CN-121998555-A - Retail cabinet replenishment analysis method and system based on multidimensional decision matrix

CN121998555ACN 121998555 ACN121998555 ACN 121998555ACN-121998555-A

Abstract

The invention discloses a retail cabinet replenishment analysis method and a system based on a multidimensional decision matrix, wherein the method collects the site characteristics of a retail cabinet, the user characteristics of a service object and local city hot-sale commodity data to construct the multidimensional decision matrix; the method comprises the steps of completing commodity compliance screening and recommendation based on a multi-dimensional decision matrix, determining a candidate commodity set of a retail cabinet adapting to a current scene and user requirements, adjusting the shelf display position of candidate commodities through dynamic allocation rules by combining the sales performance and attribute characteristics of the candidate commodities, dividing commodity categories according to historical sales frequencies of the commodities, formulating different commodity replenishment periods for different categories of commodities by adopting a dynamic classification algorithm, predicting commodity replenishment requirements by utilizing a sales volume prediction model, and planning an optimal distribution path by combining a path optimization algorithm. The invention solves the problems of low space utilization rate caused by static display of the traditional retail cabinet, backlog or backlog caused by stiff replenishment period, logistics distribution resource waste and lag in compliance management response.

Inventors

FAN JUNLONG
HUANG AIHUA
YIN JUEHUI

Assignees

上海趣致网络科技有限公司

Dates

Publication Date: 20260508
Application Date: 20260128

Claims (10)

1. The retail cabinet replenishment analysis method based on the multidimensional decision matrix is characterized by comprising the following steps of: collecting the place characteristics of a retail cabinet, the user characteristics of a service object and local city hot-sale commodity data, and constructing a multi-dimensional decision matrix integrating the place characteristics, the user characteristics of the service object and the local city hot-sale commodity data; Completing commodity compliance screening and recommendation based on the multi-dimensional decision matrix, and determining a candidate commodity set of the retail cabinet adapting to the current scene and the user requirement; combining the sales performance and attribute characteristics of the candidate commodity, and adjusting the shelf display position of the candidate commodity through a dynamic allocation rule; Dividing commodity categories by adopting a dynamic classification algorithm according to the historical commodity sales frequency, and formulating different commodity replenishment periods for commodities of different categories; and predicting the commodity replenishment demand by using a sales volume prediction model, and planning an optimal distribution path by combining a path optimization algorithm.
2. The retail cabinet restocking analysis method based on the multi-dimensional decision matrix according to claim 1, wherein the construction formula of the multi-dimensional decision matrix is: in the formula, A matrix is scored for the matching degree of the commodity, The matrix is adapted for the characteristics of the locale, For the user feature matching matrix, The matrix is adapted for the hot-marketing of cities, The weight coefficients of the three matrixes respectively are satisfied ; The location feature adaptation matrix Elements of (2) Indicating the adaptation coefficient of the ith commodity and the jth place when Prohibit selling which indicates that the commodity is the j-th place, when When the goods are encouraged to display in the j-th place; The dynamic updating formula of the multi-dimensional decision matrix is as follows: in the formula, For the updated multi-dimensional decision matrix, 、、 The method comprises the steps of respectively updating a place feature adaptation matrix, a user feature matching matrix and an urban hot sale adaptation matrix, wherein the updating triggering conditions comprise place feature information change, user feature information iteration, urban hot sale commodity data update and prohibit selling commodity policy adjustment.
3. The multi-dimensional decision matrix-based retail cabinet restocking analysis method according to claim 1, wherein in the process of adjusting the shelf display positions of candidate commodities by dynamic allocation rules, a formula for implementing the dynamic allocation rules by a hierarchical weight calculation model is: in the formula, Respectively the weight coefficients of the appointed influence factors and satisfy ; For the characteristic parameters related to the sale of the commodity, Is the relevant characteristic parameters of commodity profit, Is a commodity aging related characteristic parameter; The value is the actual sales number of the commodity in a preset statistical period, The value is the ratio of the sales profit of the commodity to the sales income, The value is the remaining effective days of the commodity from the current date to the expiration date.
4. The multi-dimensional decision matrix-based retail cabinet replenishment analysis method according to claim 1, wherein the dynamic classification algorithm is an ABC dynamic classification algorithm, and the formula for classifying commodity categories by adopting the dynamic classification algorithm is as follows: in the formula, The frequency of the sale of the commodity is the duty ratio, For a historical sales volume of a single item, Sum of historical sales total for all commodities; When (when) When the commodity is classified into A-class high-frequency consumer commodity When the commodity is classified into class B medium frequency consumer commodity When the commodity is classified into C-class low-frequency consumption commodity; in the formula, And Is a dynamic classification threshold, and The threshold value is adjusted according to the historical sales volume change trend of all commodities.
5. The multi-dimensional decision matrix-based retail cabinet restocking analysis method according to claim 4, wherein the calculation formulas of the differentiated restocking period are respectively: Class a commodity replenishment cycle: A day; Type B commodity replenishment cycle: Tiantian (Chinese character of 'Tian') C-class commodity replenishment cycle: in the formula, Is the replenishment period of the class-C commodity, For the current inventory quantity of the commodity, The daily sales quantity of the commodity is calculated.
6. The multi-dimensional decision matrix-based retail sales analysis method of claim 1, wherein the sales volume prediction model is constructed by using an LSTM neural network, and the formula of the sales volume prediction model is: in the formula, Is a commodity sales predicted value in a preset period, As a mapping function of the LSTM neural network, For the historical sales data of the merchandise, For the seasonal variation to affect the factor, As the scene fluctuation influencing factor, Is a set correlation factor affecting sales of goods.
7. The multi-dimensional decision matrix-based retail cabinet replenishment analysis method according to claim 1, wherein the path optimization algorithm adopts an ant colony algorithm, and realizes distribution path planning by combining replenishment task aggregation rules, and a task aggregation judgment formula is: in the formula, For a straight line distance between two retail cabinets, And The geographic coordinates of the two retail cabinets, Radius is aggregated for tasks.
8. The retail cabinet restocking analysis method based on the multi-dimensional decision matrix according to claim 1, wherein the user feature of the service object is normalized to construct a user feature vector, and the normalization processing formula is: in the formula, Is a normalized value of the ith user feature, For this item of feature original value for a single user, As a minimum value for this feature, Is the maximum of this feature; The user feature vector is , For the normalized value of the age characteristic, For the occupational feature normalized value, In order to normalize the value of the frequency of purchase, Values are normalized for consumption preferences.
9. The multi-dimensional decision matrix-based retail sales analysis method of claim 1, wherein the update formula of the urban hot-mix commodity data is: in the formula, For the updated set of hot-sell products, To update a pre-update set of hot-sell products, The TOP10 hot-sell merchandise collection for the latest grab, Is a weight coefficient.
10. A retail cabinet restocking analysis method based on a multi-dimensional decision matrix, adopting the retail cabinet restocking analysis method based on a multi-dimensional decision matrix as claimed in any one of claims 1 to 9, characterized by comprising: The multi-dimensional decision matrix construction module is used for acquiring the place characteristics of the retail cabinets, the user characteristics of the service objects and the local city hot-sale commodity data and constructing a multi-dimensional decision matrix integrating the place characteristics, the user characteristics of the service objects and the local city hot-sale commodity data; The candidate commodity set determining module is used for completing commodity compliance screening and recommendation based on the multi-dimensional decision matrix and determining a candidate commodity set of the retail cabinet which is adaptive to the current scene and the user requirements; The candidate commodity display adjustment module is used for adjusting the shelf display position of the candidate commodity through a dynamic allocation rule by combining the sales performance and attribute characteristics of the candidate commodity; the commodity replenishment period analysis module is used for dividing commodity categories by adopting a dynamic classification algorithm according to the historical commodity sales frequency and making different replenishment periods for commodities of different categories; And the commodity replenishment demand prediction module is used for predicting commodity replenishment demands by utilizing the sales volume prediction model and planning an optimal distribution path by combining a path optimization algorithm.

Description

Retail cabinet replenishment analysis method and system based on multidimensional decision matrix Technical Field The invention relates to the technical field of unmanned vending equipment, in particular to a retail cabinet replenishment analysis method and system based on a multidimensional decision matrix. Background With the vigorous development of the smart retail industry, unmanned retail cabinets have widely covered multiple scenes such as transportation hubs, hospitals, office parks and the like by virtue of the advantages of flexible deployment and 24-hour service, and become an important supplementary form of the retail industry. However, there are still many technical shortboards for replenishment and operation management of the current retail cabinets, which are difficult to adapt to the dynamic market demands. First, the commodity display mode is cured, and the space utilization is low. The traditional retail cabinet adopts a static display strategy distributed in a fixed level, and the distribution is optimized without combining the characteristics of commodities and the difference of demands, so that hot commodities are inconvenient to touch, and the dead commodities occupy gold positions, thereby seriously influencing sales transformation. Secondly, the replenishment cycle lacks flexibility and the supply and demand matching is unbalanced. In the existing scheme, a ABC classification method with a fixed period or a threshold value needing to be manually adjusted is mostly adopted, dynamic factors such as sales fluctuation, seasonal variation and the like cannot be responded, so that the high-frequency commodity shortage rate is high, the conventional retail is high due to untimely replenishment according to industry data, and stock backlog is easy to occur for low-frequency commodities. Furthermore, the distribution resource allocation is inefficient. The characteristic of multipoint scattered deployment of the retail cabinets ensures that the existing single-point independent distribution mode does not carry out task aggregation and path optimization, so that the distribution vehicle has high empty rate and serious resource waste, and meanwhile, the system lacks a replenishment priority judging mechanism, thereby prolonging the response time. In addition, the compliance management response is lagged, and the risk is prevented and controlled inadequately. Different scenes have clear restrictions on commodity selling, but the existing mode relies on manual inspection to check out forbidden commodities, so that policy changes cannot be synchronized in real time, and compliance risks are easily caused. Disclosure of Invention Therefore, the invention provides a retail cabinet replenishment analysis method and a system based on a multidimensional decision matrix, which solve the problems of low space utilization rate caused by static display of the traditional retail cabinet, backlog or backlog caused by replenishment period rigidness, logistics distribution resource waste and lag in compliance management response. In order to achieve the purpose, the invention provides a retail cabinet replenishment analysis method based on a multidimensional decision matrix, which comprises the following steps: collecting the place characteristics of a retail cabinet, the user characteristics of a service object and local city hot-sale commodity data, and constructing a multi-dimensional decision matrix integrating the place characteristics, the user characteristics of the service object and the local city hot-sale commodity data; Completing commodity compliance screening and recommendation based on the multi-dimensional decision matrix, and determining a candidate commodity set of the retail cabinet adapting to the current scene and the user requirement; combining the sales performance and attribute characteristics of the candidate commodity, and adjusting the shelf display position of the candidate commodity through a dynamic allocation rule; Dividing commodity categories by adopting a dynamic classification algorithm according to the historical commodity sales frequency, and formulating different commodity replenishment periods for commodities of different categories; and predicting the commodity replenishment demand by using a sales volume prediction model, and planning an optimal distribution path by combining a path optimization algorithm. As a preferred scheme of the retail cabinet replenishment analysis method based on the multi-dimensional decision matrix, the construction formula of the multi-dimensional decision matrix is as follows: in the formula, A matrix is scored for the matching degree of the commodity,The matrix is adapted for the characteristics of the locale,For the user feature matching matrix,The matrix is adapted for the hot-marketing of cities,The weight coefficients of the three matrixes respectively are satisfied。 As a preferred scheme of the retail cabinet restocking analysis method b