training, and get a satisfactory result.

The prediction model of , where G is the objective function of the prediction model, W* is the value of each parameter of the objective function of the prediction model,

; Among them, h(Xi, W) is the predicted value of the training sample Xi when the parameter value is W, yi is the label value of the training sample Xi, L is the loss function,

is a preset empirical term used to prevent over-fitting of the prediction model; when the final calculated W* can satisfy the

When the training of the prediction model is completed, and then predict the value of the candidate address under the objective function based on the value of each decision variable of the candidate address, and pass the value of the candidate address under the objective function and the preset value The difference judges whether the candidate address is suitable as a recommended address. 2 . The method according to claim 1 , wherein the generating a recommended address sequence including a plurality of recommended addresses according to the value of the decision variable of the candidate address and the optimized target model, comprises: the applying The platform generates and displays the recommended address sequence according to the basic data and the online strategy layer. 3. The method of claim 2, wherein the method further comprises: determining the total number of addresses in the recommended address sequence; The evaluation index is determined according to the total amount of addresses and the number of addresses selected by the user in the recommended address sequence, and is used to evaluate the performance of the application platform. 4. The method according to claim 3, wherein determining the evaluation index according to the total amount of addresses and the number of addresses selected by the user in the recommended address sequence, comprising: The ratio of the number of addresses selected by the user in the recommended address sequence to the total amount of addresses is determined as a user site adoption rate, and the user site adoption rate is determined as an evaluation index. 5. The method of claim 1, wherein the constraints comprise: Self-constraint conditions for constraining the geographic distribution among the recommended addresses obtained in the recommended address sequence; and/or, Constraints of this product, when there is a first-class address related to its own business in the basic data, the self-constraint is used to constrain the relationship between each recommended address in the recommended address sequence and the first-class address geographic distribution; and/or, Competing product constraints, when there is a second type of address related to a competitor in the basic data, the competing product constraints are used for each recommended address in the recommended address sequence and the second type of address. geographical distribution; and/or, point of interest constraints, used to constrain the selection method of determining candidate addresses according to points of interest in the acquired basic data; and/or, The business constraints are used to constrain the geographic environment attributes of the recommended addresses in the recommended address sequence. 6. The method of claim 5, wherein approximately optimizing the constraints comprises: Approximate optimization of the self-constraint condition, which specifically includes: after determining the i-th recommended address in the recommended address sequence, determining the remaining candidate addresses that are greater than a preset distance from the i-th recommended address, where i is greater than 1 the natural numbers; A candidate address with the smallest distance from the i-th recommended address among the remaining candidate addresses is determined as the i-1-th recommended address in the recommended address sequence. 7. The method of claim 5, wherein approximately optimizing the constraints comprises: Approximate optimization of the constraints of this product, specifically including: within a preset range, the number of the first type of address and the recommended address exceeds the preset number range; The average distance between the plurality of first-class addresses and the recommended addresses does not exceed a preset value. 8. The method of claim 5, wherein approximately optimizing the constraints comprises: Approximate optimization of the competitive product constraints specifically includes: within a preset distance range of the recommended addresses, the number of the second type of addresses does not exceed a preset number threshold. 9. The method of claim 5, wherein, when the basic data is geospatial data obtained from a geographic information system, approximately optimizing the constraints of the point of interest, specifically comprising: Perform grid division on the geospatial data, and if there is a school, community or POI in the grid, determine the school, community or POI as a candidate address; If there is no school, community or POI in the grid, a candidate address is selected according to the values of decision variables of other addresses included in the grid. 10. The method of claim 2, wherein the method further comprises: The application platform, according to business requirements, rearranges the recommended addresses in the recommended address sequence, and displays the rearranged recommended address sequence; or, the application platform, reorders the recommended addresses in the recommended address sequence Associate the recommended information, and display the recommended address and the recommended information. 11. An address selection device, comprising: Determine the module to determine the initial model containing decision variables, objective functions and constraints; an optimization module, which approximately optimizes the objective function and the constraints to generate an optimized objective model; an acquisition module to acquire basic data, the basic data includes candidate addresses characterized by the decision variable; a recommendation module, which generates a recommended address sequence including a plurality of recommended addresses according to the value of the decision variable of the candidate address and the optimized target model; The performing approximate optimization on the objective function includes: training a prediction model using the training samples characterized by the decision variable, and determining an output function of the prediction model as the objective function; The location selection device further includes: a building module, the building module is used to build an application platform with the target model as an online strategy layer; the acquiring basic data includes: the application platform acquires from a geographic information system, including Basic data of schools, communities or POIs; Described adopting the training samples characterized by the decision variable to obtain a prediction model, and determining the output function of the prediction model as the objective function, including: Given a set X of training samples characterized by the school's income or the number of students, and the decision variables corresponding to the school's address, perform model training, and get a satisfactory result.

The prediction model of , where G is the objective function of the prediction model, W* is the value of each parameter of the objective function of the prediction model,

; Among them, h(Xi, W) is the predicted value of the training sample Xi when the parameter value is W, yi is the label value of the training sample Xi, L is the loss function,

is a preset empirical term used to prevent over-fitting of the prediction model; when the final calculated W* can satisfy the

When the training of the prediction model is completed, and then predict the value of the candidate address under the objective function based on the value of each decision variable of the candidate address, and pass the value of the candidate address under the objective function and the preset value The difference judges whether the candidate address is suitable as a recommended address. 12. An electronic device comprising: processor; and memory to store programs, wherein the program comprises instructions which, when executed by the processor, cause the processor to perform the method of any of claims 1-10. 13. A non-transitory computer readable storage medium storing computer instructions for causing the computer to perform the method of any of claims 1-10.

Images