Warehouse location dynamic assignment method based on cargo turnover rate prediction

Abstract

The invention discloses a warehouse location dynamic assignment method based on cargo turnover rate prediction, and relates to the field of warehouse location dynamic assignment. The method comprises the steps that an input feature set and an analysis data set are acquired and preprocessed, and the preprocessed input feature set and the preprocessed analysis data set are generated; based on the preprocessed input feature set, constructing a cargo turnover rate prediction model to obtain a cargo turnover rate prediction value; and based on the preprocessed analysis data set and the cargo turnover rate prediction value, in combination with constraint conditions including capacity, function and priority constraints, selecting an optimal storage location for cargo assignment through an optimized storage location dynamic assignment algorithm. The problems that the existing cargo turnover rate prediction precision is insufficient, the storage location distribution lacks the'cargo turnover rate prediction value-real-time state 'dynamic association, and the circulation efficiency and the space utilization rate cannot be balanced are solved.

Description

Technical Field

[0001] This invention relates to the field of dynamic allocation of warehouse locations, and more particularly to a method for dynamic allocation of warehouse locations based on cargo turnover rate prediction. Background Technology

[0002] With the rapid development of e-commerce retail, intelligent manufacturing, and other industries, warehousing and logistics, as a core link in the supply chain, directly impact a company's market response speed and overall cost control through their operational efficiency. Current warehouse location allocation technology has gone through three development stages: "manual experience allocation," "static rule allocation," and "preliminary intelligent optimization." However, significant technical bottlenecks remain in complex and ever-changing real-world scenarios, making it difficult to adapt to the diverse and high-frequency fluctuating warehousing demands. Specifically, these limitations manifest in two aspects: First, turnover rate prediction relies solely on a single time-series feature or linear model, failing to capture non-linear patterns such as seasonal fluctuations and sudden demand changes, resulting in insufficient prediction accuracy. Second, location allocation lacks a dynamic adjustment mechanism, making it difficult to balance real-time storage needs with cargo turnover efficiency, leading to low warehouse space utilization and high cargo storage and retrieval costs. While existing technologies attempt to introduce time-series models to optimize predictions, they do not incorporate environmental factors and incremental feature changes, and location allocation lacks a dynamic correlation between predicted values ​​and location suitability, failing to meet the efficient management needs of complex warehousing scenarios. Summary of the Invention

[0003] This invention provides a method for dynamic allocation of warehouse locations based on cargo turnover rate prediction. It addresses the problems of insufficient accuracy in existing cargo turnover rate prediction due to neglecting nonlinear fluctuations and multi-dimensional environmental factors, as well as the lack of dynamic correlation between "cargo turnover rate prediction value and real-time status" in location allocation, which makes it impossible to balance circulation efficiency and space utilization. Through nonlinear modeling and dynamic optimization algorithms, it achieves accurate and real-time adjustment of warehouse location allocation.

[0004] The present invention provides a method for dynamic allocation of warehouse locations based on cargo turnover rate prediction, comprising the following steps: S1. Acquire and preprocess the input feature set and analysis dataset to generate the preprocessed input feature set and analysis dataset; based on the preprocessed input feature set, construct a cargo turnover rate prediction model to obtain the predicted cargo turnover rate value; S2. Based on the preprocessed analysis dataset and the predicted cargo turnover rate, combined with constraints including capacity, function, and priority, the optimal storage location is selected for cargo assignment through an optimized dynamic storage location assignment algorithm.

[0005] Preferably, S1 specifically includes: Historical time-series features are obtained and all historical time-series features are output as a structured historical time-series feature set in the format of "goods code + entry / exit time". Based on the structured historical time-series feature set, the incremental change feature set is calculated. At the same time, temperature and humidity features, holiday features and seasonal features are obtained to form an environmental impact feature set.

[0006] Preferably, S1 specifically includes: The input feature set consists of a structured historical time series feature set, an incremental change feature set, and an environmental impact feature set.

[0007] Preferably, S1 specifically includes: Based on the preprocessed input feature set, a cargo turnover rate prediction model is constructed using logistic regression terms.

[0008] Preferably, S1 specifically includes: In the process of implementing the cargo turnover rate prediction model, the predicted cargo turnover rate is generated by introducing a decay factor and an interaction term between preprocessed environmental impact features and historical time series features.

[0009] Preferably, S2 specifically includes: The comprehensive cost of each storage location is calculated based on the metrics of storage demand, the cost of goods turnover matching, and the physical cost of the storage location.

[0010] Preferably, S2 specifically includes: Among all storage locations, the one with the lowest overall cost is selected as the optimal storage location.

[0011] Preferably, S2 specifically includes: When there are two or more storage locations with the same overall cost, the amount of available space in the storage location will be used as a secondary screening criterion.

[0012] The beneficial effects of the technical solution of the present invention are: 1. Breaking through the technical bottleneck of the existing "static and singular" warehouse location assignment, it provides a full-process intelligent solution of "prediction-optimization-execution-feedback", enriching the technical system of intelligent warehouse logistics management.

[0013] 2. Significantly improves warehouse space utilization and cargo turnover efficiency, reduces storage and retrieval costs, and is especially suitable for warehousing scenarios with large demand fluctuations and complex cargo types, with broad prospects for industrial application.

[0014] 3. It can reduce warehousing and operating costs for enterprises, including space rental costs and labor costs for storage and retrieval, improve supply chain response speed, and enhance market competitiveness. Attached Figure Description

[0015] Figure 1 This is a flowchart of a method for dynamic allocation of warehouse locations based on cargo turnover rate prediction, as described in this invention. Detailed Implementation

[0016] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0018] The following description, in conjunction with the accompanying drawings, details a specific scheme for a dynamic warehouse location allocation method based on cargo turnover rate prediction provided by this invention.

[0019] See attached document Figure 1 The diagram illustrates a flowchart of a dynamic warehouse location allocation method based on cargo turnover rate prediction, provided by an embodiment of the present invention. The method includes the following steps: S1. Acquire and preprocess the input feature set and analysis dataset to generate the preprocessed input feature set and analysis dataset; based on the preprocessed input feature set, construct a cargo turnover rate prediction model to obtain the predicted cargo turnover rate value; First, real-time data acquisition and historical time-series feature acquisition are performed. Specifically, historical time-series features are obtained through the warehouse's existing warehouse management system (WMS), including goods turnover rate, average storage time, category-related turnover rate, goods category, goods code, and inbound / outbound time. All historical time-series features are organized and output as a structured historical time-series feature set in the format of "goods code + inbound / outbound time". Based on the structured historical time-series feature set, the incremental change feature set is calculated using the adjacent time point difference method. Environmental impact features are collected collaboratively by multiple devices. The environmental impact feature set consists of environmental impact features such as temperature and humidity features, holiday features, and seasonal features. Among them, temperature and humidity features are acquired in real time by temperature and humidity sensors in the warehouse. Holiday features are binarized and encoded after calling a third-party calendar interface, with holidays set to 1 and non-holidays set to 0. Seasonal features are encoded according to manually configured rules: spring (March-May) is set to 1, summer (June-August) to 2, autumn (September-November) to 3, and winter (December-February) to 4. The final input feature set includes the structured historical time-series feature set, the incremental change feature set, and the environmental impact feature set. Meanwhile, based on the real-time data collected by the warehouse management system, such as storage demand, inventory turnover rate, available space, distance from the warehouse to the exit, and warehouse floor height, the data is used as the analysis dataset for dynamic allocation analysis of warehouse locations.

[0020] Preprocessing is performed on the input feature set and analysis dataset. As a specific example, the 3σ criterion is used to remove outliers, linear interpolation is used to fill in continuous missing values, Z-score standardization is used to eliminate dimensional differences, and one-hot encoding is used to convert categorical features, such as goods type, holiday features, and seasonal features, into numerical values. The final output is the preprocessed input feature set and analysis dataset.

[0021] Based on the preprocessed input feature set, a cargo turnover rate prediction model is constructed to obtain the predicted cargo turnover rate. The cargo turnover rate prediction model is based on the following formula:

[0022] in, Indicates in The predicted value of cargo turnover rate at time t, with a value range of [0,1]; Indicates the current moment; Indexing historical moments; For the preprocessed first A historical time series characteristic at time The value; For the preprocessed first A historical time series characteristic at time The incremental change value is used to capture the volatility of this historical time series characteristic. , For adjacent time intervals, For the preprocessed first A historical time series characteristic at time The value; and The weighting coefficients for each preprocessed structured historical time series feature and each preprocessed incremental change feature are obtained by training based on the preprocessed input feature set using a gradient descent algorithm, such as stochastic gradient descent, and then by Min-Max normalization. The value range is [0,1]. Indicates the first preprocessed... The time decay coefficient of each historical time series feature is obtained through cross-validation of the preprocessed input feature set. For the preprocessed first The key periodic points of the historical time series features are obtained through autocorrelation analysis based on the preprocessed historical time series features. It is a time scaling factor, used to control the intensity of the impact of time difference on the cargo turnover rate prediction model; The total number of preprocessed historical time-series features used in cargo turnover rate prediction; An index for preprocessed historical time-series features; This is a combined contribution of preprocessed historical time series characteristics and short-term fluctuations; It is the attenuation factor; For the first Environmental impact characteristics, such as temperature and humidity characteristics, and holiday characteristics, at any given time The value at time; The total number of environmental impact characteristics used in the prediction of cargo turnover rate; An index for environmental impact characteristics; In order to be in The first time after time preprocessing Historical time series characteristics With the Environmental impact characteristics The interaction coefficients between them are obtained by training the preprocessed input feature set through a gradient descent algorithm, such as stochastic gradient descent, followed by regularization constraints and Min-Max normalization. Regularization constraints include L2 regularization. This is the interaction term between the preprocessed environmental impact features and historical time series features. It can capture the nonlinear relationship between the environmental impact features and historical time series features. For example, the goods category in the historical time series features is fresh produce, and the temperature in the environmental impact features is high temperature. The combined effect of the two leads to a decrease in the turnover rate of this goods. This is a logistic regression term used to control the range of values ​​for the output result.

[0023] S2. Based on the preprocessed analysis dataset and the predicted cargo turnover rate, combined with constraints including capacity, function, and priority, the optimal storage location is selected for cargo assignment through an optimized dynamic storage location assignment algorithm.

[0024] Before implementing dynamic allocation of storage locations, input data preparation and constraint definition are performed. The input data includes preprocessed storage location demand, storage location inventory turnover rate, storage location free space, distance from the storage location to the exit, and storage location floor height. This is the preprocessed analysis dataset, along with the predicted inventory turnover rate. In addition, constraints need to be clearly defined, including capacity constraints, functional constraints, and priority constraints: capacity constraints require that the amount of available space in the storage location is greater than the volume of goods stored; functional constraints stipulate that refrigerated goods can only be assigned to cold chain storage locations, and fragile goods can only be assigned to storage locations on no more than the second floor; priority constraints require that goods urgently needing to be shipped out should be assigned to storage locations near the outbound exit.

[0025] Next, based on the input data and constraints, an optimized dynamic storage location allocation algorithm is used to select the optimal storage location. The optimization formula is:

[0026] in, The comprehensive cost of the optimal storage location is calculated by taking the comprehensive cost of each storage location and selecting the storage location with the lowest comprehensive cost as the optimal storage location. The storage location with the lowest comprehensive cost is the optimal storage location. This is the set of available storage locations within the warehouse that satisfy all constraints. For storage location index; , These are the weighting coefficients for storage demand and inventory turnover rate, respectively. They are obtained through linear regression analysis based on historical warehouse operation data from the warehouse management system, and their values ​​range from [0,1]. ; This represents the pre-processed storage requirements for warehouse locations; Pre-processed storage location At any moment The warehouse inventory turnover rate is obtained from the preprocessed analysis dataset; To avoid the zero-value problem and ensure computational stability, it is a constant, taking values ​​such as... ; This is a metric for storage demand in warehouse locations. Warehouse locations with high storage demand have higher usage frequency and higher storage costs. It is used to quantify the deviation between the predicted value of cargo turnover rate and the actual turnover rate of the warehouse location. Its function is to convert the deviation into a calculable operating cost, so as to help to prioritize the selection of warehouse locations with suitable turnover rates when assigning warehouse locations, and reduce the extra time and resource losses caused by poor matching. Matching cost items to inventory turnover rate: When inventory turnover rate is not well matched, the warehouse needs to spend more time and resources to allocate inventory or update inventory, resulting in increased operating costs. It is a weighting coefficient used to adjust the physical storage cost. The initial value is set according to the actual storage environment and business needs and optimized by random forest. The value range is [0,1]. This is the distance from the pre-processed storage location to the outlet. This refers to the height of the storage location after preprocessing. , These are the unit distance handling cost and the unit floor height handling cost, respectively. They are calculated using a genetic algorithm based on historical warehouse data and corresponding actual operating costs in the warehouse management system. Historical warehouse data includes records of past goods handling distances for different storage locations, and data on the manpower input time and equipment energy consumption for each handling operation. This is the physical cost item for warehouse locations, which reflects the physical storage and retrieval costs of warehouse locations and takes into account the actual impact of the handling distance and floor height of warehouse locations on warehouse operations. Indicates the set of available storage locations In this process, the storage location that minimizes the overall cost is selected.

[0027] Ultimately, the optimized dynamic warehouse location allocation algorithm calculates the comprehensive cost of each warehouse location by considering the differences in warehouse storage demand and goods turnover rate, as well as the actual physical cost, and selects the optimal warehouse location for goods allocation. If multiple warehouse locations with the same cost exist, the amount of available space in the warehouse location will be used as a secondary screening criterion, prioritizing the warehouse location with more available space to improve the utilization rate of warehouse space.

[0028] In summary, a method for dynamic allocation of warehouse locations based on cargo turnover rate prediction has been developed.

[0029] The order of the embodiments is for illustrative purposes only and does not represent the superiority or inferiority of the embodiments. The processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0030] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0031] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for dynamic allocation of warehouse locations based on cargo turnover rate prediction, characterized in that, Includes the following steps: S1. Obtain and preprocess the input feature set and analysis dataset to generate the preprocessed input feature set and analysis dataset; Based on the preprocessed input feature set, a cargo turnover rate prediction model is constructed to obtain the predicted cargo turnover rate. S2. Based on the preprocessed analysis dataset and the predicted cargo turnover rate, combined with constraints including capacity, function, and priority, the optimal storage location is selected for cargo assignment through an optimized dynamic storage location assignment algorithm.

2. The method for dynamic allocation of warehouse locations based on cargo turnover rate prediction according to claim 1, characterized in that, S1 specifically includes: Historical time-series features are obtained and all historical time-series features are output as a structured historical time-series feature set in the format of "goods code + entry / exit time". Based on the structured historical time-series feature set, the incremental change feature set is calculated. At the same time, temperature and humidity features, holiday features and seasonal features are obtained to form an environmental impact feature set.

3. The method for dynamic allocation of warehouse locations based on cargo turnover rate prediction according to claim 2, characterized in that, S1 specifically includes: The input feature set consists of a structured historical time series feature set, an incremental change feature set, and an environmental impact feature set.

4. The method for dynamic allocation of warehouse locations based on cargo turnover rate prediction according to claim 1, characterized in that, S1 specifically includes: Based on the preprocessed input feature set, a cargo turnover rate prediction model is constructed using logistic regression terms.

5. The method for dynamic allocation of warehouse locations based on cargo turnover rate prediction according to claim 4, characterized in that, S1 specifically includes: In the process of implementing the cargo turnover rate prediction model, the predicted cargo turnover rate is generated by introducing a decay factor and an interaction term between preprocessed environmental impact features and historical time series features.

6. The method for dynamic allocation of warehouse locations based on cargo turnover rate prediction according to claim 1, characterized in that, S2 specifically includes: The comprehensive cost of each storage location is calculated based on the metrics of storage demand, the cost of goods turnover matching, and the physical cost of the storage location.

7. The method for dynamic allocation of warehouse locations based on cargo turnover rate prediction according to claim 6, characterized in that, S2 specifically includes: Among all storage locations, the one with the lowest overall cost is selected as the optimal storage location.

8. The method for dynamic allocation of warehouse locations based on cargo turnover rate prediction according to claim 7, characterized in that, S2 specifically includes: When there are two or more storage locations with the same overall cost, the amount of available space in the storage location will be used as a secondary screening criterion.

Images