CN114943407A - Area planning method, device, equipment, readable storage medium and program product - Google Patents

Abstract

The present application discloses a regional planning method, apparatus, device, readable storage medium and program product, and relates to the field of machine learning. The method includes: acquiring a geographic location set, where the geographic location set includes geographic location coordinates corresponding to at least two target type entities within a candidate area; determining an environmental data set corresponding to the at least two target type entities based on the geographic location coordinates; combining the entities Business distribution and entity location relationship Perform location analysis on at least two target-type entities to obtain candidate centers within the candidate area; divide the candidate area based on the candidate centers to obtain at least two target-type entities within the candidate area. Regional planning results. That is, by analyzing the location of the target type entity in combination with the entity business distribution to determine the regional planning result, the accuracy and planning efficiency of the regional planning can be better improved.

Description

Regional planning method, apparatus, device, readable storage medium and program product

technical field

The embodiments of the present application relate to the field of machine learning, and in particular, to a regional planning method, apparatus, device, readable storage medium, and program product.

Background technique

Regional planning usually refers to the process of dividing a certain area into several non-overlapping sub-areas, such as dividing an urban area into several non-overlapping sub-areas. Wherein, during the division, a plurality of center points are usually determined in the area, so that the sub-area range corresponding to the point is determined according to each center point.

In the related art, after determining the n location points (n≥2) contained in a current area, a city is divided into a density-based clustering method with noise (Density-Based Spatial Clustering of Applications with Noise, DBSCAN). The n location points included in the cluster are clustered according to their corresponding geographical location distribution relationships to obtain the final clustering result, and the sub-region planning result is determined according to the clustering result.

However, in the above method, because the clustering method only relies on the geographical location distribution relationship of the location points, it cannot meet different business requirements, resulting in a low accuracy rate of regional planning results.

SUMMARY OF THE INVENTION

The embodiments of the present application provide a regional planning method, apparatus, device, readable storage medium, and program product, which can improve the accuracy of regional planning. The technical solution is as follows:

In one aspect, a regional planning method is provided, the method comprising:

Obtain a geographic location set, the geographic location set includes geographic location coordinates corresponding to at least two target type entities within the scope of the candidate area;

Based on the geographic location coordinates, determine the environmental data set corresponding to the at least two target type entities, where the environmental data set is used to indicate the entity location relationship and the entity business distribution corresponding to the at least two target type entities, and the The entity business distribution is used to indicate the spatial and temporal distribution of the business of the target type entity;

Performing a location analysis on the at least two target type entities in combination with the entity business distribution and the entity location relationship to obtain a candidate center within the candidate area;

Region division is performed on the candidate region range based on the candidate center to obtain a region planning result of the at least two target type entities within the candidate region range.

In another aspect, an apparatus for regional planning is provided, the apparatus comprising:

an acquisition module, configured to acquire a geographic location set, the geographic location set includes geographic location coordinates corresponding to at least two target type entities within the scope of the candidate area;

a determining module, configured to determine, based on the geographic location coordinates, an environmental data set corresponding to the at least two target type entities, where the environmental data set is used to indicate the entity location relationship and entity corresponding to the at least two target type entities Service distribution, where the entity service distribution is used to indicate the spatiotemporal distribution of the service of the target type entity;

an analysis module, configured to perform location analysis on the at least two target type entities in combination with the entity business distribution and the entity location relationship, to obtain a candidate center within the candidate area;

A division module, configured to perform area division on the candidate area range based on the candidate center, and obtain the area planning result of the at least two target type entities within the candidate area range.

In another aspect, a computer device is provided, the computer device includes a processor and a memory, the memory stores at least one instruction, at least a section of a program, a code set or an instruction set, the at least one instruction, the at least one A piece of program, the code set or the instruction set is loaded and executed by the processor to implement the regional planning method as described in any of the foregoing embodiments of the present application.

In another aspect, a computer-readable storage medium is provided, wherein the storage medium stores at least one instruction, at least one piece of program, code set or instruction set, the at least one instruction, the at least one piece of program, the code The set or instruction set is loaded and executed by the processor to implement the regional planning method as described in any of the above embodiments of the present application.

In another aspect, a computer program product or computer program is provided, the computer program product or computer program comprising computer instructions stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the regional planning method described in any of the foregoing embodiments.

The beneficial effects brought by the technical solutions provided in the embodiments of the present application include at least:

By analyzing the geographic location coordinates of the target type entities in the geographic location, the entity location relationship and entity business distribution corresponding to each target type entity are determined. The candidate center used to divide the scope of the candidate area, and finally obtain the regional planning result, that is, by determining the entity location relationship and entity business distribution corresponding to the target type entity, it is possible to combine the target type entity in the process of determining the candidate center in the regional planning. The corresponding relevant business information can meet regional planning schemes under different business needs, thereby improving the accuracy of regional planning.

Description of drawings

In order to illustrate the technical solutions in the embodiments of the present application more clearly, the following briefly introduces the drawings that are used in the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can also be obtained from these drawings without creative effort.

1 is a schematic diagram of a regional planning method provided by an exemplary embodiment of the present application;

FIG. 2 is a schematic diagram of an implementation environment provided by an exemplary embodiment of the present application;

3 is a flow chart of regional planning provided by an exemplary embodiment of the present application;

4 is a flowchart of a regional planning method provided by another exemplary embodiment of the present application;

5 is a schematic diagram of an actor network analysis process provided by an exemplary embodiment of the present application;

6 is a schematic diagram of a regional planning result provided by another exemplary embodiment of the present application;

7 is a flowchart of a regional planning method provided by an exemplary embodiment of the present application;

8 is a schematic diagram of a training process of a target strategy model provided by an exemplary embodiment of the present application;

9 is a flowchart of a regional planning method provided by another exemplary embodiment of the present application;

10 is a structural diagram of a regional planning apparatus provided by an exemplary embodiment of the present application;

11 is a structural diagram of a regional planning apparatus provided by another exemplary embodiment of the present application;

FIG. 12 is a structural block diagram of a computer device provided by an embodiment of the present application.

Detailed ways

In order to make the objectives, technical solutions and advantages of the present application clearer, the embodiments of the present application will be further described in detail below with reference to the accompanying drawings.

First, briefly introduce the terms involved in the embodiments of the present application:

Markov Decision Process (MDP): It is a mathematical model of sequential decision-making (Sequential Decision), which is mainly constructed through a set of interactive objects, namely Agent and Environment. The elements include State, Action, Policy and Reward. In the MDP model, the agent obtains the current system state and performs corresponding actions on the environment according to the policy, thereby changing the environment. state and get the corresponding expected value, and finally determine the synthesis of all expected values as a return.

Actor-Critic (AC): A reinforcement learning algorithm that combines a policy-based algorithm and an expectation-based algorithm, in which the policy-based algorithm is used as an actor network for The corresponding behavior is selected based on the probability. The algorithm based on the expectation value is used as the critic network to judge the score of the behavior corresponding to the "actor", and then the "actor" will adjust the probability of the behavior according to the score of the "critic", and finally used for training policy-based algorithm to obtain the target model.

Machine Learning (ML): It is a multi-field interdisciplinary subject involving probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory and other subjects. Machine learning is the core of artificial intelligence and the fundamental way to make computers intelligent, and its applications are in all fields of artificial intelligence. Machine learning and deep learning usually include artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, teaching learning and other techniques.

Illustratively, please refer to FIG. 1 , which shows a schematic diagram of a regional planning method provided by an exemplary embodiment of the present application. As shown in FIG. 1 , a candidate region range 100 is determined, and the candidate region range 100 includes at least two target types. The geographic location coordinates corresponding to the entities (represented by hollow circles in the figure), wherein the target type entities are: shops, urban buildings, etc., and the environmental data set 110 corresponding to each target type entity is determined according to the geographic location coordinates, wherein the environmental data The set 110 includes the entity location relationship 111 and the entity business distribution 112 corresponding to at least two target type entities. The location analysis is performed on the target type entity in combination with the entity location relationship 111 and the entity business distribution 112 to obtain the candidate center 120 ( 1), the area center 121 is determined according to the candidate center 120, the area of the candidate area is divided by the area center 121, and the area planning result 130 within the candidate area area 100 is determined.

The regional planning method provided by the present application, by determining the entity location relationship and entity business distribution corresponding to the target type entity, and performing location analysis on it, can be combined with the target type entity in the process of determining the candidate center in the regional planning. related entity business information, thereby improving the accuracy of regional planning.

In conjunction with the above-mentioned nomenclature introduction, the application scenarios of the embodiments of the present application are illustrated with examples:

1. Applied to the distribution area planning scenario. In the distribution area planning scenario, such as: express transportation area planning, takeaway distribution area planning, etc., take out food delivery area planning as an example, obtain the geographical coordinates corresponding to n corresponding businesses in a city, n is a positive integer, according to n The geographic location coordinates corresponding to each merchant determine the corresponding location relationship among n merchants and the related business distribution corresponding to the merchants (such as historical order data, the merchant's expectation of the number of orders, the merchant's delivery difficulty, etc.), which are not limited here. ), as the environmental data set corresponding to n merchants, analyze the location of the merchants in combination with the corresponding location relationships among the n merchants and the related business distribution corresponding to the merchants, and obtain candidate merchants. The takeaway distribution area is divided into regions, and the final result of the takeaway distribution area division is obtained;

2. Applied to road planning scenarios. When planning the roads in a city, the geographic location coordinates corresponding to n urban buildings contained in a city are obtained, and the corresponding geographic location coordinates between the n urban buildings are determined according to the geographic location coordinates corresponding to the n urban buildings. The location relationship and the business distribution corresponding to the urban buildings (such as the traffic flow situation corresponding to the location of the urban buildings, the number of residents, etc.), as the environmental data set corresponding to the n urban buildings, combined with the positions corresponding to the n urban buildings The relationship and the business distribution corresponding to the city buildings are used to analyze the location of n city buildings, determine the candidate city buildings, divide the area within a city according to the candidate city buildings to obtain the division result, and specify the road according to the contour corresponding to the divided area. Proposal.

It is worth noting that the embodiments of the present application provide a regional planning method, which may be implemented by a terminal, may also be implemented by a server, or may be implemented collaboratively by a terminal and a server.

Wherein, when the terminal and the server cooperate to implement the solution provided by the embodiments of the present application, the terminal and the server may be directly or indirectly connected through wired or wireless communication, which is not limited in the embodiments of the present application.

Illustratively, please refer to FIG. 2 , which shows a schematic diagram of an implementation environment provided by an exemplary embodiment of the present application. As shown in FIG. 2 , the implementation environment includes a terminal 210 , a server 220 , and a connection between the terminal 210 and the server 220 Connected through the communication network 230 .

Wherein, the terminal 210 is installed with an application program that provides a regional planning function. When the terminal 210 runs the application program, the terminal 210 sends the regional planning request to the server 220.

The target policy model 221 in the server 220 determines the corresponding environmental data set according to the regional planning area sent by the terminal 210, wherein the environmental data set includes the entity location relationship and entity business distribution corresponding to at least two target type entities, and the server also Including the target strategy model 221 , the environment data set is input into the target strategy model and the output is obtained to obtain candidate centers, at least two target type entities are divided into regions according to the candidate centers, and the final regional planning result is determined and fed back to the terminal 210 .

Among them, the target policy model 221 contained in the server 220 is obtained by training the AC framework, wherein the actor network 223 (Actor network) and the critic network 224 (Critic network) in the AC framework are trained through the sample data set 222 .

The terminal 210 includes at least one of a smart phone, a tablet computer, a portable laptop computer, a desktop computer, a smart speaker, a smart wearable device, a vehicle-mounted terminal, and the like.

It is worth noting that the above-mentioned communication network 230 may be implemented as a wired network or a wireless network, and the communication network 230 may be implemented as any one of a local area network, a metropolitan area network, or a wide area network, which is not covered in this embodiment of the present application. limited.

In some embodiments, the above-mentioned server 220 may be implemented as an independent single server, or may be implemented as a group of servers, and the group of servers includes a plurality of servers working together.

In some embodiments, the above server 220 may be implemented as a cloud server in the cloud, or may be implemented as a node in a blockchain system, which is not limited herein.

It is worth noting that the above-mentioned communication network 230 may be implemented as a wired network or a wireless network, and the communication network 230 may be implemented as any one of a local area network, a metropolitan area network, or a wide area network, which is not covered in this embodiment of the present application. limited.

The area planning method provided by the embodiment of the present application is described in conjunction with the above-mentioned nomenclature introduction and application scenario, and the method is executed by a server to complete the description as an example. For an illustration, please refer to FIG. 3 , which shows an example of the present application. The flow chart of the regional planning method provided by the exemplary embodiment, the method includes the following steps:

Step 301, obtaining a geographic location set.

The geographic location set includes geographic location coordinates corresponding to at least two target type entities within the range of the candidate area.

Illustratively, the range of the candidate area refers to the designated area corresponding to the current need for regional planning. For example, if a city is used as the designated area, the range of the candidate area is the urban area corresponding to the city, and if a park is used as the designated area, the range of the candidate area is For the park area corresponding to the park, the set of areas corresponding to multiple cities is used as the designated area, and the range of the candidate area is the set of areas corresponding to multiple cities, etc., which is not limited in this application.

In some embodiments, the target type entity includes entities corresponding to the specified type within the candidate area, for example: when the specified type is a merchant, the target type entity is a merchant entity (such as a store corresponding to the merchant); when the specified type is a landmark When the target type entity is the building corresponding to the landmark, when the specified type is the traffic location, the target type entity is the relevant entity (such as a building or other reference objects) corresponding to the location of the traffic location, which is not limited here. .

Wherein, when the target type entity is a business entity, the geographic location coordinates corresponding to the target type entity are the store location coordinates corresponding to the business; when the target type entity is a building, the geographic location coordinates corresponding to the target type entity are the locations corresponding to the building Coordinates; when the target type entity is a traffic location, the geographic location coordinates corresponding to the target type entity are the location coordinates within the candidate area corresponding to the traffic location.

Optionally, at least two target type entities within the scope of the candidate area are entities of the same type, or entities of different types, which are not limited herein.

Step 302: Determine environment data sets corresponding to at least two target type entities based on the geographic location coordinates.

The environment data set is used to indicate the entity location relationship and entity service distribution corresponding to at least two target type entities, and the entity service distribution is used to indicate the spatiotemporal distribution of services of the target type entities.

Optionally, the entity location relationship includes a route distance between two target-type entities; or, a distribution relationship of corresponding locations between multiple target-type entities, such as: entity a, entity b, and entity c present a circular distribution; Or, the location distribution status of multiple target-type entities within the candidate area, for example: entity 1, entity 2 and entity 3 correspond to the location points in a cluster, entity 4 is the corresponding isolated point outside the cluster, in This is not limited, wherein the entity location relationship is determined by the respective geographic location coordinates corresponding to at least two target type entities.

Illustratively, the entity business distribution includes the time distribution or spatial distribution of the business of the target type entity, wherein, when the business types corresponding to the target type entities are different, the corresponding entity business distribution may be the same or different. .

Among them, when the target type entity is a merchant entity, the business types corresponding to the merchant entity include delivery business, operation business, etc. The delivery business is used as an example to illustrate, for example: when multiple merchants within a specified area are used for delivery area When dividing, the distribution in time of the physical business corresponding to the merchant entity includes: the number of historical orders of the merchant in the historical time period, the business hours corresponding to the merchant, the expected number of orders set by the merchant, and the historical turnover of the merchant, etc. The spatial distribution of the physical business distribution corresponding to the merchant entity includes: the corresponding order volume in the area where the merchant is located, the difficulty of picking up and delivering goods corresponding to the merchant, and the location outline corresponding to the merchant's location, etc., which are not limited. It is worth noting that the entity business distribution and entity location relationship corresponding to one or more target type entities are selected as the environment dataset.

Optionally, the entity business distribution corresponding to each target type entity in the environment data set is the same or different, which is not limited here.

Step 303: Perform location analysis on at least two target type entities in combination with entity business distribution and entity location relationship to obtain candidate centers within the candidate area.

In some embodiments, the candidate center refers to the position center corresponding to each divided area when the area is divided in the current area.

Optionally, the candidate center includes selecting one or more target type entities from at least two target type entities as candidate centers; or, according to the geographic locations corresponding to the at least two target type entities, at least one candidate center within the scope of the candidate area is determined. (or no candidate center is selected), that is, the candidate center is a certain target type entity, or a certain geographic location between multiple target type entities, which is not limited here.

In some embodiments, the location analysis refers to analyzing the adaptation results of at least two target-type entities as candidate centers when performing region division.

Illustratively, the location analysis method includes at least one of the following methods:

1. Establish a target strategy model, input the entity business distribution and entity location relationship into the target strategy model, and output the location analysis results corresponding to each target type entity to determine the candidate center. The target strategy model is used for location analysis of target type entities. model;

2. Divide the range of candidate areas in advance, determine the area characteristics corresponding to each candidate area, extract the location characteristics of the entity location relationship corresponding to the target type entity, and the parameter characteristics corresponding to the entity business distribution, and combine the location characteristics and parameters corresponding to the target type entity. The features are fused to obtain the fusion features corresponding to the target type entities, and the correlation degree analysis is carried out according to the fusion features corresponding to each target type entity and the regional features to obtain the central correlation degree corresponding to each target type entity, and the corresponding correlation degree with the candidate area according to the central correlation degree The target type entity with the highest degree is used as the candidate center. Subsequently, the division of the candidate region is adjusted based on the candidate center.

It should be noted that the above position analysis method is only a schematic example, which is not limited in this embodiment of the present application.

In some embodiments, the environment data set is input into the target strategy model, and the output is to obtain candidate centers. The target strategy model is used to perform location analysis on at least two target type entities in combination with entity business distribution and entity location relationship.

Optionally, the method of performing location analysis by using the target policy model includes at least one of the following methods:

1. After inputting the environmental data set into the target strategy model, the target strategy model performs location analysis according to the entity business distribution corresponding to each target type entity, obtains the location analysis score corresponding to each target type entity, and selects the location analysis score with the highest score. at least one target type entity as candidate center;

2. The parameter threshold is preset in the target strategy model. After inputting the environmental data set into the target strategy model, the target strategy model compares the entity business distribution corresponding to each target type entity with the parameter threshold, and the entity business distribution that reaches the parameter threshold corresponds to the The target type entity is used as the candidate center.

It should be noted that the above-mentioned method of performing location analysis by using the target strategy model is only a schematic example, which is not limited in the embodiment of the present application.

Step 304 , perform regional division on the candidate region based on the candidate center, and obtain regional planning results of at least two target type entities within the candidate region.

In some embodiments, the area division refers to dividing the candidate area into at least one sub-area, and each sub-area includes an area center as the center position corresponding to the sub-area.

Illustratively, the area planning result is used to indicate the division result corresponding to at least one sub-area within the range of the candidate area.

Optionally, the candidate center is used as the regional center, and the candidate area range is divided according to the regional center; or, according to the candidate center, the regional center corresponding to the candidate area range is determined, so as to perform regional division on the candidate area range, that is, the candidate center It may be an entity of a target type, or it may be a geographic location coordinate, which is not limited here.

Optionally, within the scope of the candidate area, there is an overlapping portion between the planning areas corresponding to the regional planning result, or the planning areas are closely fitted and have no overlapping portion, which is not limited.

It should be noted that the information (including but not limited to user equipment information, user personal information, etc.), data (including but not limited to data for analysis, stored data, displayed data, etc.) and signals involved in this application, All are authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data need to comply with the relevant laws, regulations and standards of relevant countries and regions. For example, the entity location relationship, entity business distribution, regional center and other related entity data corresponding to the target type entity involved in this application are all obtained with full authorization.

To sum up, in the regional planning method provided by the embodiments of the present application, by analyzing the geographic location coordinates of the target type entities in the geographical location, the entity location relationship and entity business distribution corresponding to each target type entity are determined, and the target type entities are analyzed. In the process of location analysis, the entity business distribution and entity location relationship are combined, so as to use the candidate center for dividing the range of candidate regions, and finally obtain the regional planning result, that is, by determining the entity location relationship and entity business distribution corresponding to the target type entity, Therefore, in the process of determining the candidate center in the regional planning, the relevant business information corresponding to the entity of the target type can be combined, and the regional planning scheme under different business requirements can be satisfied, thereby improving the accuracy of the regional planning.

In an optional embodiment, please refer to FIG. 4 , which shows a flowchart of a regional planning method provided by an exemplary embodiment of the present application. The method is described by taking the method being executed by the server as an example. That is, on the basis of the embodiment shown in FIG. 3 , step 303 further includes step 3031:

Step 301, obtaining a geographic location set.

The geographic location set includes geographic location coordinates corresponding to at least two target type entities within the range of the candidate area.

Illustratively, the candidate area range refers to a designated area corresponding to the current need for area planning.

Illustratively, the regional planning includes at least one of the following planning contents:

1. Divide the designated area. For example: divide a city into regions, and divide the city into a specified number of different regions according to business needs (such as: regional planning for takeaway delivery, regional planning for express transportation, etc.) for business management of different regions;

2. Route planning for the designated area. For example, for route planning in a park, by dividing the park into several different areas, the route laying plan is determined according to the area outline corresponding to each area;

3. Assemble the designated area. For example, the design of the transportation network between provinces and cities, through the planning of multiple cities in a certain province, determine the aggregated results corresponding to the cities, which are used to construct the transportation network.

It should be noted that the above planning content about regional planning is only a schematic example, which is not limited in this embodiment of the present application.

Optionally, the geographic location coordinates of the target type entity are plane coordinates, or, the geographic location coordinates of the target type entity are three-dimensional coordinates; the geographic location coordinates of the target type entity include those determined according to the range of the candidate location area, or, the The geographic location coordinates are the latitude and longitude coordinates corresponding to the entity, and are not limited here.

Step 302: Determine environment data sets corresponding to at least two target type entities based on the geographic location coordinates.

The environment data set is used to indicate the entity location relationship and entity service distribution corresponding to at least two target type entities, and the entity service distribution is used to indicate the spatiotemporal distribution of services of the target type entities.

Illustratively, the manner of determining the entity business distribution includes at least one of the following manners:

1. Obtain the historical data corresponding to the target type entity as the entity business distribution of the target type entity;

2. Preset the time range, and obtain the historical data corresponding to the target type entity within the time range as the entity business distribution;

3. After determining the parameter type of the target type entity, a parameter threshold is preset for the parameter as the entity business distribution corresponding to the target type entity, that is, the expected value of the target type entity corresponding to the parameter type.

It should be noted that the above-mentioned determination method of entity service distribution is only a schematic example, which is not limited in this embodiment of the present application.

Step 3031: Input the environment data set into the target policy model, and output the candidate center.

The target strategy model is used to perform location analysis on at least two target type entities in combination with entity business distribution and entity location relationship.

In some embodiments, the at least two target type entities include a merchant entity, the context data set is used to indicate the entity location relationship of the merchant entity, and the business distribution of the merchant.

Optionally, in an achievable scenario, the target type entity is implemented as a merchant entity, such as a merchant store, etc., and the environment data set includes the entity location relationship corresponding to the merchant entity and the business distribution corresponding to the merchant.

Wherein, the entity location relationship corresponding to the business entity includes the corresponding route distance between two business entities; or, the location distribution relationship corresponding to multiple business entities within the candidate area, which is not limited herein.

Among them, the business distribution corresponding to the merchant refers to the relevant business distribution information contained by the merchant under the specified business type. For example, when the merchant is in the takeaway business scenario, the business distribution corresponding to the merchant includes the number of takeaway orders of the merchant, the difficulty of picking up meals corresponding to the merchant, At least one of the business distribution information such as the difficulty of picking up meals in the designated area where the merchant is located, the outline of the area corresponding to the merchant's location, and the order expectations preset by the merchant manager for the merchant.

In some embodiments, based on the environmental data set, the historical regional center of the candidate area is determined, and the historical regional center includes the business entity as the divided regional center in the historical time period; through the target strategy model, the historical regional center and the environmental data are analyzed. The set is analyzed, and the output is the candidate center.

Illustratively, in the environmental data set, it also includes the corresponding regional center selection status of at least two business entities in the historical time period, that is, each business entity contains a regional center selection status: selected as a regional center or not. Selected as area center.

Among them, being selected as the regional center means that the business entity has been selected as the regional center one or more times during the process of regional planning for the candidate area within the historical time period; not selected as the regional center refers to the business entity in the historical time period. In the process of regional planning for the candidate region range in the segment, it is not selected as the historical record of the regional center. Therefore, the target type entity selected as the regional center will be the historical regional center.

In some embodiments, the historical time period includes a certain historical time range, or includes at least one historical area planning process, which is not limited herein.

Optionally, in the current regional planning process, one or a historical regional center may be determined according to the environmental data set, or there is no historical regional center, which is not limited herein.

In this embodiment, the environment dataset is used as Environment, and the historical regional center is the State corresponding to the merchant entity, that is, the current merchant entity corresponds to the state included in the historical time period as the regional center.

The environment data sets corresponding to at least two business entities and the historical regional center are input into the target strategy model to analyze the location of the business entity, and the business entity as the candidate center is obtained as output.

In some embodiments, the target policy model includes an actor network in a reinforcement learning architecture, the actor network is used for center prediction, and the actor network includes an encoder and a decoder in an encoder-decoder architecture.

In this embodiment, an actor network is included in the reinforcement learning architecture, and center prediction refers to analyzing whether the target type entity can be used as a candidate center. The actor network is used to predict and analyze the relevant data of the target type entity input into the target strategy model, and determine it as a candidate center. The object type entity of the center, where the actor network implements the center prediction function through the encoder and decoder in the encoder-decoder architecture.

In some embodiments, based on the environment data set, determine the historical area center of the candidate area range, the historical area center includes the target type entity as the divided area center in the historical time period; input the historical area center and the environment data set into the actor network The encoder outputs the encoding feature and the center feature set; decode and analyze the encoded feature through the decoder to obtain the decoded feature; obtain the candidate center based on the decoded feature and the center feature set.

In this embodiment, the center selection state of each target type entity within the historical time range is determined through the environment data set, and the target type entity whose center selection state is selected as a candidate center is determined as a historical regional center. Input the environmental data set and the historical regional center into the target strategy model. The target strategy model combines the entity business distribution corresponding to the target type entity in the environmental data set and the historical regional center, and analyzes the location of at least two target type entities to determine the final candidate center. The target type entity.

Illustratively, please refer to FIG. 5, which shows a schematic diagram of the actor network analysis process provided by an exemplary embodiment of the present application. As shown in FIG. 5, the center selection state 510 and the geographic location coordinates 520 corresponding to the target type entity are input. The encoder 530, wherein the center selection state 510 is used to indicate whether the target type entity is selected as a candidate center in the historical time period, that is, the center selection state 510 includes two states, selected and unselected, and the value range of the center selection state 510 It is [0, t], 0 indicates that the target type entity was not selected as the candidate center in the historical time period, t indicates that it was selected as the candidate center in the t-th selection process, and the geographic location coordinate 520 is the target type entity in the candidate center. 2D plane coordinates (lng, lat) within the area.

The encoder 530 includes two neural networks (531 and 532), the neural network 531 is used to extract the state feature 541 corresponding to the center selection 510, and the neural network 532 is used to extract the location feature 542 corresponding to the geographic location coordinates 520. After selecting the state 510 and the geographic location coordinates 520 and inputting the neural network 531 and the neural network 532 at the same time, the state feature 541 and the position feature 542 corresponding to each target type entity are obtained, and the state feature representation 541 and the position feature representation 542 corresponding to each target type entity are obtained. as the central feature set 540.

The target policy model also includes a decoder, and the decoder includes a neural network 550 . The coding feature 543 corresponding to the historical regional center is determined by the neural network 532. In addition, according to the entity location relationship and the entity business distribution corresponding to the target type entity, the business feature 544 corresponding to the historical regional center is determined, and the coding feature 543 and the business feature 544 are input into the neural network. The network 550 outputs the decoded feature 551. The decoded feature 551 is used to indicate the feature representation corresponding to the candidate center 560. The center feature set 540 and the decoded feature 551 are used to determine the final candidate center 560 through the attention mechanism.

In this embodiment, the neural network 531 and the neural network 532 are implemented as a network with a convolutional neural network (Convolutional Neural Networks, CNN) structure, and the neural network 550 is implemented as a network with a long short-term memory (Long Short-Term Memory, LSTM) structure, This application does not limit this.

Step 304 , perform regional division on the candidate region based on the candidate center, and obtain regional planning results of at least two target type entities within the candidate region.

Illustratively, in the process of determining the regional center according to the candidate center, the determination method includes:

1. Take the candidate center output from the target strategy model as the regional center;

2. When there are m candidate centers in the same planning area, connect the position points corresponding to the m candidate centers, and use the intersection obtained by connecting the lines as the area center, where m is a positive integer;

3. After the target policy model obtains the candidate centers, use the weighted centroid algorithm to traverse the candidate centers to determine the final regional center.

It should be noted that the above-mentioned determination method of the regional center is only a schematic example, which is not limited in this embodiment of the present application.

Illustratively, the manner of determining the regional planning result includes at least one of the following manners:

1. Preset the specified number, after determining the regional center, take the regional center as the benchmark, and take the area corresponding to the specified number of target type entities as the planned area;

2. Preset the area outline, after determining the area center, take the area center as the benchmark, and determine the planning area corresponding to the area center according to the preset area outline;

3. Determine the results of regional planning according to business needs. For example, when planning for businesses, distribute businesses with the same type of content in different planning areas, and place businesses with business connections (such as movie theaters and restaurants) with in the same planning area to meet the final business needs.

It should be noted that the above content about the regional planning result is only a schematic example, and the comparison of the embodiments of the present application is not limited.

Illustratively, the regional planning result includes at least one of the following results:

1. The regional planning result is to display each planning area divided according to the regional center within the scope of the candidate area, that is, to display each planning area within the scope of the candidate area;

2. The regional planning result is to display the regional outline corresponding to each planned area within the candidate area, that is, to display the graphic shape and corresponding contour line corresponding to each planned area within the candidate area;

3. The regional planning result is to display the clustering results corresponding to at least two target-type entities within the candidate region, that is, to display at least one clustering result corresponding to the target-type entity within the candidate region.

It should be noted that the above content about the regional planning result is only a schematic example, which is not limited in this embodiment of the present application.

In some embodiments, centroid positioning is performed on the candidate center to obtain the area center corresponding to the candidate area range; based on the area center, entity clustering results corresponding to at least two target type entities in the candidate area range are determined; as a result of regional planning.

Illustratively, after obtaining candidate centers corresponding to at least two target-type entities, at least one regional center is determined from the candidate centers by a weighted centroid algorithm, and a maximum expectation algorithm (Expectation-MximizationAgorithm, EM algorithm) is used to calculate the distance between the target-type entities and their distances. The nearest regional center is clustered to obtain the entity clustering results corresponding to each target type entity in the candidate area. Among them, one target type entity is clustered with one or more regional centers. According to the entity clustering results, a graph outsourcing algorithm is used. Determine the final regional planning result, wherein, using the graph outsourcing algorithm refers to taking the polygon outline corresponding to the geographic location of the target type entity as the basic unit, and determining the outsourcing outline of the planning area corresponding to the center of each area through the basic unit, that is, the current candidate Generate at least one outsourcing drawing within the area as a result of area planning.

Illustratively, please refer to FIG. 6, which shows a schematic diagram of a regional planning result provided by an exemplary embodiment of the present application. As shown in FIG. 6, FIG. 6 shows a regional planning result interface 600. In the regional planning result interface 600 At least one planning area 601 is included (the area delineated by the bold line in FIG. 6 is exemplified), and the outer contour of the planning area 601 refers to the contour line corresponding to the range where the planning area 601 is located.

To sum up, in the regional planning method provided by the embodiments of the present application, by analyzing the geographic location coordinates of the target type entities in the geographical location, the entity location relationship and entity business distribution corresponding to each target type entity are determined, and the target type entities are analyzed. In the process of location analysis, the entity business distribution and entity location relationship are combined, so as to use the candidate center for dividing the range of candidate regions, and finally obtain the regional planning result, that is, by determining the entity location relationship and entity business distribution corresponding to the target type entity, Therefore, in the process of determining the candidate center in the regional planning, the relevant business information corresponding to the entity of the target type can be combined, and the regional planning scheme under different business requirements can be satisfied, thereby improving the accuracy of the regional planning.

In this embodiment, the historical regional center is determined by the environmental data set, and the historical regional center and the environmental data set are jointly input into the target strategy model for location analysis of the target type entity, and the analysis is carried out in combination with the center selection state of the target type entity , which can improve the analysis accuracy of the target strategy, thereby improving the accuracy of regional planning.

In an optional embodiment, please refer to FIG. 7 , which shows a flowchart of a regional planning method provided by an exemplary embodiment of the present application. This embodiment is used to describe the training side of the area planning method, and the method is executed by a server as an example for description. That is, on the basis of the embodiment shown in FIG. 3 , before step 304, the following steps are further included:

Step 701, obtaining a sample data set.

The sample data set is used to indicate the sample location relationship and sample business distribution corresponding to at least two sample type entities.

In some embodiments, the sample type entities include: businesses, buildings, traffic locations, etc., wherein at least two sample type entities are located within the sample candidate area.

Wherein, when the sample type entity is a business, the geographic location coordinates corresponding to the sample type entity are the store location coordinates corresponding to the business; when the sample type entity is a building, the geographic location coordinates corresponding to the sample type entity are the location coordinates corresponding to the building ; When the sample type entity is a traffic location, the geographic location coordinates corresponding to the sample type entity are the location coordinates within the sample candidate area corresponding to the traffic location.

Optionally, the sample location relationship includes a route distance between two sample type entities; or, the distribution relationship of the corresponding locations of multiple sample type entities; or, the location distribution state of multiple sample type entities within the sample candidate area. , which is not limited here, wherein the sample location relationship is determined by the respective geographic location coordinates corresponding to at least two sample type entities.

In some embodiments, the sample business distribution refers to the relevant business parameters corresponding to a single target type entity. According to different business requirements, the sample business distribution is also different. The business needs, the sample business distribution includes the number of historical orders of the merchant in the historical time period, the business type of the merchant, the business hours, the difficulty of picking up and delivering the merchant, the expected number of orders of the merchant, and the corresponding regions where multiple merchants are located. The area outline, etc., are not limited here. Wherein, in the process of regional planning, one or more sample business distributions are selected as the sample data set.

Illustratively, the geographic location coordinates corresponding to the sample type entity are plane location coordinates.

Step 702: Predict the sample data set through the actor network to obtain sample candidate centers of at least two sample-type entities.

Illustratively, first determine the sample historical center corresponding to the sample data set, wherein, before the first training, the number of centers is preset, and a random selection method is used to determine the sample type entity corresponding to the preset number of centers in the sample data set as the sample The candidate center is used as the sample history center for the input actor network in the first training process. In the first training process, the sample data set and the sample history center are input into the actor network, and the output is generated corresponding to the first training. The sample candidate center of , as the sample historical center in the second training process.

Starting from the second training, in each training process, the sample candidate center obtained from the previous training is used as the corresponding sample historical center in this training process.

Step 703: Evaluate the sample candidate centers through the critic network in the reinforcement learning architecture to obtain the expected deviation.

Among them, the expected deviation is used to indicate the prediction effect of the sample candidate center.

In this embodiment, the reinforcement learning architecture further includes a critic network, and the critic network is used to generate the predicted value of the corresponding result of the estimated sample candidate center.

First, the sample candidate center is substituted into a preset expectation function to obtain a first expected value corresponding to the sample candidate center, where the first expected value is used to indicate the aggregation relationship between the sample candidate centers.

Illustratively, the preset expectation function is a preset function determined according to the sample service type corresponding to the sample type entity, and is used to determine the first expected value corresponding to the sample candidate center.

In this embodiment, in the scenario of planning a merchant delivery area, please refer to Formula 1 for the method of determining the first expected value:

Formula 1: Rt=α·sum of orders within a circle with a certain radius +β·distance between two merchants at the center point+γ·(1-regional average delivery difficulty)+δ·(KL (regional preset order distribution, merchants Estimated single-quantity distribution)

Among them, α, β, γ, and δ correspond to different adjustable weights, and the sum of the unit quantities within a certain radius circle refers to the order corresponding to the sample type entity (merchant) in the sample candidate planning area (within a certain radius circle) The sum of the quantity, the distance between the two merchants of the center point refers to the corresponding route distance between the two sample candidate centers, the regional average distribution difficulty refers to the average value of the corresponding distribution difficulty in the sample candidate planning area, KL (regional preset order distribution, Merchant's estimated single-volume distribution) refers to the relative entropy between the business manager's preset single-volume distribution and the merchant's estimated single-volume distribution, where the business manager's preset single-volume distribution refers to the planning results corresponding to the regional planning results. The distribution of the preset order volume corresponding to the region, the merchant's estimated order volume distribution refers to the expected order volume estimated by each merchant.

Among them, assign different weights to each sample business distribution and sample location relationship corresponding to each sample candidate center, and calculate the weighted sum result of each sample business distribution and sample location relationship as the first expected value corresponding to the sample candidate center cluster, where , the sample candidate center cluster refers to the center set corresponding to the sample candidate center.

It is worth noting that the parameters used to determine the first expected value in the above formula 1 are only illustrative examples, and different parameters can be set according to actual requirements in the application process.

Illustratively, the aggregation relationship refers to the degree of dispersion of the positions of the sample candidate centers within the regional planning range.

Secondly, an expectation analysis is performed on the sample data set through the critic network, and a second expectation value corresponding to the sample data set is obtained, and the second expectation value is used to indicate the estimated aggregation relationship corresponding to the sample data set.

In some embodiments, the sample data set is input into a candidate prediction model, the candidate prediction model includes a critic network, and the candidate prediction model is a machine learning model obtained by pre-training; the sample data set is expected to be predicted through the critic network. Estimate, the output obtains the second expected value.

In this embodiment, a candidate prediction model including a critic network is constructed, and the second expected value corresponding to the sample data set is obtained through the output of the critic network, wherein the critic network in the candidate prediction model can be implemented as a two-layer CNN network structure and a two-layer fully connected structure, and the output result is a scalar network structure, which is used to represent the estimated value corresponding to the first expected value, that is, the prediction result of the aggregation relationship corresponding to the sample data set.

Finally, the expected deviation is obtained based on the first expected value and the second expected value.

In some embodiments, a gradient relationship corresponding to the first expected value and the second expected value is determined; the gradient relationship between the first expected value and the second expected value is determined as the expected deviation.

Illustratively, the parameter adjustment direction of the actor network is determined by analyzing the gradient relationship between the first expected value and the second expected value, wherein the specific method of determining the gradient relationship can refer to formula 2:

Formula two:

指梯度权重。Among them, ω(Z ⁱ |G) refers to the first expected value, b(G) refers to the second expected value,

refers to the gradient weight.

Step 704: Train the actor network based on the expected deviation to obtain a target policy model.

In this embodiment, the expected deviation corresponding to the first expected value and the second expected value is determined by the above formula 2. When the first expected value is closer to the second expected value, it means that the parameter accuracy of the actor network is higher. When the first expected value is greater than the second expected value When the expected value is two, the gradient weight is set to the positive direction of the first expected value, and the training direction of the actor network is adjusted; when the first expected value is less than the second expected value, the gradient weight is set to the negative direction of the first expected value, and the training direction of the actor network is adjusted. Finally, the target policy model is obtained.

Step 705 , based on the first expected value and the second expected value, adjust the model parameters of the candidate prediction model to obtain the expected prediction model.

Wherein, the expectation prediction model is used to determine the second expectation value.

Illustratively, by determining the mean square error value corresponding to the first expected value and the second expected value, the model parameters of the candidate prediction model are adjusted, wherein the calculation method of the mean square error value can refer to Formula 3:

Formula three:

Among them, ω(Z ⁱ |G) refers to the first expected value, b(G) refers to the second expected value, and L _c represents the mean square error value. The smaller the mean square error value, the higher the accuracy of the model parameters of the candidate prediction model.

It is worth noting that in the process of training the actor network, the critic network is also being trained, and only the actor network is involved in the final target policy model.

Illustratively, please refer to FIG. 8 , which shows a schematic diagram of the training process of the target policy model provided by an exemplary embodiment of the present application. As shown in FIG. 8 , a sample data set 810 is obtained, and the sample data set 810 includes at least two The geographic location coordinates corresponding to the sample type entity (t sample data sets required for t training are shown in the figure), the sample historical center corresponding to the sample type entity is determined according to the sample data set 810, and the sample data set is determined according to the sample historical center The center selection state 820 corresponding to each sample type entity in 810, that is, the center selection state 820 includes the sample type entity selected as the sample historical center, and the sample type entity not selected as the sample historical center, wherein the selected There are two different states corresponding to not being selected.

Extract the parameter features corresponding to the sample business distribution of the sample type entities in the sample data set 810, and input the sample position relationship, parameter features, and center selection status 820 in the sample data set 810 into the actor network 830 (the input of the sample position relationship and parameter features). The process is not shown), and the output obtains the sample candidate center 840 .

Aggregate the sample candidate centers output by the actor network to generate sample candidate clusters, combine the sample business distribution and sample location relationship corresponding to the sample candidate centers, perform aggregation degree analysis on the sample candidate clusters, and obtain the aggregation analysis results corresponding to the sample candidate clusters. as the first expected value. As shown in FIG. 8 , the corresponding sample candidate cluster 850 is determined by the sample candidate center 840 , and the first expected value thereof is determined according to the sample candidate cluster 850 .

Input the sample data set 810 into the candidate estimation model 860, and output the second expected value, wherein the expected estimation of the sample data set 810 is performed by the critic network included in the candidate estimation model, and the obtained estimation result is output as the second expected value. expected value. The expected deviation is determined by the first expected value and the second expected value, which is used to train the actor network 830 to finally obtain the target policy model.

To sum up, in the regional planning method provided by the embodiments of the present application, by analyzing the geographic location coordinates of the target type entities in the geographical location, the entity location relationship and entity business distribution corresponding to each target type entity are determined, and the target type entities are analyzed. In the process of location analysis, the entity business distribution and entity location relationship are combined, so as to use the candidate center for dividing the range of candidate regions, and finally obtain the regional planning result, that is, by determining the entity location relationship and entity business distribution corresponding to the target type entity, Therefore, in the process of determining the candidate center in the regional planning, the relevant business information corresponding to the entity of the target type can be combined, and the regional planning scheme under different business requirements can be satisfied, thereby improving the accuracy of the regional planning.

In this embodiment, the critic network is trained by adding a candidate prediction model in the training process, and the critic network is also trained in the process of training the actor network, which can improve the parameter accuracy of the actor network, and the obtained target strategy model Accuracy is also higher.

Illustratively, please refer to FIG. 9 , which shows a flowchart of a regional planning method provided by an exemplary embodiment of the present application. That is, FIG. 9 is a detailed description of the training of the regional planning method in the merchant delivery scenario, and the method includes the following steps:

910, data preparation.

In this embodiment, the set of merchants in a city is used as the set of location points on a two-dimensional plane, that is, the training process is to carry out regional planning for the merchants in a city, and a single merchant is used as a sample type entity to obtain the corresponding information of each merchant. geographic coordinates.

Illustratively, according to the geographic location coordinates corresponding to each business, a sample data set corresponding to each business in the city is determined, wherein the sample data set includes a sample geographic location relationship and a sample business distribution corresponding to the business, and the sample geographic location relationship includes two. Corresponding route distance between merchants, sample business distribution of the number of orders of each merchant in the historical time period, the final outsourcing outline corresponding to the planning area within the regional planning range determined by the distribution business, the distribution difficulty corresponding to the merchant, the management’s opinion on the merchant The preset expected quantity of orders, etc., is not limited here, and the feature representation corresponding to each data included in the sample data set is determined as the environmental feature representation set (G).

The number B of samples corresponding to a single training process and the number of preset training times E are preset to determine the number N corresponding to the planning area when planning the city.

920, the model is initialized.

In this embodiment, in the process of starting training, initialize various parameters, including initializing the environmental feature set (G), initializing the actor network (θ) and the critic network (θ _c ), and determining the corresponding merchants. The parameter characteristics of , the number of planning areas and the second expected value in different states. Among them, the critic network is the network included in the expectation prediction model.

Among them, the actor network includes two convolutional neural networks and a long short-term memory network, and the actor network includes two layers of convolutional neural networks and two fully connected layers.

930. Determine the sample candidate center and the first expected value.

In this embodiment, the training process includes determining the corresponding sample candidate center for each sample in each training process. Determine the corresponding merchants in the city as sample historical centers, mark them as selected, and mark the rest of the merchants as unselected, and use them as central features.

The center feature and the environment feature are input into the actor network, and the sample candidate center (Z ⁱ |G) is obtained from the output, and the first expected value (ω(Z ⁱ |G)) corresponding to the sample candidate center is determined according to formula 1.

The sample candidate center and the first expected value corresponding to the sample candidate center are stored.

940. Determine a second expected value corresponding to the sample data set.

After inputting the sample data set corresponding to each merchant into the critic network, the output obtains the second expected value (b|G) corresponding to the sample data set.

950, the model parameters are adjusted and updated.

After the first expected value and the second expected value are determined, the gradient relationship corresponding to the first expected value and the second expected value is determined according to formula 2, the actor network is adjusted according to the gradient relationship, and the parameters are updated to obtain the target strategy model.

According to formula 3, the mean square error value corresponding to the first expected value and the second expected value is determined, the parameters of the candidate prediction model are adjusted, and the parameters are updated.

To sum up, in the regional planning method provided by the embodiments of the present application, by analyzing the geographic location coordinates of the target type entities in the geographical location, the entity location relationship and entity business distribution corresponding to each target type entity are determined, and the target type entities are analyzed. In the process of location analysis, the entity business distribution and entity location relationship are combined, so as to use the candidate center for dividing the range of candidate regions, and finally obtain the regional planning result, that is, by determining the entity location relationship and entity business distribution corresponding to the target type entity, Therefore, in the process of determining the candidate center in the regional planning, the relevant business information corresponding to the entity of the target type can be combined, and the regional planning scheme under different business requirements can be satisfied, thereby improving the accuracy of the regional planning.

The beneficial effects of this program also include:

1. The method of determining candidate centers by clustering entities is transformed into a sequential decision-making scheme, and the training result is judged only according to the historical state;

2. The way of determining the second expected value can integrate different business needs, better combine the actual situation, and adapt to the development of the business;

3. The model training efficiency can be improved by constructing a training model.

FIG. 10 is a structural block diagram of an apparatus for regional planning provided by an exemplary embodiment of the present application. As shown in FIG. 10 , the apparatus includes:

The obtaining module 1010 is configured to obtain a geographic location set, the geographic location set includes geographic location coordinates corresponding to at least two target type entities within the scope of the candidate area:

The determining module 1020 is configured to determine, based on the geographic location coordinates, an environmental data set corresponding to the at least two target type entities, where the environmental data set is used to indicate the entity location relationship and the corresponding entity position relationship of the at least two target type entities. Entity business distribution The entity business distribution is used to indicate the distribution situation of the business of the target type entity in space and time;

An analysis module 1030, configured to perform location analysis on the at least two target type entities in combination with the entity business distribution and the entity location relationship to obtain a candidate center within the candidate area;

The division module 1040 is further configured to perform area division on the candidate area range based on the candidate center, to obtain a regional planning result of the at least two target type entities within the candidate area range.

In an optional embodiment, the analysis module 1030 includes:

The output unit 1031 is configured to input the environment data set into a target strategy model, and output to obtain the candidate center, and the target strategy model is used to combine the entity business distribution and the entity location relationship with the at least two targets. Type entity for location analysis.

In an optional embodiment, the at least two target type entities include a merchant entity, and the environment data set is used to indicate the entity location relationship of the merchant entity and the business distribution of the merchant;

The output unit 1031 is further configured to determine the historical area center of the candidate area based on the environment data set, where the historical area center includes the business entity as the divided area center in the historical time period; The target strategy model analyzes the historical regional center and the environmental data set, and outputs the candidate center.

In an optional embodiment, the target policy model includes an actor network in a reinforcement learning architecture, the actor network is used for center prediction, and the actor network includes an encoder and a decoder in an encoder-decoder architecture ;

The output unit 1031 is further configured to determine the historical area center of the candidate area based on the environmental data set, where the historical area center includes the target type entity as the divided area center in the historical time period; The historical area center and the environmental data set are input into the encoder of the actor network, and the output obtains the encoding feature and the center feature set; the encoding feature is decoded and analyzed by the decoder to obtain the decoding feature; The candidate center is obtained from the decoded feature and the center feature set.

In an optional embodiment, the target policy model includes an actor network in a reinforcement learning architecture, and the actor network is used for center prediction;

The device also includes:

The obtaining module 1010 is further configured to obtain a sample data set, where the sample data set is used to indicate the sample location relationship and sample business distribution corresponding to at least two sample type entities;

A prediction module 1050, configured to predict the sample data set through the actor network to obtain sample candidate centers of the at least two sample type entities;

An evaluation module 1060, configured to evaluate the sample candidate center through the critic network in the reinforcement learning architecture to obtain an expected deviation, where the expected deviation is used to indicate the prediction effect of the sample candidate center;

A training module 1070, configured to train the actor network based on the expected deviation to obtain the target policy model.

In an optional embodiment, the evaluation module 1060 is further configured to substitute the sample candidate center into a preset expectation function to obtain a first expected value corresponding to the sample candidate center, where the first expected value is used to indicate The aggregation relationship between the sample candidate centers; the expectation analysis is performed on the sample data set through the critic network, and the second expectation value corresponding to the sample data set is obtained, and the second expectation value is used to indicate the sample The estimated aggregation relationship corresponding to the data set; the expected deviation is obtained based on the first expected value and the second expected value.

In an optional embodiment, the evaluation module 1060 is further configured to determine a gradient relationship corresponding to the first expected value and the second expected value; A gradient relationship is determined as the desired deviation.

In an optional embodiment, the evaluation module 1060 is further configured to input the sample data set into a candidate prediction model, where the candidate prediction model includes the critic network, and the candidate prediction model is a machine learning model obtained by pre-training; the sample data set is expected to be estimated through the critic network, and the second expected value is obtained by outputting.

In an optional embodiment, the apparatus further includes:

The adjustment module 1080 is configured to adjust the model parameters of the candidate prediction model based on the first expected value and the second expected value to obtain an expected prediction model, and the expected prediction model is used to determine the first prediction model. Two expectations.

In an optional embodiment, the analysis module 1030 is further configured to perform centroid location on the candidate center to obtain the area center corresponding to the candidate area range; based on the area center, determine the at least two The entity clustering result corresponding to the target type entity in the candidate area; the entity clustering result is used as the area planning result.

To sum up, the regional planning device provided by the embodiment of the present application determines the entity location relationship and entity business distribution corresponding to each target type entity by analyzing the geographic location coordinates of the target type entities in the geographical location concentration, and performs the analysis on the target type entities. In the process of location analysis, the entity business distribution and entity location relationship are combined, so as to use the candidate center for dividing the range of candidate regions, and finally obtain the regional planning result, that is, by determining the entity location relationship and entity business distribution corresponding to the target type entity, Therefore, other parameter information of the target type entity other than the geographic location coordinates can be combined in the process of determining the candidate center in the regional planning, thereby improving the accuracy of the regional planning.

It should be noted that: the regional planning device provided in the above-mentioned embodiments is only illustrated by the division of the above-mentioned functional modules. In practical applications, the above-mentioned functions can be allocated to different functional modules according to needs. It is divided into different functional modules to complete all or part of the functions described above. In addition, the regional planning apparatus and the regional planning method embodiments provided by the above embodiments belong to the same concept, and the specific implementation process thereof is detailed in the method embodiments, which will not be repeated here.

Please refer to FIG. 12 , which shows a structural block diagram of a computer device 1200 provided by an embodiment of the present application. The computer device 1200 may be the electronic device described above, for implementing the above-mentioned regional planning method.

Generally, computer device 1200 includes: processor 1201 and memory 1202 .

The processor 1201 may include one or more processing cores, such as a 4-core processor, an 8-core processor, and the like. The processor 1201 may be implemented by at least one hardware form among DSP (Digital Signal Processing, digital signal processing), FPGA (FieldProgrammable Gate Array, field programmable gate array), and PLA (Programmable Logic Array, programmable logic array). The processor 1201 may also include a main processor and a coprocessor. The main processor is a processor used to process data in a wake-up state, also called a CPU (Central Processing Unit, central processing unit); A low-power processor for processing data in a standby state. In some embodiments, the processor 1201 may be integrated with a GPU (Graphics Processing Unit, image processor), and the GPU is used for rendering and drawing the content that needs to be displayed on the display screen. In some embodiments, the processor 1201 may further include an AI (Artificial Intelligence, artificial intelligence) processor, where the AI processor is used to process computing operations related to machine learning.

Memory 1202 may include one or more computer-readable storage media, which may be non-transitory. Memory 1202 may also include high-speed random access memory, as well as non-volatile memory, such as one or more disk storage devices, flash storage devices.

Those skilled in the art can understand that the structure shown in FIG. 12 does not constitute a limitation on the computer device 1200, and may include more or less components than the one shown, or combine some components, or adopt different component arrangements.

In an example embodiment, there is also provided a computer device comprising a processor and a memory having a computer program stored in the memory. The computer program is configured to be executed by one or more processors to implement the area planning method described above.

In an exemplary embodiment, a computer-readable storage medium is also provided, and a computer program is stored in the storage medium, and the computer program implements the above-mentioned regional planning method when executed by a processor of a computer device.

Optionally, the above-mentioned computer-readable storage medium can be ROM (Read-Only Memory, read-only memory), RAM (Random Access Memory, random access memory), CD-ROM (Compact Disc Read-Only Memory, read-only CD-ROM) , tapes, floppy disks and optical data storage devices.

In an exemplary embodiment, there is also provided a computer program product which, when executed on a computer device, causes the computer device to perform the area planning method as described above.

It should be understood that references herein to "a plurality" means two or more. "And/or", which describes the association relationship of the associated objects, means that there can be three kinds of relationships, for example, A and/or B, which can mean that A exists alone, A and B exist at the same time, and B exists alone. The character "/" generally indicates that the associated objects are an "or" relationship. In addition, the numbering of the steps described in this document only exemplarily shows a possible execution sequence between the steps. In some other embodiments, the above steps may also be executed in different order, such as two different numbers. The steps are performed at the same time, or two steps with different numbers are performed in a reverse order to that shown in the figure, which is not limited in this embodiment of the present application.

The above are only exemplary embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application shall be included in the protection of the present application. within the range.

Claims (14)

1. A method of area planning, the method comprising:

acquiring a geographical position set, wherein the geographical position set comprises geographical position coordinates corresponding to at least two target type entities in a candidate area range;

determining environment data sets corresponding to the at least two target type entities based on the geographic position coordinates, wherein the environment data sets are used for indicating entity position relations and entity service distributions corresponding to the at least two target type entities, and the entity service distributions are used for indicating the distribution conditions of the services of the target type entities on space and time;

performing position analysis on the at least two target type entities by combining the entity service distribution and the entity position relation to obtain a candidate center in the candidate area range;

and performing region division on the candidate region range based on the candidate center to obtain region planning results of the at least two target type entities in the candidate region range.

2. The method of claim 1, wherein the performing location analysis on the at least two target type entities in combination with the entity service distribution and the entity location relationship comprises:

and inputting the environment data set into a target strategy model, and outputting to obtain the candidate center, wherein the target strategy model is used for carrying out position analysis on the at least two target type entities by combining the entity service distribution and the entity position relation.

3. The method of claim 2, wherein the at least two target type entities comprise merchant entities, the set of environmental data is indicative of entity location relationships of the merchant entities, and business distributions of the merchant;

the inputting the environment data set into a target strategy model and outputting to obtain a candidate center comprises:

determining historical region centers of the candidate region ranges based on the environment data set, wherein the historical region centers comprise merchant entities which are divided into region centers in historical time periods;

and analyzing the historical region center and the environment data set through the target strategy model, and outputting to obtain the candidate center.

4. The method of claim 2, wherein the target strategy model comprises an actor network in a reinforcement learning architecture, the actor network being used for center prediction, the actor network comprising an encoder and a decoder in a codec architecture;

the inputting the environment data set into a target strategy model and outputting to obtain the candidate center comprises:

determining a historical region center of the candidate region range based on the environment data set, wherein the historical region center comprises a target type entity which is used as a divided region center in a historical time period;

inputting the historical region center and the environment data set into the encoder of the actor network, and outputting to obtain an encoding feature and a center feature set;

decoding and analyzing the coding characteristics through the decoder to obtain decoding characteristics;

and obtaining the candidate center based on the decoding characteristics and the center characteristic set.

5. The method of claim 2, wherein the target strategy model comprises an actor network in a reinforcement learning architecture, the actor network being used for center prediction;

before inputting the environment data set into a target policy model and outputting to obtain the candidate center, the method further includes:

acquiring a sample data set, wherein the sample data set is used for indicating the sample position relationship and the sample service distribution corresponding to at least two sample type entities;

predicting the sample data set through the actor network to obtain sample candidate centers of the at least two sample type entities;

evaluating the sample candidate center through a critic network in the reinforcement learning architecture to obtain an expected deviation, wherein the expected deviation is used for indicating the prediction effect of the sample candidate center;

and training the actor network based on the expected deviation to obtain the target strategy model.

6. The method of claim 5, wherein said evaluating said prediction center through a network of critics for expected deviation comprises:

substituting the sample candidate centers into a preset expectation function to obtain first expectation values corresponding to the sample candidate centers, wherein the first expectation values are used for indicating the aggregation relation among the sample candidate centers;

performing expected analysis on the sample data set through the critic network to obtain a second expected value corresponding to the sample data set, wherein the second expected value is used for indicating a pre-estimated aggregation relation corresponding to the sample data set;

the expected deviation is derived based on the first expected value and the second expected value.

7. The method of claim 6, wherein said deriving the expected deviation based on the first expected value and the second expected value comprises:

determining a gradient relation corresponding to the first expected value and the second expected value;

determining a gradient relationship between the first expected value and the second expected value as the expected deviation.

8. The method of claim 6, wherein said performing a desired analysis on the sample data set via the critic network to obtain a second desired value for the sample data set comprises:

inputting the sample data set into a candidate pre-estimation model, wherein the candidate pre-estimation model comprises the critic network and is a machine learning model obtained through pre-training;

and performing expected prediction on the sample data set through the critic network, and outputting to obtain the second expected value.

9. The method of claim 8, further comprising:

and adjusting model parameters of the candidate prediction model based on the first expectation value and the second expectation value to obtain an expected prediction model, wherein the expected prediction model is used for determining the second expectation value.

10. The method according to any one of claims 1 to 9, wherein the performing the area division on the candidate area range based on the candidate center to obtain the area planning result of the at least two target type entities in the candidate area range comprises:

carrying out mass center positioning on the candidate center to obtain a region center corresponding to the candidate region range;

determining entity clustering results corresponding to the at least two target type entities in the candidate region range based on the region center;

and taking the entity clustering result as the area planning result.

11. An area planning apparatus, the apparatus comprising:

the acquisition module is used for acquiring a geographical position set, and the geographical position set comprises geographical position coordinates corresponding to at least two target type entities in a candidate area range;

a determining module, configured to determine, based on the geographic position coordinates, environment data sets corresponding to the at least two target type entities, where the environment data sets are used to indicate entity position relationships and entity service distributions corresponding to the at least two target type entities, and the entity service distributions are used to indicate temporal and spatial distribution conditions of services of the target type entities;

the analysis module is used for carrying out position analysis on the at least two target type entities by combining the entity service distribution and the entity position relation to obtain a candidate center in the candidate area range;

and the dividing module is used for carrying out region division on the candidate region range based on the candidate center to obtain a region planning result of the at least two target type entities in the candidate region range.

12. A computer device comprising a processor and a memory, the memory having stored therein at least one instruction, at least one program, a set of codes, or a set of instructions, which is loaded and executed by the processor to implement a region planning method according to any one of claims 1 to 10.

13. A computer readable storage medium having stored therein at least one instruction, at least one program, a set of codes, or a set of instructions, which is loaded and executed by a processor to implement a method of area planning as claimed in any of claims 1 to 10.

14. A computer program product comprising a computer program or instructions which, when executed by a processor, implements the area planning method according to any one of claims 1 to 10.

Images