Warehouse goods allocation layout optimization method and system based on big data visual analysis

Abstract

The invention relates to the technical field of data processing, in particular to a warehouse goods allocation layout optimization method and system based on big data visualization analysis, and the method comprises the steps: obtaining an access probability thermodynamic diagram of goods in recent periods of a warehouse and the same-batch warehouse-out category data, and according to a high-correlation goods combination determined based on the same batch of ex-warehouse category data, goods layout in a plurality of high-frequency turn-back routes determined based on the access probability thermodynamic diagram is adjusted, and after adjustment is completed, based on the smoothness degree of each arrangement sequence, the target sequence of goods placement in each high-frequency turn-back route is screened, and the target sequence of goods placement in each high-frequency turn-back route is obtained. And the optimization degree of each high-frequency turn-back route is calculated based on the smoothness degree of the target sequence and the length of the high-frequency turn-back route, so that the warehouse goods allocation layout is adjusted based on the target sequence of the high-frequency turn-back route with the relatively high optimization degree. According to the invention, the pick-up efficiency of a pick-up person in a warehouse can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing. More specifically, the present invention relates to a method and system for optimizing the layout of warehouse storage locations based on big data visualization analysis. Background Art

[0002] With the rapid development of the e-commerce and logistics industries, the operation mode of a large number of product categories and high-frequency inbound and outbound orders in large-scale warehousing scenarios has become the norm. However, there are some problems in the actual application of traditional warehousing management technologies: for example, congestion and blockage often occur in high-heat areas (areas with a high pick-up frequency) or high-heat paths, and there is a situation where goods are scattered, resulting in an increase in the ineffective moving distance, and thus a lower picking efficiency.

[0003] In related technologies, the above problems are often solved by combining big data visualization processing with intelligent algorithms. Specifically, on the one hand, based on a large amount of data related to the inbound and outbound of goods, the demand pattern of a certain product is predicted, and the goods are classified into ABC categories according to the inventory value and turnover rate of the goods themselves. The inventory items are divided into three levels: particularly important inventory (category A), generally important inventory (category B), and unimportant inventory (category C) according to the variety and the amount of occupied funds, and then different levels are managed and controlled respectively. On the other hand, the picking density is displayed through a warehouse heat map, and combined with a hybrid optimization genetic algorithm, the goods are stored in a concentrated manner to reduce the picking time of pickers.

[0004] However, the ABC classification method only relies on a single-dimensional index and ignores the order co-occurrence law between goods, resulting in a low correlation between goods on the same picking path, so that pickers need to pass through multiple paths to complete the picking of goods, and the picking path is redundant; and the hybrid optimization algorithm (such as the genetic algorithm) mostly generates a fixed layout based on a static data set and cannot respond to the impact of order fluctuations in real time. It may concentrate highly correlated goods, resulting in congestion when pickers pick up goods in the warehouse, and thus affecting the picking efficiency. Summary of the Invention

[0005] In order to solve the problem of low picking efficiency caused by redundant picking paths and the concentration of high-frequency goods, the present invention provides a method and system for optimizing the layout of warehouse storage locations based on big data visualization analysis.

[0006] According to a first aspect of the present invention, there is provided a method for optimizing the layout of warehouse storage locations based on big data visualization analysis, including: Obtaining an access probability heat map of goods in a target warehouse in the recent period and data on the categories of goods shipped in the same batch; Based on the access probability heat map, multiple high-frequency return routes are generated, and according to the outbound category data of the same batch, multiple highly correlated product combinations are generated. The high-frequency return routes are combined with the highly correlated product combinations to adjust the product layout in each high-frequency return route so that the highly correlated products are distributed in the same high-frequency return route; Obtain all the permutation orders of the products in the same high-frequency return route after adjusting the product layout, and calculate the smoothness of each permutation order according to the dispersion degree of the products with similar outbound frequencies, so as to take the permutation order with the maximum smoothness as the target order of the corresponding high-frequency return route; Calculate the optimization degree of each high-frequency return route. The optimization degree is positively correlated with the smoothness of the target order of the corresponding high-frequency return route and negatively correlated with the length of the corresponding high-frequency return route, so as to adjust the warehouse location layout according to the target order of the high-frequency return route with an optimization degree greater than the preset threshold.

[0007] The present invention combines the high-frequency return route with the highly correlated product combination and adjusts the product layout in each high-frequency return route, which can adjust the highly correlated products to the same high-frequency return route, reduce the redundancy of the pick-up path of the picker, and thus improve the pick-up efficiency; moreover, the present invention selects the target order by evaluating the smoothness of different permutation orders of the products in each high-frequency return route, which can avoid concentrating the high-frequency products and reduce the congestion degree, thereby further improving the pick-up efficiency. After adjusting the warehouse location layout according to the target order of the high-frequency return route selected based on the optimization degree, the pick-up effect can be significantly improved.

[0008] Preferably, calculating the smoothness of each permutation order according to the dispersion degree of the products with similar outbound frequencies includes: Take any high-frequency return route after adjusting the product layout as the target route, number the products distributed in the target route in any permutation order, and after numbering, cluster the numbered products based on the outbound frequencies of the products to obtain multiple clustering clusters; Calculate the variance of the data point numbers in each clustering cluster and sum them, and take the normalized value of the obtained sum as the smoothness of any permutation order.

[0009] Through the clustering algorithm, the present invention can quickly and accurately find the products with similar outbound frequencies, thus providing a reliable data basis for calculating the smoothness of each permutation order.

[0010] Preferably, take the variance of the data point numbers in each clustering cluster as the dispersion degree of the corresponding clustering cluster, and the smoothness of any permutation order satisfies the following relationship: ; In the formula, For the th high-frequency return route after adjusting the layout of goods, the th arrangement order's smoothness; When clustering the goods numbered according to the th arrangement order in the high-frequency return route, the th clustering cluster's dispersion degree; is the number of clustering clusters in the clustering result; is the normalization function.

[0011] Preferably, combine the high-frequency return route with the high-correlation goods combination, and adjust the layout of goods in each high-frequency return route, including: Mark the goods with recent outbound frequency greater than the preset frequency threshold in each high-frequency return route, and based on the high-correlation goods combination, obtain one or more associated goods of each marked good; For any marked good in any high-frequency return route, judge whether all the associated goods of any marked good exist in any high-frequency return route. If so, continue to judge the next marked good. Otherwise, record the missing associated goods as the goods to be moved in; Based on the comparison between the total number of goods to be moved in and the total number of vacant storage locations in the same high-frequency return route, adjust the layout of goods in each high-frequency return route according to the preset rules.

[0012] The method of adjusting the layout of goods in each high-frequency return route according to the present invention can place the high-correlation goods in the same high-frequency return route, thereby reducing the time spent on picking goods.

[0013] Preferably, adjusting the layout of goods in each high-frequency return route according to the preset rules includes: If the total number of goods to be moved in any high-frequency return route is greater than or equal to the total number of vacant storage locations, adjust the layout of goods in any high-frequency return route according to the first preset rule; If the total number of goods to be moved in any high-frequency return route is less than the total number of vacant storage locations, adjust the layout of goods in any high-frequency return route according to the second preset rule; Among them, the first preset rule is to move out the low-frequency goods in any high-frequency return route and completely move the goods to be moved in into any high-frequency return route; the second preset rule is to directly move the goods to be moved in into any high-frequency return route.

[0014] By performing the moving-in and moving-out operations in different ways according to the present invention, it can ensure that the goods to be moved in each high-frequency return route are completely moved into the corresponding high-frequency return route, thereby reducing the number of low-correlation goods in each high-frequency return route.

[0015] Preferably, adjusting the goods layout of any high-frequency return route according to the first preset rule includes: Calculating the difference between the total quantity of goods to be moved in and the total quantity of vacant storage locations in any high-frequency return route; In the order from the lowest to the highest outbound frequency, sequentially move out several goods from any high-frequency return route until the quantity of the moved-out goods is the same as the difference, so as to move the goods to be moved in into any high-frequency return route.

[0016] Preferably, calculating the optimization degree of each high-frequency return route, satisfying the following relational expression: ; In the formula, is the optimization degree of the th high-frequency return route; is the length of the th high-frequency return route; is the smoothness degree of the target order in the th high-frequency return route after adjusting the goods layout; is the ordinal number of the target order; is the normalization function.

[0017] In the present invention, adding 1 to the denominator can avoid the denominator being zero, ensuring the effectiveness of the optimization degree calculation formula.

[0018] Preferably, generating multiple high-frequency return routes based on the access probability heat map, including: Regarding each position in the access probability heat map as a node, and mapping the access probability of each node as a weight, the weight is negatively correlated with the access probability; Based on the weights of each node, using the Dijkstra algorithm to obtain a shortest high-frequency return path from the starting point to the ending point, and combining with the genetic algorithm, generating multiple paths to obtain multiple high-frequency return routes.

[0019] Preferably, generating multiple highly correlated goods combinations according to the same-batch outbound category data, including: According to the same-batch outbound category data, using the Apriori algorithm to generate multiple highly correlated goods combinations.

[0020] According to the second aspect of the present invention, there is provided a warehouse storage location layout optimization system based on big data visualization analysis. The system includes a memory and a processor. A computer program is stored on the memory, and the processor executes the computer program to implement the steps of the first aspect of the present invention.

[0021] The present invention has the following effects: On the one hand, the present invention adjusts the product categories in the high-frequency return route, concentrates the highly correlated products in the same route, can reduce the picking time, and improve the picking efficiency; on the other hand, by optimizing the product placement order and avoiding the concentrated distribution of products with similar frequencies, it can further improve the picking efficiency while ensuring the layout of highly correlated products. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 It is a schematic flow chart of the steps of the warehouse location layout optimization method based on big data visualization analysis in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0023] The following will describe in detail the specific embodiments of the present invention with reference to the accompanying drawings.

[0024] Refer to Figure 1 , the warehouse location layout optimization method based on big data visualization analysis includes steps S1 - S5, specifically as follows: S1: Obtain the access probability heat map of products in the target warehouse in the recent period and the same-batch outbound category data.

[0025] Among them, the target warehouse refers to any warehouse selected for analysis; the access probability heat map is generated by analyzing the access probability and picking distribution of products in the warehouse. The color of the heat map represents the size of the access probability, and the darker the color, the higher the access probability; the same-batch outbound category data refers to the number of categories included in the products in any outbound order. For example, when products A, B, and C are on the same outbound order, products A, B, and C are used as the same-batch outbound category data.

[0026] Specifically, the picking data and access data of products in the target warehouse in a recent period, such as within the recent month, can be collected. Then, calculate the access probability of each product (i.e., the proportion of the number of times each product is accessed to the total number of accesses) and the picking distribution (i.e., the proportion of the number of picking times at each product location to the total number of picking times), and synthesize them into a heat index (such as through weighted synthesis). Finally, generate a heat map based on each heat index to obtain the access probability heat map of products in the target warehouse in the recent period. It should be noted that the construction process of the heat map is an existing technology, and this embodiment will not elaborate on it here.

[0027] Optionally, the same-batch outbound category data of the target warehouse in the recent period, such as within one month, can be obtained by counting the outbound order information of the target warehouse in the recent period; or the same-batch outbound category data of the target warehouse in the recent period can be obtained through the warehouse management system. Among them, the time range for counting the same-batch outbound category data is the same as the time range for constructing the access probability heat map. This embodiment does not make a special limitation on the method for determining the same-batch outbound category data.

[0028] S2: Generate multiple high-frequency return routes based on the access probability heat map, and generate multiple highly correlated product combinations according to the product category data shipped in the same batch.

[0029] Among them, the high-frequency return route refers to the route with a relatively high access probability of the product; the highly correlated product combination refers to the product combination with a strong correlation. For example, when the correlation between product A, product B, and product C is relatively high, then product A, product B, and product C are used as a group of highly correlated product combinations.

[0030] In an exemplary embodiment of the present invention, the determination of the high-frequency return route can be achieved through the following steps: Take each position in the access probability heat map as a node, and map the access probability of each node to a weight, where the weight is negatively correlated with the access probability; based on the weights of each node, use the Dijkstra algorithm to obtain a shortest high-frequency return path from the starting point to the ending point, and combine with the genetic algorithm to generate multiple paths to obtain multiple high-frequency return routes.

[0031] It should be noted that since the access probability heat map is determined based on the access probability and picking distribution of the product, the obtained access probability heat map can reflect the access probability of the products at each position in the warehouse. Therefore, the present invention takes each position in the access probability heat map as a node.

[0032] It should be further noted that the Dijkstra algorithm itself can only generate one shortest path and cannot directly generate multiple routes. However, in practical applications, by combining other algorithms (such as A*, genetic algorithm, etc.) or improvement strategies (such as dynamically adjusting weights, bidirectional search, etc.), the generation of multiple routes can be indirectly achieved.

[0033] Specifically, the access probability of each node in the access probability heat map can be mapped to a weight (such as taking the reciprocal of the access probability as the weight, so that the nodes with a high access probability have a lower weight). After that, based on the determined weights, use the Dijkstra algorithm to find a shortest path starting from the node (starting point) located at the warehouse entrance in the access probability heat map to the node (ending point) located at the warehouse exit in the access probability heat map, so as to obtain the shortest high-frequency return path.

[0034] Furthermore, based on the obtained shortest high-frequency return path, multiple high-frequency return routes can be generated by combining the genetic algorithm or the A* algorithm. This process is prior art and will not be described in detail in this embodiment.

[0035] In an exemplary embodiment of the present invention, the determination of the highly correlated product combination can be achieved through the following steps: Generate multiple highly associated product combinations using the Apriori algorithm based on the product category data shipped in the same batch.

[0036] Specifically, all the product category data shipped in the same batch collected in S1 can be used as input. Then, set the minimum support to the lowest proportion (such as 20%) of the product combination appearing in all orders, and set the minimum confidence to 50%. After that, based on the two set thresholds (minimum support and minimum confidence), use the Apriori algorithm to iteratively generate frequent item sets. After several (such as 3) iterations, use the frequent item set generated in the last iteration as the highly associated product combination, thereby obtaining multiple highly associated product combinations.

[0037] It should be noted that the minimum support and the minimum confidence are professional terms in the Apriori algorithm, and the process of iteratively generating frequent item sets using the Apriori algorithm is prior art, and this embodiment will not be described in detail here.

[0038] S3: Combine the high-frequency return routes with the highly associated product combinations, and adjust the product layout in each high-frequency return route to distribute the highly associated products in the same high-frequency return route.

[0039] It should be noted that not all the products in the high-frequency return route are necessarily highly associated product combinations. There may also be some products with relatively low outbound frequencies or empty storage locations. In order to shorten the relevant picking time of the highly associated product combinations, the present invention adjusts the product layout in each high-frequency return route to place the highly associated products in the same high-frequency return route.

[0040] In an exemplary embodiment of the present invention, the adjustment of the product layout in each high-frequency return route can be achieved through the following steps: Step 1: Mark the products in each high-frequency return route with an outbound frequency greater than the preset frequency threshold in the recent period, and based on the highly associated product combinations, obtain one or more associated products for each marked product.

[0041] It should be noted that, for the convenience of screening marked products, the present invention normalizes the outbound frequencies of the products in each high-frequency return route in the recent period (such as within one month), so as to screen the products to be marked based on the normalized values. The outbound frequency here refers to the number of outbound times of any product in the recent period.

[0042] Optionally, the frequency threshold can be set to 0.7, so that based on this frequency threshold and the normalized values of the outbound frequencies of the products in each high-frequency return route in the recent period, the products to be marked can be screened and marked. After that, based on the highly associated product combinations determined in step S2, obtain one or more associated products for each marked product.

[0043] Step 2: For any marked item in any high-frequency return route, determine whether all associated items of any marked item exist in any high-frequency return route. If so, continue to judge the next marked item. Otherwise, record the missing associated items as items to be moved in.

[0044] Exemplarily, assume that the associated items of item A are item B and item C, and item A is a marked item in the th high-frequency return route. At this time, if item B is in the th high-frequency return route, but item C is not in the th high-frequency return route, then item C is recorded as the item to be moved into the th high-frequency return route, so that all items to be moved into each high-frequency return route can be determined.

[0045] Step 3: Based on the comparison between the total quantity of items to be moved into a high-frequency return route and the total quantity of available spaces, adjust the layout of items in each high-frequency return route according to a preset rule.

[0046] Among them, the total quantity of items to be moved in refers to the sum of the inventory quantities of all items to be moved into any high-frequency return route; the total quantity of available spaces refers to the number of items that can be accommodated in any high-frequency return route. And for the convenience of calculation, in the present invention, items of different categories are regarded as the same independent unit, without considering the size and dimensions of the items.

[0047] In an exemplary embodiment of the present invention, the layout of items in each high-frequency return route can be adjusted according to a preset rule through the following steps: (1) If the total quantity of items to be moved into any high-frequency return route is greater than or equal to the total quantity of available spaces, adjust the layout of items in any high-frequency return route according to the first preset rule; wherein, the first preset rule is to move out the items with low frequency in any high-frequency return route and completely move the items to be moved into any high-frequency return route; In an exemplary embodiment of the present invention, the layout of items can be adjusted according to the first preset rule through the following steps: Calculate the difference between the total quantity of items to be moved into any high-frequency return route and the total quantity of available spaces; screen several marked items from the marked items in any high-frequency return route in ascending order of the outbound frequency and move them out in turn until the quantity of items moved out is the same as the difference, so as to move the items to be moved into any high-frequency return route.

[0048] Exemplarily, the total quantity of items to be moved into the th high-frequency return route can be recorded as , record the total number of available storage locations in this high-frequency return route as , then the difference of this high-frequency return route is .

[0049] Then, for the th high-frequency return route, normalize the outbound frequencies of each item in the most recent month, and in ascending order, sequentially move out a number of items until the total quantity of the moved-out items is . Thus, the total number of available storage locations in this high-frequency return route is the same as the total quantity of the items to be moved into this high-frequency return route, so as to move the items to be moved into this high-frequency return route and realize the adjustment of the item layout in this high-frequency return route.

[0050] (2) If the total quantity of the items to be moved into any high-frequency return route is less than the total number of available storage locations, then adjust the item layout of any high-frequency return route according to the second preset rule; wherein, the second preset rule is to directly move the items to be moved into any high-frequency return route.

[0051] Optionally, after completing the operation of moving items into or out of each high-frequency return route according to the first preset rule or the second preset rule, items with high correlation can be distributed to the same high-frequency return route, so as to realize the adjustment of the item layout in each high-frequency return route.

[0052] S4: Obtain all the permutation orders of the items in the same high-frequency return route after adjusting the item layout, and calculate the smoothness of each permutation order according to the dispersion degree of the items with similar outbound frequencies, so as to take the permutation order with the greatest smoothness as the target order of the corresponding high-frequency return route.

[0053] It should be noted that the placement order of the items will also affect the picking efficiency of the picker. For example, when the high-frequency items are placed too densely, it will cause multiple pickers to gather in this area to pick up items at the same time, resulting in congestion. Therefore, the present invention evaluates the smoothness of each permutation order, so that the item placement order with the greatest smoothness can be selected to adjust the item placement method in each high-frequency return route after adjusting the item layout.

[0054] In an exemplary embodiment of the present invention, the smoothness of each item placement order can be determined through the following steps: (1) Take any high-frequency return route after adjusting the item layout as the target route, number the items distributed in any permutation order in the target route, and after numbering, cluster the numbered items based on the outbound frequencies of the items to obtain a plurality of clustering clusters; Specifically, clustering algorithms such as the K-means algorithm and the DBSCAN algorithm can be used to cluster all the numbered goods on the target route, so that the goods with similar outbound frequencies can be grouped into a clustering cluster, and multiple (such as 3) clustering clusters can be obtained. The number of clustering clusters obtained in this embodiment can be adjusted based on the actual scenario. Among them, the process of clustering using the clustering algorithm is a prior art, and this embodiment will not elaborate on it here.

[0055] It should be noted that the goods on the target route are numbered in the present invention to evaluate the distance between different goods through the number difference, so that the dispersion degree of the goods with similar outbound frequencies can be evaluated based on the distribution of the data point numbers within each clustering cluster.

[0056] (2) Calculate the variance of the data point numbers within each clustering cluster, sum them up, and use the normalized value of the obtained cumulative sum as the smoothness degree of any permutation order.

[0057] In an exemplary embodiment of the present invention, the variance of the data point numbers within each clustering cluster is used as the dispersion degree of the corresponding clustering cluster. Then, the smoothness degree of each permutation order satisfies the following relational expression: ; In the formula, is the smoothness degree of the th high-frequency return route after adjusting the goods layout, for the th permutation order; is the dispersion degree of the th clustering cluster obtained when clustering the goods numbered in the th permutation order in this high-frequency return route; is the number of clustering clusters in the clustering result. In this embodiment, ; is the normalization function.

[0058] Among them, The smaller the value of, the closer the numbers of the data points within the corresponding clustering cluster are, which further indicates that the high-frequency goods are more densely distributed under the corresponding permutation order. If the goods are placed in this permutation order, the congestion situation of the corresponding high-frequency return route is more serious, and the corresponding smoothness degree is lower; on the contrary, the larger the value, the more dispersed the high-frequency goods are distributed under the corresponding permutation order. If the goods are placed in this permutation order, the probability of congestion of the corresponding high-frequency return route is lower, and the corresponding smoothness degree is higher.

[0059] Optionally, the standard deviation of the data point numbers within the clustering cluster can also be calculated to evaluate the dispersion degree of the clustering cluster.

[0060] In another embodiment, instead of using the clustering method, goods with similar outbound frequencies can be obtained; alternatively, by directly setting a threshold, goods with an outbound frequency difference less than the set threshold are regarded as goods with similar outbound frequencies.

[0061] Furthermore, the smoothness calculation formula can be used to calculate the smoothness of the discharge order of each type of goods in each high-frequency return route. Then, the discharge order with the highest smoothness is used as the target order for the corresponding high-frequency return route. Here, the high-frequency return route is the high-frequency return route after adjusting the goods layout through step S3.

[0062] S5: Calculate the optimization degree of each high-frequency return route. The optimization degree is positively correlated with the smoothness of the target order of the corresponding high-frequency return route and negatively correlated with the length of the corresponding high-frequency return route. Then, adjust the warehouse storage location layout according to the target order of the high-frequency return route whose optimization degree is greater than the preset threshold.

[0063] Specifically, the optimization degree of each high-frequency return route satisfies the following relationship: ; In the formula, is the optimization degree of the th high-frequency return route; is the length of the th high-frequency return route; is the smoothness of the target order in the th high-frequency return route after adjusting the goods layout; is the ordinal number of the target order; is the normalization function.

[0064] Among them, with a smaller value indicates that the time spent by the picker when picking goods in the th high-frequency return route is smaller, which means the picker has a higher picking efficiency in this high-frequency return route, and the corresponding optimization degree is higher; with a larger value indicates that it is not easy for the picker to get stuck when picking goods in the th high-frequency return route after adjusting the goods layout, and the corresponding optimization degree of this high-frequency return route is higher.

[0065] Furthermore, when determining the optimization degree of each high-frequency return route, high-frequency return routes with an optimization degree greater than the preset threshold, such as 0.8, can be screened. Then, according to the target order of these screened high-frequency return routes (high-frequency return routes with an optimization degree greater than 0.8), adjust the placement order of the goods in the corresponding high-frequency return route, thereby achieving the optimization of the warehouse storage location layout.

[0066] The present invention also provides a warehouse location layout optimization system based on big data visual analysis. The system includes a memory and a processor, and a computer program is stored on the memory. The computer program integrates the functions of the warehouse location layout optimization method based on big data visual analysis. When the computer program is executed, the picking efficiency of pickers in the warehouse can be improved through the warehouse location layout optimization method based on big data visual analysis.

[0067] It should be understood that various alternative solutions to the embodiments of the present invention described herein can be adopted in the practice of the present invention.

Claims

1. A warehouse cargo space layout optimization method based on big data visualization analysis, characterized in that: include: Obtain the access probability heat map of the target warehouse's goods in the recent period, and the category data of the same batch of goods shipped; Based on the access probability heat map, multiple high-frequency return routes are generated, and according to the same batch of outbound product category data, multiple high-correlation product combinations are generated, the high-frequency return routes are combined with the high-correlation product combinations, and the product layout in each high-frequency return route is adjusted to distribute the highly correlated products in the same high-frequency return route; Obtain all the arrangement sequences of the goods in the same high-frequency return route after adjusting the goods layout, and calculate the smoothness of each arrangement sequence according to the discreteness of goods with similar outbound frequencies, so as to take the arrangement sequence with the greatest smoothness as the target sequence of the corresponding high-frequency return route; The optimization degree of each high-frequency return route is calculated. The optimization degree is positively correlated with the smoothness of the target sequence of the corresponding high-frequency return route, and negatively correlated with the length of the corresponding high-frequency return route, so as to adjust the warehouse cargo layout according to the target sequence of the high-frequency return routes with an optimization degree greater than a preset threshold.

2. The warehouse cargo space layout optimization method based on big data visualization analysis according to claim 1 is characterized in that: The calculation of the smoothness of each arrangement sequence based on the discreteness of the goods with similar delivery frequencies includes: Any high-frequency return route after adjusting the layout of the goods is used as the target route, and the goods distributed in any arrangement order in the target route are numbered. After the numbering is completed, the numbered goods are clustered based on the frequency of goods leaving the warehouse to obtain multiple clusters; The variance of the data point numbers in each cluster is calculated and summed, and the normalized value of the obtained cumulative sum is used as the smoothness of any one of the arrangement sequences.

3. The warehouse cargo space layout optimization method based on big data visualization analysis according to claim 2 is characterized in that: The variance of the data point numbers in each cluster is used as the discreteness of the corresponding cluster. The smoothness of any arrangement order satisfies the following relationship: ； In the formula, To adjust the product layout after Among the high-frequency return routes, The smoothness of the arrangement sequence; For this high-frequency return route, When clustering the goods numbered in the order of arrangement, the The degree of dispersion of clusters; is the number of clusters in the clustering results; is the normalization function.

4. The warehouse cargo space layout optimization method based on big data visualization analysis according to claim 1 is characterized in that: The step of combining the high-frequency return routes with the high-correlation product combinations and adjusting the product layout in each high-frequency return route includes: Mark the goods whose recent delivery frequency is greater than a preset frequency threshold in each high-frequency return route, and obtain one or more related goods of each marked goods based on the combination of highly related goods; For any marked product in any high-frequency return route, determine whether all associated products of the marked product exist in the high-frequency return route. If so, proceed to determine the next marked product. Otherwise, record the missing associated products as products to be moved in. Based on the total number of goods to be moved in and the total number of vacant storage spaces in the same high-frequency return route, the goods layout of each high-frequency return route is adjusted according to preset rules.

5. The warehouse cargo space layout optimization method based on big data visualization analysis according to claim 4 is characterized in that: The method of adjusting the goods layout of each high-frequency return route according to preset rules includes: If the total number of goods to be moved into any high-frequency return route is greater than or equal to the total number of vacant cargo spaces, the goods layout of any high-frequency return route is adjusted according to the first preset rule; If the total number of goods to be moved into any high-frequency return route is less than the total number of vacant cargo spaces, adjusting the goods layout of any high-frequency return route according to the second preset rule; Among them, the first preset rule is to move out the low-frequency goods in any of the high-frequency return routes, and completely move the goods to be moved in to any of the high-frequency return routes; the second preset rule is to move the goods to be moved in directly into any of the high-frequency return routes.

6. The warehouse cargo space layout optimization method based on big data visualization analysis according to claim 5 is characterized in that: The adjusting the product layout of any one of the high-frequency return routes according to the first preset rule includes: Calculate the difference between the total quantity of goods to be moved in and the total quantity of vacant cargo spaces in any high-frequency return route; In order of the frequency of outbound delivery from low to high, a number of goods are removed from any of the high-frequency return routes in sequence until the number of goods removed is the same as the difference amount, so as to move the goods to be removed into any of the high-frequency return routes.

7. The warehouse cargo space layout optimization method based on big data visualization analysis according to claim 1 is characterized in that: The optimization degree of each high-frequency return route is calculated to satisfy the following relationship: ； In the formula, For the The degree of optimization of the high-frequency return routes; For the The length of the high-frequency return route; To adjust the product layout after The smoothness of the target sequence in the high-frequency return route; is the ordinal number of the target sequence; is the normalization function.

8. The warehouse cargo space layout optimization method based on big data visualization analysis according to claim 1 is characterized in that: The generating of a plurality of high-frequency return routes based on the access probability heat map includes: Each position in the access probability heat map is regarded as a node, and the access probability of each node is mapped to a weight, where the weight is negatively correlated with the access probability; Based on the weight of each node, the Dijkstra algorithm is used to obtain a high-frequency return shortest path from the starting point to the end point, and combined with the genetic algorithm, multiple paths are generated to obtain multiple high-frequency return routes.

9. The warehouse cargo space layout optimization method based on big data visualization analysis according to claim 1 is characterized in that: The generating of multiple highly correlated product combinations based on the same batch of shipped product category data includes: Based on the same batch of outbound product category data, the Apriori algorithm is used to generate multiple high-correlation product combinations.

10. Warehouse cargo layout optimization system based on big data visualization analysis, characterized by: The warehouse cargo location layout optimization system based on big data visualization analysis includes a memory and a processor, the memory stores a computer program, and the processor executes the computer program to implement the steps of the warehouse cargo location layout optimization method based on big data visualization analysis as described in any one of claims 1 to 9.

Images