Store site selection method, device, and storage medium

Abstract

The application discloses a store site selection method and device and a storage medium, relates to the technical field of data processing, and comprises the following steps: acquiring at least one site selection data based on an execution module; delivering the site selection data to an intelligent agent as input data, and acquiring decision action data output by the intelligent agent based on the input data; determining a target store site selected by the intelligent agent based on the decision action data, and feeding back the target store site to a user end. Thus, the store site selection can be performed through the trained intelligent agent, so that the accuracy of store site selection decision can be improved.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to methods, equipment and storage media for store location selection. Background Technology

[0002] Currently, in the process of store site selection, business personnel typically analyze available store locations based on simple statistical analysis. This analysis process largely relies on personal experience and intuition to assess the merits of potential locations, lacking the ability to learn and iterate independently, and failing to adapt to the complex and ever-changing market environment. This results in low accuracy in store site selection decisions.

[0003] The above content is only used to help understand the technical solution of this application and does not represent an admission that the above content is prior art. Summary of the Invention

[0004] This application provides a method, device, and storage medium for store location selection, aiming to solve the problem of low accuracy in store location selection decisions.

[0005] To achieve the above objectives, this application provides a store location selection method. The store location selection system includes an execution module, an intelligent agent, and a training module. The store location selection method includes the following steps:

[0006] At least one addressing data is obtained based on the execution module;

[0007] The location data is transmitted to the agent as input data, and the decision action data output by the agent based on the input data is obtained.

[0008] Based on the decision action data, the target store location selected by the intelligent agent is determined, and the target store location is fed back to the user.

[0009] In one embodiment, before the step of transmitting the location data as input data to the agent and obtaining the decision action data output by the agent based on the input data, the method includes:

[0010] Scene state data is obtained based on the training module;

[0011] The scene state data is transmitted to the intelligent agent as input data, and the decision action data output by the intelligent agent based on the input data is obtained;

[0012] The Q value is calculated based on the scene state data and the decision action data, and according to a preset formula.

[0013] The reward value is determined based on the scenario state data and the reward function, and the target Q value is determined based on the reward value and the Q value.

[0014] When the difference between the target Q value and the target Q value is less than a preset threshold for a continuous number of iterations, the agent is output.

[0015] In one embodiment, before outputting the agent when the difference between the target Q value and the target Q value is less than a preset threshold for a continuous number of iterations, the following steps are included:

[0016] The agent's hyperparameters are updated based on the Q-value and the target Q-value;

[0017] The updated agent performs the step of acquiring scene state data.

[0018] In one embodiment, before the step of obtaining the decision action data output by the agent based on the input data, the method includes:

[0019] The agent determines the target strategy based on the received input data;

[0020] Based on the input data, the target strategy selects and outputs corresponding decision action data, wherein the decision action data includes classifying the store location as a prime location, a high-quality location, an ordinary location, or a worthless location.

[0021] In one embodiment, the step of determining the reward value based on the scene state data and the reward function includes:

[0022] Obtain various business indicator data and corresponding weight values ​​from the scenario status data, wherein the business indicator data includes cost indicators, revenue indicators and risk indicators;

[0023] The instant reward for each business indicator is calculated based on the instant reward function, business indicator data, and corresponding weight values.

[0024] The reward value is calculated based on the reward function and the real-time rewards of various business indicator data.

[0025] In one embodiment, before the step of obtaining scene state data based on the training module, the method further includes:

[0026] Collect and integrate information on store locations, competitors, residential communities, administrative districts, and main roads, and determine them as the original dataset;

[0027] Geographic information coding technology is used to convert the address information in the original dataset into latitude and longitude coordinates, and the store locations, competitor locations, and community locations are marked on the map;

[0028] Based on the preset business district range, calculate the various business indicator data corresponding to the original dataset within the radius, and determine them as scene status data.

[0029] In one embodiment, after the step of calculating various business indicator data within a preset business district range and determining them as the scene state data, the method further includes:

[0030] The scene state data is sorted according to its numerical value, and the first quartile and the third quartile are determined. The first quartile is the value located at the 25th percentile of the sort, and the third quartile is the value located at the 75th percentile of the sort.

[0031] The fourth quartile is determined based on the first quartile and the third quartile, and the abnormal range is determined based on the fourth quartile.

[0032] Clear the data within the specified anomaly range and output the processed scene status data.

[0033] In one embodiment, after determining the target store location selected by the intelligent agent based on the decision action data and feeding back the target store location to the user terminal, the method further includes:

[0034] Predict the potential customer traffic for the target store location;

[0035] Based on the forecast results and the corresponding cost-benefit analysis, business recommendations for the target store locations are generated.

[0036] In addition, to achieve the above objectives, this application also provides a store location selection device, which includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the store location selection method as described above.

[0037] In addition, to achieve the above objectives, this application also provides a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the store location selection method as described above.

[0038] This application provides a store location selection method, store location selection device, and storage medium. Based on the execution module, at least one location selection data is obtained, and then the location selection data is transmitted to the intelligent agent as input data. The decision action data output by the intelligent agent based on the input data is obtained, and finally the target store location selected by the intelligent agent is determined based on the decision action data. The target store location is then fed back to the user terminal, thereby improving the accuracy of store location selection decisions by using a trained intelligent agent. Attached Figure Description

[0039] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0040] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0041] Figure 1 This is a flowchart illustrating the first embodiment of the store location selection method of this application;

[0042] Figure 2 This is a flowchart illustrating the second embodiment of the store location selection method of this application;

[0043] Figure 3 This is a quantitative modeling diagram of the scene state data involved in the solution of this embodiment;

[0044] Figure 4 This is a diagram showing the specific parameters of the scene state data involved in the embodiments of this application;

[0045] Figure 5 This is a data table of decision-making actions involved in the embodiments of this application;

[0046] Figure 6 This is a Q-value matrix diagram related to the embodiments of this application;

[0047] Figure 7 This is a hyperparameter assignment table for the embodiments of this application;

[0048] Figure 8 This is a hyperparameter adjustment table related to the embodiments of this application;

[0049] Figure 9 This is a diagram showing the relationship between the reward function and business indicator data involved in the solution of this embodiment;

[0050] Figure 10 This is a flowchart illustrating the third embodiment of the store location selection method of this application;

[0051] Figure 11 This is a point calibration diagram involving the scheme of the embodiments of this application;

[0052] Figure 12 This is a statistical table of data related to the embodiments of this application;

[0053] Figure 13This is a schematic diagram of the data preprocessing process involved in the embodiments of this application;

[0054] Figure 14 This is a schematic diagram of the hardware operating environment of the store location selection equipment involved in the embodiments of this application.

[0055] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0056] It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit this application.

[0057] To better understand the above technical solutions, exemplary embodiments of this application will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of this application are shown in the drawings, it should be understood that this application can be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of this application and to fully convey the scope of this application to those skilled in the art.

[0058] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0059] The main solution of this application is: to obtain at least one site selection data based on the execution module; to transmit the site selection data as input data to the intelligent agent, and to obtain the decision action data output by the intelligent agent based on the input data; to determine the target store location selected by the intelligent agent based on the decision action data, and to feed back the target store location to the user terminal.

[0060] Currently, in the process of store site selection, business personnel typically analyze available store locations based on simple statistical analysis. This analysis process largely relies on personal experience and intuition to assess the merits of potential locations, lacking the ability to learn and iterate independently, and failing to adapt to the complex and ever-changing market environment. This results in low accuracy in store site selection decisions.

[0061] This application obtains at least one site selection data based on the execution module, then transmits the site selection data as input data to the intelligent agent, obtains the decision action data output by the intelligent agent based on the input data, and finally determines the target store location selected by the intelligent agent based on the decision action data, and feeds back the target store location to the user terminal. Thus, the trained intelligent agent can be used to select store locations, thereby improving the accuracy of store location decisions.

[0062] It should be noted that the executing entity in this embodiment can be a store location selection system, or a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, or mobile phone, or a store location selection device capable of performing the above functions. This embodiment does not specifically limit it in this way. The following uses a store location selection system as the executing entity as an example to describe this embodiment and the following embodiments.

[0063] Example 1

[0064] Please refer to Figure 1 In the first embodiment, the store location selection method includes the following steps:

[0065] Step S10: Obtain at least one addressing data based on the execution module.

[0066] In this embodiment, the processing actions are performed by a store location selection system, which includes an execution module, an agent, and a training module. After the store system detects that the agent has completed training, the execution module acquires and deploys the agent in the store location selection system. The front-end of the store location selection system displays a location selection interface. When a user uploads a location selection request based on the interface, the front-end sends the request back to the store location selection system. The store location selection system determines the corresponding location selection data based on the request. This location selection data is at least one piece of data, and it is preprocessed data. The specific steps are the same as the data preprocessing steps described in Embodiment 3, and will not be repeated here. Furthermore, the location selection data includes location selection data for multiple store locations. The execution module inputs the location selection data for at least one store location into the agent, and the agent outputs decision action data for at least one store location.

[0067] Step S20: Transmit the location data as input data to the agent, and obtain the decision action data output by the agent based on the input data.

[0068] In this embodiment, the preprocessed site selection data is used as input data and transmitted to the agent. The agent selects a corresponding target strategy based on the input data and outputs corresponding decision action data. The decision action data output by the agent is obtained, wherein the decision action data is the agent's rating of the target points in the site selection data.

[0069] Step S30: Determine the target store location selected by the intelligent agent based on the decision action data, and feed back the target store location to the user terminal.

[0070] In this embodiment, the site selection data corresponds to data from multiple target locations. Based on the decision action data output by the intelligent agent to each target location, the location with the highest rating among the target locations is determined and identified as the target store location. The target store location is then fed back to the user, thereby completing the store site selection.

[0071] Optionally, in this embodiment, after the step of determining the target store location selected by the agent based on the target strategy and feeding back the target store location to the user, the method further includes:

[0072] Predict the potential customer traffic for the target store location; based on the prediction results and the corresponding cost-benefit analysis, generate business recommendations for the target store location.

[0073] Specifically, predicting potential customer traffic for a target store location involves multiple factors, including but not limited to geographical location, surrounding demographics, traffic conditions, competition, and seasonal variations. A data-driven approach, combined with data analytics, is employed. This involves analyzing the specific location of the store and its surrounding environment, including demographic data, traffic data, peak pedestrian hours, the location, type, price range, and historical sales data of nearby competitors, to determine potential customer traffic. Based on the analysis results or model outputs, business recommendations are then made. For example, if the predicted customer traffic indicates that the location is unsuitable, alternative locations need to be considered. Furthermore, a marketing plan is developed based on the characteristics of the target customer group, such as attracting younger customers through social media promotion, thereby accurately predicting potential customer traffic and conducting cost-benefit analysis.

[0074] In addition, potential customer traffic can be predicted by building a model. The specific steps are as follows: first, select a suitable prediction model, such as linear regression, time series analysis, or machine learning model (such as random forest, neural network, etc.); then, create useful feature variables based on the collected data; and finally, use the selected model to predict customer traffic in different time periods.

[0075] In the technical solution provided in this embodiment, at least one site selection data is obtained based on the execution module, and then the site selection data is transmitted to the intelligent agent as input data. The decision action data output by the intelligent agent based on the input data is obtained, and finally the target store location selected by the intelligent agent is determined based on the decision action data. The target store location is then fed back to the user terminal, thereby improving the accuracy of store site selection decisions by using a trained intelligent agent.

[0076] Example 2

[0077] Please refer to Figure 2In the second embodiment, after step S20, the method further includes:

[0078] Step S40: Obtain scene state data based on the training module.

[0079] In this embodiment, the processing action is performed by the training module in the store location system. The scene status data is preprocessed data, which includes store rent, transfer fee, management fee, average wage in the city, number of residential communities within a 250-meter radius, occupancy rate of the residential community, rent per square meter of the residential community, second-hand housing price per square meter of the residential community, number of competing stores within a 250-meter radius, and number of roads with more than four lanes in both directions within a 250-meter radius.

[0080] Optionally, the scenario status data in this embodiment is divided into cost indicators, revenue indicators, and risk indicators, each with a fixed unit, such as... Figure 3 As shown, Figure 3 This embodiment provides a quantitative modeling diagram of the scenario state data indicators. The cost indicators include shop rent, transfer fees, management fees, and the average wage in the city. Correspondingly, lower overall monthly rent results in lower costs, lower transfer fees result in lower sunk costs, lower management fees per unit area result in lower costs, and lower city wages result in lower labor costs. The revenue indicators include the number of residential communities within a 250-meter radius, community occupancy rates, rent per square meter in the community, and resale price per square meter in the community. Correspondingly, more communities mean more potential sales opportunities, higher occupancy rates mean more potential sales opportunities, higher rents mean higher average transaction prices, and higher property prices mean higher average transaction prices. The risk indicators include the number of floors in the shop, the number of competing shops within a 250-meter radius, and the number of roads with four or more lanes within a 250-meter radius. Correspondingly, more floors mean lower sales per unit area, more competitors mean fewer sales opportunities, and more lanes and proximity to main roads mean fewer potential sales opportunities and less time spent in the shop.

[0081] Specifically, major cost items include labor costs, shop rental expenses, and fixed-percentage management fees, which are the foundation for establishing a store. In today's commercial and real estate environment, controlling store costs has a more significant and far-reaching impact on store cash flow and profits than marketing growth. Major revenue items can be categorized into the number of potential consumers in the surrounding area, the level of potential consumer spending power in the surrounding area, and the scope of business coverage. These factors qualitatively, rather than quantitatively, affect store revenue. Dependent on the physical retail environment, foot traffic determines the exposure of sales channels; the more foot traffic a store has, the higher its sales volume will be. The more prosperous the business district where the store is located, the higher the premium that can be added to the goods, and the higher the transaction amount per transaction. In addition to direct cost and revenue factors, there are also risk items that indirectly affect store profitability. These risks are both known and unknown; that is, we know the risk exists, but we don't know the probability of its occurrence or the magnitude of its impact. If there are many main roads / expressways around the store, even with high foot traffic and many residential areas, passersby may not be able to enter the store and stop to make a purchase. From a consumer psychology perspective, the flow of customers in fast retail / fresh food retail stores should be as smooth and continuous as possible. If customers need to go up and down stairs frequently and there are many floors, it will greatly affect the customer shopping experience and cause a loss in sales. The number of similar competing stores already existing near the location will also greatly affect the store's sales, and may even create a monopoly of customer traffic, which is a direct threat to the store's sales.

[0082] Furthermore, due to the different units (dimensions) and magnitudes of the evaluation indicators, a unified evaluation mechanism is introduced to perform dimensionless processing. Here, a scoring model with a maximum score of 10 is used. After unified quantitative modeling, all state s values ​​are converted into dimensionless data of the same magnitude. Specific scenario state data s parameters are as follows: Figure 4 The scenario status data is determined based on the identified cost, revenue, and risk indicators, such as... Figure 4 As shown, Figure 4 This is a specific parameter diagram of the scene state data involved in the embodiments of this application. Let C1~C4 represent cost indicators, P1~P4 represent revenue indicators, R1~R3 represent risk indicators, and s represent scene state data. Then the mathematical set of scene state data s is:

[0083] s={(C1,C2,C3,...,R3)|C1,C2,C3,...,R2∈{0,1,2,...,10},R3∈{0,3,5}}

[0084] Step S50: Transmit the scene state data as input data to the intelligent agent, and obtain the decision action data output by the intelligent agent based on the input data.

[0085] In this embodiment, the intelligent agent refers to an entity capable of perceiving the environment and influencing it through actions. It can be a software program, a module deployed in a store location selection system, or other modules capable of making autonomous decisions. After the store location selection system acquires the scene state data, it transmits the scene state data as input data to the intelligent agent. When the intelligent agent's output action is detected, the system acquires the decision action data output by the intelligent agent based on the input data.

[0086] Understandably, an intelligent agent maximizes accumulated rewards through interaction with scene state data. An intelligent agent comprises a perception module, a decision-making module, and an action module. The agent acquires information about the current environment through the perception module. This information is typically provided in the form of scene state data, which can be a partial or complete description of the environment. Then, the decision-making module takes corresponding actions based on the current state information, i.e., decision-making action data. This decision-making process is based on a corresponding policy or can be based on a complex machine learning model. Finally, the action module executes specific behaviors according to the action instructions output by the decision-making module. The workflow of an intelligent agent is as follows: The agent acquires information about the current state from the environment. Based on the current state, the agent decides which action to take. This decision may be based on a predefined policy or a policy learned through learning. The agent executes the selected action. The environment provides feedback based on the agent's actions, usually a formalized reward signal. The agent updates its internal state or policy based on the feedback to make better decisions in the future.

[0087] For example, the decision-making data output by the intelligent agent includes prime locations, high-quality locations, ordinary locations, and worthless locations, corresponding to classifying stores into three levels: A, B, and C. Location A is considered a prime location where a franchise agreement can be signed within one month; location B is defined as a high-quality location where a franchise agreement can be signed within one to three months; and location C is defined as an ordinary location where a franchise agreement can be signed after three months. Additionally, there are worthless locations X that are abandoned by actual business operations, lacking the potential for signing agreements and generating profits. Figure 5 As shown, Figure 5 This is a decision action data table related to the embodiments of this application. The decision action a of the intelligent agent when facing different scenario states s has only four types: A, B, C, and X, determined by... Figure 5 The mathematical set of decision action 'a' can be derived as: a = {A, B, C, X}.

[0088] Optionally, in this embodiment, before the step of obtaining the decision action data output by the intelligent agent based on the input data, the method includes:

[0089] The intelligent agent determines a target strategy based on the received input data; through the target strategy, it selects and outputs corresponding decision action data based on the input data, wherein the decision action data includes classifying the store location as a prime location, a high-quality location, an ordinary location, or a worthless location.

[0090] Specifically, after the execution module or the training module acquires the scene state data, the execution module or the training module transmits the scene state data as input data to the agent. Upon receiving the input data, the decision-making module within the agent automatically selects a target strategy based on the scene state data and executes specific decision actions according to the action instructions output by the target strategy, thereby outputting decision action data. This decision action data includes classifying store locations as prime locations, high-quality locations, ordinary locations, and worthless locations.

[0091] Step S60: Calculate the Q value based on the scene state data and the decision action data, and based on a preset formula.

[0092] In this embodiment, after the agent outputs decision action data, the Q value is calculated based on Q(st,at), where s t a represents the scene state data at time step t. t Let t represent the decision action of the agent in response to the scene state data s at time step t, and calculate the Q value based on the above data and Q(st,at).

[0093] Specifically, based on the set expressions for states s and actions a, the Q-value table can be constructed using s × a, i.e., the Cartesian product of the two sets, as shown below. Figure 6 The Q-value matrix shown is Figure 6 This is a Q-value matrix diagram related to the embodiments of this application. This matrix is ​​the core Q-value table. Based on the updates of the values ​​in this matrix, the action decisions of the intelligent agent in different states are recorded. By continuously updating this table through iterative formulas, the learning process of the intelligent agent can be realized.

[0094] Step S70: Determine the reward value based on the scene state data and the reward function, and determine the target Q value based on the reward value and the Q value.

[0095] In this embodiment, the scene state data corresponds to a reward function. In the steps described above, the scene state data is divided into cost indicators, revenue indicators, and risk indicators. Each type of indicator data is identified by a status code, and each code is associated with a reward function. The reward value is obtained by substituting the numerical value of the scene state data into the reward function. Then, the target Q value is calculated based on the reward value and the Q value.

[0096] Specifically, in this embodiment, the target Q value is calculated based on the reward value and the Q value through iterative calculation. Further, in this embodiment, before outputting the agent when the difference between the target Q value and the Q value for a consecutive number of iterations is less than a preset threshold, the following steps are included:

[0097] Update the hyperparameters of the agent based on the Q value and the target Q value; and perform the step of acquiring scene state data based on the updated agent.

[0098] Specifically, each type of scenario state data corresponds to a time step. After this time step, the agent's hyperparameters need to be updated based on the Q-value and the target Q-value. These hyperparameters include the learning rate, immediate reward, discount factor, exploration rate, exploration rate decay, minimum exploration rate, and exploration strategy. After updating the agent, the store location selection system randomly selects another target scenario state data, allowing the agent to execute the next time step. The agent calculates the Q-value for this time step based on the target state data and the corresponding strategy, then calculates the difference. This is achieved by iteratively calculating the Q-value using the Q-learning algorithm and the Bellman equation, with the following formula:

[0099] Q(s t ,a t )←Q(s t ,a t )+α[r t+1 +γQ a (s t+1 ,a)-Q(s t ,a t )]

[0100] Q(s t ,a t ) is the state s at time step t. t Next, take action a t The value of is (0,1). Here, "s" represents the state of the external environment, i.e., the scene state data; "a" represents the action the agent can take in a given state, i.e., the decision action data; t represents the time step. In Q-learning, time slices are discrete concepts of time steps, i.e., the time nodes when an action must be taken in the face of the environment. Numerous discrete time slices are fitted into a continuous decision process. α is the learning rate, determining the degree to which the agent updates the old Q value with a new one, i.e., the proportion of new knowledge replacing old knowledge, and its value is (0,1]. r is the immediate reward obtained after performing the action, the main reward and punishment factor after performing the action, and is used as the main feedback to adjust the agent's action selection. γ is the discount factor, determining the present value of future rewards, and its value is (0,1).t and s t+1 These represent the states at time steps t and t+1, respectively. When iteratively calculating the Q-value based on the Q-learning algorithm and the Bellman equation, the state function is first initialized:

[0101] s={(C1,C2,C3,...,R3)|C1,C2,C3,...,R2∈{0,1,2,...,10},R3∈{0,3,5}}

[0102] All states are initially at position 0, and the Q-value table is also initialized to 0. The time step t is divided into two steps: t is the time node for making a judgment based on historical experience, and t+1 is the time node for comparing the agent's decision based on the actual location rating. Combined with r... learning =r process +r final The reward function substitutes process reward and outcome reward into the formula. After using historical Q-values ​​for decision-making, the process reward and outcome reward are coefficient-reduced, and the iterative portion is then updated into the Q-index, thus iterating cyclically until the optimal solution is approximated. Furthermore, when determining the agent's exploration strategy, there is another hyperparameter—the exploration rate ε—not reflected in the formula, but equally important for agent learning. When using a greedy strategy, ε controls the frequency of the agent's exploration of unknown actions, similar to the aggressiveness of trying new actions, and takes values ​​(0,1]. The learning rate α, discount factor γ, immediate reward r, and exploration rate ε have default values ​​at the start of iteration, such as... Figure 7 As shown, Figure 7 This is the hyperparameter assignment table involved in the embodiments of this application. Subsequently, after the agent converges, the production parameters are optimized to better preserve existing model features and avoid decision divergence. Specific adjustments are as follows: Figure 8 As shown, Figure 8 This is a hyperparameter adjustment table related to the embodiments of this application.

[0103] As an alternative implementation method for acquiring scene state data, after updating the agent's hyperparameters, target scene state data different from the previous training step can be acquired and input into the updated agent to continue the agent's training iteration process. By using multiple different types of scene state data, the accuracy of agent training can be improved.

[0104] As another alternative implementation for acquiring scene state data, after updating the agent's hyperparameters, the scene state data from the previous training step can continue to be used until the decision action data output by the agent is calculated. When the difference between the target Q-value and the target Q-value is less than a preset threshold, the target scene decision action data is then replaced and input into the agent for training. By using the same scene state data to train the agent multiple times, the amount of training data prepared can be reduced, thereby improving the agent's training efficiency.

[0105] Optionally, in this embodiment, the step of determining the reward value based on the scene state data and the reward function includes:

[0106] Obtain various business indicator data and corresponding weight values ​​from the scenario status data, wherein the business indicator data includes cost indicators, revenue indicators, and risk indicators; calculate the instant reward for each business indicator data based on the instant reward function, the business indicator data, and the corresponding weight values; and calculate the reward value based on the reward function and the instant rewards for each business indicator data.

[0107] Specifically, the business metrics data include cost metrics, revenue metrics, and risk metrics. Each business metric data has a corresponding weight value. The reward function includes providing immediate rewards for the actions taken by the agent at each time step. The reward value is calculated from each of the immediate rewards, such as... Figure 9 As shown, Figure 9 This is a diagram showing the relationship between the reward function and business indicator data involved in this embodiment. The reward function can be described as follows:

[0108]

[0109] The immediate reward for each business indicator is calculated by multiplying each immediate reward function by its weight value. Finally, the evaluation of the agent's decision-making results is primarily based on the action 'a' given by the agent after evaluating the current single-dimensional state 's'. If the action 'a' given by the agent differs from the actual results in the training set, the decision is considered incorrect. Since this method's model training is result-oriented, the agent should be given an overwhelming penalty, sufficient to offset the rewards during the process guidance while still having a margin exceeding an order of magnitude, and vice versa. This reward function is only applicable during the agent's learning phase. After the model converges and is deployed, there are no decision-making errors. The specific judgment is as follows:

[0110]

[0111] In summary, the reward function of the Q-learning algorithm during agent training is r. learning =r process +r final .

[0112] Step S80: When the difference between the target Q value and the target Q value is less than a preset threshold for a continuous number of iterations, the agent is output.

[0113] In this embodiment, the target number of iterations can be preset, specifically set to 500, and the preset threshold is set to 0.01. For the dataset, the convergence is determined by the iterative stability of the observed Q-value differences, achieving a balance between data volume and efficiency. When the variation difference of Q-values ​​in 500 consecutive iterations is less than 0.01, the agent can be considered to have converged, and then the agent is output and deployed in the store location selection system.

[0114] In the technical solution provided in this embodiment, scene state data is acquired based on the training module, and the scene state data is input into the intelligent agent. The decision action data output by the intelligent agent is determined. Then, based on the scene state data and the decision action data, the Q value is calculated based on a preset formula. Next, the reward value is determined based on the scene state data and the reward function. Finally, the target Q value is determined based on the reward value and the Q value. When the difference between the target Q value and the Q value is continuously iterated for a target number of times and is less than a preset threshold, the intelligent agent is output. Thus, the trained intelligent agent can be used for store site selection to improve the accuracy of store site selection decisions.

[0115] Example 3

[0116] Please refer to Figure 10 In the second embodiment, before the step of obtaining scene state data based on the training module, the method further includes:

[0117] Step S60: Collect and integrate store locations, competitor information, community information, administrative district information, and main road information, and determine them as the original dataset.

[0118] Step S70: Use geographic information coding technology to convert the address information in the original dataset into latitude and longitude coordinates, and mark the store locations, competitor locations, and community locations on the map.

[0119] Step S80: Based on the preset business district range, calculate the various business indicator data corresponding to the original dataset within the radius, and determine them as the scene status data.

[0120] In this embodiment, various raw datasets are first acquired before the data source, including store locations, competitor information, community information, administrative district information, and main road information. The data types of the raw datasets are integers, floating-point numbers, and strings. The collected raw data may not be usable directly and needs to be transformed into quantifiable and definable data. Specifically, this involves data transformation and the use of front-end visual geographic information tools for information collection, combined with automated scripts for batch processing of the indicator data. Then, using the geocoding function provided by the geographic information service provider, the address strings are converted into latitude and longitude floating-point numbers, and the selected locations, competitor locations, and community locations are rendered on the map.

[0121] Specifically, such as Figure 11 As shown, Figure 11 This is a point location mapping diagram related to the embodiments of this application. In the diagram, crosses represent competitors, stars represent target points, solid circles represent the center point of the residential area, gray dashed lines represent lane boundaries, black solid lines represent the residential area and road boundaries, and arrows indicate the direction of traffic flow on main roads. A semi-transparent blue layer representing the business coverage area is drawn with a radius of 250 meters centered on the target point, thereby transforming some counting indicators into a planar geometric problem.

[0122] Based on the above-mentioned calibrated map, determine the number of residential areas within a 250-meter radius. This is done by intersecting the polygonal line segments of the residential areas within the blue circular area; an intersection increments the count by 1, with no duplicate entries for a single residential area. The value is 5. For the number of competitor shops within a 250-meter radius, determine the geometric containment relationship between the blue circular area and the competitor's location; a containment increments the count by 1, resulting in a value of 2. For the number of roads with four or more lanes within a 250-meter radius, determine the number of roads with more than four lanes, and increment the count by 1 if any road's edge line segment intersects the circle. Based on the map, this is determined to be 1. After data completion and indicator transformation, the data is linked through foreign keys in each table. Unnecessary auxiliary linked data is discarded, resulting in a unified data format with site selection locations as the dimension. Based on the scoring scheme, scores are entered into each field to obtain the final result. Figure 12 , Figure 12 This is a statistical table of data related to the embodiments of this application.

[0123] For example, such as Figure 13 As shown, Figure 13This is a schematic diagram of the data preprocessing process involved in the embodiments of this application. The user selects or inputs a target location in the front-end interface. The front-end obtains the raw dataset of the target location through an interface and determines its geographical information. Geographic information encoding technology is used to convert the address information in the raw dataset into latitude and longitude coordinates, and the store location, competitor location, and community location are marked on a map and displayed to the front-end. After calculation, the planar geometric relationships and accumulated index counts are determined, the index data of the selected location is updated, and it is identified as scene state data and stored in the database.

[0124] Optionally, in this embodiment, after the step of calculating various business indicator data within a preset business district range and determining them as the scene state data, the method further includes:

[0125] The scene state data is sorted according to its numerical value, and the first quartile and the third quartile are determined. The first quartile is the value located at the 25th percentile of the sort, and the third quartile is the value located at the 75th percentile of the sort. The fourth quartile distance is determined based on the first quartile and the third quartile, and the anomaly range is determined based on the fourth quartile distance. Data within the anomaly range is cleared according to the anomaly range, and the processed scene state data is output.

[0126] Specifically, after data processing, outliers need to be removed to refine the data. The specific cleaning method is the quartile method. First, the interquartile range (IQR) is calculated. The first quartile (Q1, 25th quartile) and the third quartile (Q3, 75th quartile) are calculated, thus determining the IQR: IQR = Q3 - Q1. Then, the range of outliers is determined. Specifically, data points below Q1 - 1.5 × IQR or above Q3 + 1.5 × IQR are considered outliers. Outlier data points in the dataset are then deleted based on this range. Furthermore, after removing outliers, the remaining dataset is re-analyzed using the quartile method. Once no values ​​exceed the normal range, the data cleaning is considered complete.

[0127] In the technical solution provided in this embodiment, store locations, competitor information, community information, administrative district information, and main road information are collected and integrated to form the original dataset. Then, geographic information coding technology is used to convert the address information in the original dataset into latitude and longitude coordinates, and store locations, competitor locations, and community locations are marked on a map. Finally, based on a preset business district range, the various business indicator data corresponding to the original dataset within the radius are calculated and determined as the scene state data. This embodiment preprocesses the data to ensure the reliability of the training data for the intelligent agent, thereby improving the training efficiency of the intelligent agent.

[0128] Since the system described in the embodiments of this application is a system used to implement the methods of the embodiments of this application, those skilled in the art can understand the specific structure and variations of the system based on the methods described in the embodiments of this application, and therefore will not be described in detail here. All systems used in the methods of the embodiments of this application fall within the scope of protection of this application.

[0129] This application provides a store location selection device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, which are executed by the at least one processor to enable the at least one processor to perform the store location selection method in the above embodiment 1.

[0130] The following is for reference. Figure 14 The diagram illustrates a structural schematic suitable for implementing the store location selection device in the embodiments of this application. The store location selection device in the embodiments of this application may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 14 The store location selection device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0131] like Figure 14As shown, the store location selection device may include a processing unit 1001 (e.g., a core processor, graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage device 1003 into a random access memory (RAM) 1004. The RAM 1004 also stores various programs and data required for the operation of the store location selection device. The processing unit 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to the I / O interface 1006: input devices 1007 including, for example, a touchscreen, touchpad, keyboard, mouse, image sensor, microphone, accelerometer, gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), speaker, vibrator, etc.; storage devices 1003 including, for example, magnetic tape, hard disk, etc.; and communication devices 1009. Communication device 1009 allows the store location selection device to communicate wirelessly or wiredly with other devices to exchange data. Although the figures show store location selection devices with various systems, it should be understood that implementation or possession of all the systems shown is not required. More or fewer systems may be implemented alternatively.

[0132] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from ROM 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0133] The store location selection device provided in this application, employing the store location selection method described in the above embodiments, can solve the technical problem of low accuracy in store location selection decisions of related solutions. Compared with the prior art, the beneficial effects of the store location selection device provided in this application are the same as those of the store location selection method provided in the above embodiments, and other technical features of this store location selection device are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0134] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0135] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0136] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, the computer-readable program instructions being used to execute the store location selection method in the above embodiments.

[0137] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0138] The aforementioned computer-readable storage medium may be included in the store location selection device; or it may exist independently and not be assembled into the store location selection device.

[0139] The aforementioned computer-readable storage medium carries one or more programs. When the aforementioned one or more programs are executed by the store location selection device, the store location selection device causes the following: to acquire at least one location selection data based on the execution module; to transmit the location selection data as input data to the intelligent agent, and to acquire the decision action data output by the intelligent agent based on the input data; to determine the target store location selected by the intelligent agent based on the decision action data, and to feed back the target store location to the user terminal.

[0140] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a Local Area Network (LAN) or a Wide Area Network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0141] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0142] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0143] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described store location selection method, which can solve the technical problem of low accuracy in store location decision-making in related schemes. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as the beneficial effects of the store location selection method provided in the above embodiments, and will not be repeated here.

[0144] This application provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the store location selection method described above.

[0145] The computer program product provided in this application can solve the technical problem of low accuracy in store location decision-making in related solutions. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as the beneficial effects of the store location selection method provided in the above embodiments, and will not be repeated here.

[0146] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent scope of this application.

Claims

1. A method for selecting store locations, characterized in that, The method, applied to a store location selection system, which includes an execution module, an intelligent agent, and a training module, comprises the following steps: Collect and integrate information on store locations, competitors, residential communities, administrative districts, and main roads, and determine them as the original dataset; Geographic information coding technology is used to convert the address information in the original dataset into latitude and longitude coordinates, and the store locations, competitor locations, and community locations are marked on the map; Based on the preset business district range, calculate the various business indicator data corresponding to the original dataset within the radius, and determine them as scene status data; The scene state data is sorted according to its numerical value, and the first quartile and the third quartile are determined. The first quartile is the value located at the 25th percentile of the sort, and the third quartile is the value located at the 75th percentile of the sort. The fourth quartile is determined based on the first quartile and the third quartile, and the abnormal range is determined based on the fourth quartile. Clear the data within the specified abnormal range and output the processed scene status data. The scene state data is obtained based on the training module; The scene state data is transmitted to the intelligent agent as input data, and the decision action data output by the intelligent agent based on the input data is obtained; Based on the scene state data and the decision action data, and calculated according to a preset formula, the Q value is calculated based on Q(st, at) after the agent outputs the decision action data, where st represents the scene state data at time step t, and at represents the decision action of the agent for the scene state data s at time step t. According to the set expression of state s and action a, the Q value table can be s×a, that is, the Cartesian product of the two sets. Acquire various business indicator data and corresponding weight values ​​from the scene status data. The business indicator data includes cost indicators, revenue indicators, and risk indicators, denoted by C1 to C4 for cost indicators, P1 to P4 for revenue indicators, and R1 to R3 for risk indicators. The scene status data is represented by s. The cost indicators include store rent, transfer fees, management fees, and the average wage in the city. The revenue indicators include the number of residential communities within a 250-meter radius, the occupancy rate of the communities, the rent per square meter of the communities, and the second-hand housing price per square meter of the communities. The risk indicators include the number of floors of the shops, the number of competitor shops within a 250-meter radius, and the number of roads with four or more lanes in both directions within a 250-meter radius. The instant reward for each business indicator is calculated based on the instant reward function, business indicator data, and corresponding weight values. The reward value is calculated based on the reward function and the real-time rewards of various business indicators, and the target Q value is determined based on the reward value and the Q value. When the difference between the target Q value and the target Q value is less than a preset threshold for a continuous number of iterations, the agent is output. At least one addressing data is obtained based on the execution module; The location selection data is transmitted to the agent as input data, and the decision action data output by the agent based on the input data is obtained. Based on the decision action data, the target store location selected by the intelligent agent is determined, and the target store location is fed back to the user terminal; The reward function is: 。 2. The method as described in claim 1, characterized in that, Before outputting the agent when the difference between the target Q value and the target Q value is less than a preset threshold for a continuous number of iterations, the following steps are included: The agent's hyperparameters are updated based on the Q-value and the target Q-value; The updated agent performs the step of acquiring scene state data.

3. The method as described in claim 1, characterized in that, Before the step of obtaining the decision action data output by the intelligent agent based on the input data, the following steps are included: The agent determines the target strategy based on the received input data; Based on the input data, the target strategy selects and outputs corresponding decision action data, wherein the decision action data includes classifying the store location as a prime location, a high-quality location, an ordinary location, or a worthless location.

4. The method as described in claim 1, characterized in that, After the step of determining the target store location selected by the intelligent agent based on the decision action data and feeding back the target store location to the user terminal, the method further includes: Predict the potential customer traffic for the target store location; Based on the forecast results and the corresponding cost-benefit analysis, business recommendations for the target store locations are generated.

5. A store location selection device, characterized in that, The store location selection device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the store location selection method as described in any one of claims 1 to 4.

6. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the store location selection method as described in any one of claims 1 to 4.

Images