Cold-chain shelf position optimization method and system based on digital twinning

Abstract

The application relates to the technical field of intelligent vertical storage, and discloses a cold-chain shelf position optimization method and system based on digital twinning, which comprises the following steps: extracting historical order data to obtain a historical order data set; performing goods purchase correlation analysis on the historical order data set to obtain a correlation feature set of a purchase correlation goods pair; calculating a spatial adjacency coefficient of the purchase correlation goods pair according to the correlation feature set; establishing a digital twinning model of a cold-chain vertical storage; planning a shelf position in the digital twinning model according to the spatial adjacency coefficient to obtain a plurality of shelf position adjustment schemes; evaluating the plurality of shelf position adjustment schemes based on path planning in the digital twinning model, and obtaining an optimal shelf position layout according to an evaluation result; and executing the optimal shelf position layout in a physical domain cold-chain vertical storage. The method and system solve the limitation that a cold-chain vertical storage relies on manual experience to perform shelf position layout, and significantly reduce order picking distance and time.

Description

Technical Field

[0001] This invention relates to the field of intelligent automated warehouse technology, and in particular to a method and system for optimizing the location of cold chain shelves based on digital twins. Background Technology

[0002] Cold chain logistics is a crucial part of the modern logistics system, especially in the transportation and storage of temperature-controlled products (such as fresh food). The efficiency of cold chain logistics directly impacts product quality and distribution costs. Currently, automated cold chain warehouses (AS / RS), as an important component of cold chain logistics, play a vital role in improving warehouse space utilization, reducing material picking time, minimizing cargo damage, and enhancing overall operational efficiency through the rational planning of their shelving locations.

[0003] However, traditional cold chain automated warehouse management often relies on manual experience for shelf placement, lacking data-driven scientific optimization methods. This approach can lead to low space utilization and operational inconvenience, and the lack of systematic optimization thinking can also negatively impact overall work efficiency and the quality of warehouse environment management. With the continuous development of information technology and the Internet of Things (IoT), digital twin technology is gradually emerging in the industrial field, becoming an effective tool for improving warehouse management and optimizing resource allocation.

[0004] Digital twin technology creates virtual models of physical objects, enabling real-time reflection and analysis of warehouse operations. This helps decision-makers simulate different scenarios in a virtual environment, optimizing various warehouse management metrics. Particularly in automated cold chain warehouses (AS / RS), digital twin technology can optimize shelf placement, rationally plan warehouse space, and improve operational and retrieval efficiency. Currently, AS / RS optimization solutions based on digital twin technology are still in the exploratory stage, focusing primarily on improving warehouse management and operational efficiency, but lacking specific shelf placement optimization strategies and implementation paths in practical applications. Summary of the Invention

[0005] Therefore, the purpose of this invention is to overcome the limitations of cold chain warehouse shelf layout in the prior art and provide a method and system for optimizing cold chain shelf location based on digital twins, which significantly reduces order picking distance and time.

[0006] In a first aspect, to solve the above-mentioned technical problems, the present invention provides a method for optimizing the location of cold chain shelves based on digital twins, including:

[0007] Extract historical order data to obtain a historical order dataset; the historical order data includes order time, product category, quantity, and storage method;

[0008] Perform a product purchase correlation analysis on the historical order dataset to obtain a set of correlation features for purchase-related product pairs;

[0009] Calculate the spatial adjacency coefficient of the purchased related goods pair based on the set of associated features;

[0010] A digital twin model of a cold chain automated warehouse is established, and the shelf positions are planned in the digital twin model according to the spatial adjacency coefficient to obtain multiple shelf position adjustment schemes;

[0011] In the digital twin model, multiple shelf location adjustment schemes are evaluated based on path planning, and the optimal shelf location layout is obtained based on the evaluation results;

[0012] The optimal shelf location layout is executed in the physical domain cold chain automated warehouse.

[0013] In one embodiment of the present invention, a product purchase association analysis is performed on the historical order dataset to obtain a set of association features for purchase-related product pairs. This includes assigning a time decay weight to each historical order data point to obtain a weighted order dataset; mining frequent itemsets of product combinations in the weighted order dataset layer by layer based on a prior algorithm to obtain a set of candidate association rules; performing multi-indicator screening on the set of candidate association rules to obtain a preliminary set of screening rules; the screening indicators include support, confidence, and lift; performing density clustering analysis on the preliminary set of screening rules to obtain a set of strong association rules; and extracting the set of association features for purchase-related product pairs from the set of strong association rules.

[0014] In one embodiment of the present invention, the historical order dataset is subjected to a product purchase correlation analysis to obtain a set of correlation features for purchase-related product pairs. The method further includes calculating the correlation strength of each product pair in the set of strong correlation rules; marking product pairs with correlation strength greater than or equal to a preset correlation strength as purchase-related product pairs; and mapping the correlation rules of the purchase-related product pairs to dynamic connection lines between virtual shelves in the digital twin model; wherein the line width of the connection lines is positively correlated with the correlation strength.

[0015] In one embodiment of the present invention, the spatial adjacency coefficient of the purchase-related goods pair is calculated based on the set of associated features, including each associated feature in the set of associated features including the support, confidence, and lift of the purchase-related goods pair; the spatial adjacency coefficient is obtained based on the support, confidence, and lift.

[0016] In one embodiment of the present invention, calculating the spatial adjacency coefficient of the purchased related goods pair further includes, if the purchased related goods pair has a bidirectional association feature, merging the bidirectional spatial adjacency coefficients of the purchased related goods pair to obtain a merged spatial adjacency coefficient; using the merged spatial adjacency coefficient as the actual spatial adjacency coefficient; wherein the merged spatial adjacency coefficient is obtained based on the following method:

[0017] ;

[0018] Among them, W XY ' represents the spatial adjacency coefficient of the purchase of related goods pair XY; X and Y represent the goods purchased in the related goods pair; W XY W represents the spatial adjacency coefficient of X as the premise product and Y as the conclusion product. YX This represents the spatial adjacency coefficient of Y as the premise item and X as the conclusion item; This represents the attenuation coefficient.

[0019] In one embodiment of the present invention, the shelf location is planned in the digital twin model according to the spatial adjacency coefficient, including determining whether the storage methods of the purchased related goods pair are the same: if the storage methods are the same, the shelves of the purchased related goods pair are adjusted to storage units in the same logical partition according to the spatial adjacency coefficient; if the storage methods are different, the shelves of the purchased related goods pair are adjusted to storage units in different logical partitions according to the spatial adjacency coefficient.

[0020] In one embodiment of the present invention, adjusting the shelves of the purchased related goods pairs to storage units in the same logical partition according to the spatial adjacency coefficient includes classifying the purchased related goods pairs into strongly related goods pairs, moderately related goods pairs, and weakly related goods pairs according to the spatial adjacency coefficient; in the digital twin model, adjusting the shelves of the strongly related goods pairs to adjacent storage units; adjusting the shelves of the moderately related goods pairs to non-adjacent storage units; and filling empty spaces with shelves for the weakly related goods pairs; wherein, the adjacent storage units are defined as nodes directly connected in the virtual shelf topology; and the non-adjacent storage units are defined as nodes connected in the virtual shelf topology through at least one intermediate node.

[0021] In one embodiment of the present invention, in the digital twin model, the rationality of the multiple shelf location adjustment schemes is evaluated based on path planning to obtain the optimal shelf location layout. This includes loading multiple shelf location adjustment schemes into the digital twin model; for each shelf location adjustment scheme, simulating AGVs to perform picking of multiple order data in the digital twin model to obtain picking evaluation results of the multiple shelf location adjustment methods; the evaluation indicators of the picking evaluation results include picking distance, temperature control stability, and AGV energy consumption; and extracting the shelf location adjustment scheme with the best picking evaluation results as the optimal shelf location layout.

[0022] In one embodiment of the present invention, the AGV performs picking according to an effective path. Determining the effective path includes: generating AGV picking paths based on a genetic algorithm in the digital twin model to obtain a candidate path set; introducing a path planning cost function and calculating the average path cost and standard deviation of the candidate paths; filtering effective paths from the candidate path set according to the average path cost and standard deviation to obtain a filtered path set; sorting the path costs of the filtered path set in descending order, and extracting the first path as the effective path.

[0023] Secondly, based on the same inventive concept, to solve the aforementioned technical problems, this invention provides a cold chain shelf location optimization system based on digital twins, comprising:

[0024] The data extraction module extracts historical order data and obtains a historical order dataset, which includes order time, product category, quantity, and storage method.

[0025] The correlation analysis module performs product purchase correlation analysis on the historical order dataset to obtain a set of correlation features for purchase-related product pairs;

[0026] The adjacency coefficient calculation module calculates the spatial adjacency coefficient of the purchased related goods pair based on the set of associated features;

[0027] The digital twin modeling module is used to build digital twin models of cold chain automated warehouses;

[0028] The intelligent planning module interacts with the adjacency coefficient calculation module and the digital twin modeling module, and is configured to generate multiple shelf location adjustment schemes in the digital twin model based on the spatial adjacency coefficient.

[0029] The scheme evaluation module is configured to evaluate the multiple shelf location adjustment schemes based on path planning in the digital twin model, and determine the optimal shelf location layout based on multi-objective optimization.

[0030] The drive execution module is configured to convert the optimal shelf location layout into warehouse equipment control commands, and drive the cold chain automated storage and retrieval system shelves to adjust according to the optimal shelf location layout.

[0031] Compared with the prior art, the above-described technical solution of the present invention has the following advantages:

[0032] The cold chain shelving location optimization method and system based on digital twins described in this invention provides optimized solutions for shelving location layout based on actual needs through purchase correlation analysis and spatial adjacency coefficient calculation of historical order data. This significantly reduces order picking distance and time, and lowers operating costs. Furthermore, different shelving layout schemes are simulated and verified using digital twin models, ensuring the scientific validity and operability of the final solution while reducing implementation risks.

[0033] By exploring the correlation between goods purchases and quantifying spatial adjacency coefficients, the shelves are adjusted to store frequently occurring goods nearby, significantly reducing picking paths, reducing the travel distance and operation time of stacker cranes or AGV equipment, and reducing the empty running rate of equipment and the risk of temperature fluctuations in the cold chain environment.

[0034] By dynamically adjusting the shelf layout based on historical order data characteristics, the rigidity of static zoning strategies can be avoided, allowing for flexible responses to seasonal order fluctuations or changes in product category demand, thereby improving the utilization rate of warehouse resources.

[0035] By using digital twin models to perform virtual simulation and path planning evaluation of multiple shelving adjustment schemes, the optimal layout scheme can be quickly selected before physical implementation, reducing the number of actual handling operations and trial and error costs, while also reducing cold energy loss caused by frequent inbound and outbound operations in cold chain warehousing. Attached Figure Description

[0036] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings, wherein...

[0037] Figure 1 This is a flowchart of a cold chain shelf location optimization method based on digital twins in a preferred embodiment of the present invention;

[0038] Figure 2 This is a flowchart illustrating the process of obtaining the associated feature set in a preferred embodiment of the present invention;

[0039] Figure 3 This is a flowchart of obtaining the spatial adjacency coefficient in a preferred embodiment of the present invention;

[0040] Figure 4 This is a flowchart illustrating the planning of shelf locations based on spatial adjacency coefficients in a preferred embodiment of the present invention.

[0041] Figure 5 This is a flowchart illustrating the optimal shelf location layout in a preferred embodiment of the present invention;

[0042] Figure 6 This is a flowchart for determining an effective path in a preferred embodiment of the present invention;

[0043] Figure 7This is a structural block diagram of a cold chain shelf location optimization system based on digital twins in a preferred embodiment of the present invention. Detailed Implementation

[0044] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand and implement the present invention. However, the embodiments described are not intended to limit the present invention.

[0045] Example 1: The purpose of this embodiment of the invention is to overcome the limitations of cold chain automated storage and retrieval systems (AS / RS) relying on manual experience for shelf placement. (Refer to...) Figure 1 As shown, this embodiment of the invention discloses a method for optimizing the location of cold chain shelves based on digital twins, including:

[0046] S10. Extract historical order data to obtain a historical order dataset; the historical order data includes order time, product category, quantity, and storage method;

[0047] S20. Perform a product purchase correlation analysis on the historical order dataset to obtain a set of correlation features for purchase-related product pairs;

[0048] S30. Calculate the spatial adjacency coefficient of the purchased related goods pair based on the set of associated features;

[0049] S40. Establish a digital twin model of the cold chain automated warehouse, and plan the shelf positions in the digital twin model according to the spatial adjacency coefficient to obtain multiple shelf position adjustment schemes;

[0050] S50. In the digital twin model, the multiple shelf location adjustment schemes are evaluated based on path planning, and the optimal shelf location layout is obtained based on the evaluation results.

[0051] S60. Execute the optimal shelf location layout in the physical domain cold chain automated warehouse.

[0052] In specific application scenarios, historical order data is extracted from the cold chain automated warehouse management system or other relevant platforms. Each historical order record includes the order time, product category, quantity, and storage method. For example, the order time can include the specific year, month, day, and time period; the product category can cover various cold chain products, such as cola, ice cream, iced tea, dairy products, beef, and steak; the quantity represents the number of products in each order; and the storage method includes refrigerated, frozen, or ambient temperature storage. By extracting historical order data, real-time operational information of the cold chain automated warehouse can be obtained, and a background dataset of product storage can be constructed, providing a foundation for subsequent analysis.

[0053] By analyzing order content in historical order datasets, association rule analysis algorithms (such as Apriori and FP-growth) are used to identify frequently purchased product pairs and extract association features between each pair. Association features can include purchase frequency, overlap in purchase time, and similarity between product categories. For example, if a customer frequently buys cola along with refrigerated meat over a certain period, a strong association can be considered between these two product categories. This information is of significant reference value for the shelving layout of cold chain automated warehouses. By establishing a set of association features between products, it is possible to effectively identify which products should be stored close together, thereby reducing the time required for goods to be stored and retrieved.

[0054] The spatial adjacency coefficient of item pairs within a set of associated features is used to assess their relative positions in the warehouse. The spatial adjacency coefficient quantifies the priority of items to be stored in adjacent shelf locations. For example, if two items are frequently purchased together, their spatial adjacency coefficient is high, meaning they should be placed in adjacent shelf locations. Calculating the spatial adjacency coefficient allows for the allocation of appropriate relative positions to items in the cold chain automated warehouse, optimizing shelf layout. Placing items with high adjacency coefficients close together helps improve inventory retrieval efficiency and warehouse management.

[0055] A digital twin model of the cold chain automated warehouse is constructed. This model is a three-dimensional virtual warehouse model created based on information such as the layout, spatial constraints, and storage conditions of the actual warehouse. Subsequently, using the calculated spatial adjacency coefficient, the shelf positions are planned based on the spatial characteristics of the warehouse (such as the number, location, and size of the shelves). The digital twin model will automatically plan the shelf layout according to the spatial adjacency coefficient, generating multiple shelf position adjustment schemes.

[0056] The rationality of multiple shelf location adjustment schemes is evaluated using path planning algorithms (such as A* algorithm, Dijkstra algorithm, etc.). The core of the path planning evaluation is to consider factors such as picking distance, temperature control stability and AGV energy consumption, optimize the path of goods storage and retrieval, and thus obtain the optimal shelf layout scheme.

[0057] Based on the above optimization results, the optimal shelf location layout is applied to the physical domain cold chain automated warehouse. Warehouse staff adjust the shelves and store goods according to the planned optimal shelf location layout, thereby realizing the execution of the optimization scheme in the digital twin model in the actual warehouse.

[0058] The cold chain shelving location optimization method and system based on digital twins described in this invention provides optimized solutions for shelving location layout based on actual needs through purchase correlation analysis and spatial adjacency coefficient calculation of historical order data. This significantly reduces order picking distance and time, and lowers operating costs. Furthermore, different shelving layout schemes are simulated and verified using digital twin technology models, ensuring the scientific validity and operability of the final solution while reducing implementation risks.

[0059] By exploring the correlation between goods purchases and quantifying spatial adjacency coefficients, the shelves are adjusted to store frequently co-occurring goods nearby, significantly reducing picking paths, decreasing the travel distance and operation time of stacker cranes or AGV equipment, and reducing the empty running rate of equipment and the risk of temperature fluctuations in cold chain environments.

[0060] By dynamically adjusting the shelf layout based on historical order data characteristics, the rigidity of static zoning strategies can be avoided, allowing for flexible responses to seasonal order fluctuations or changes in product category demand, thereby improving the utilization rate of warehouse resources.

[0061] By using digital twin models to perform virtual simulation and path planning evaluation of multiple shelving adjustment schemes, the optimal layout scheme can be quickly selected before physical implementation, reducing the number of actual handling operations and trial and error costs, while also reducing cold energy loss caused by frequent inbound and outbound operations in cold chain warehousing.

[0062] Specifically, refer to Figure 2 As shown, a product purchase association analysis is performed on the historical order dataset to obtain a set of association features for purchase-related product pairs. This includes assigning time decay weights to each historical order data point to obtain a weighted order dataset; mining frequent itemsets of product combinations in the weighted order dataset layer by layer based on a prior algorithm to obtain a set of candidate association rules; performing multi-indicator screening on the candidate association rule set to obtain a preliminary set of screening rules; the screening indicators include support, confidence, and lift; performing density clustering analysis on the preliminary set of screening rules to obtain a set of strong association rules; and extracting the set of association features for purchase-related product pairs from the set of strong association rules.

[0063] In specific application scenarios, to ensure that more recent order data has a greater impact on product correlation analysis, a time decay model is used to weight historical order data. By assigning a decay weight to each order based on its time, it ensures that more recent order data has a stronger influence on the analysis results, while the influence of more distant order data gradually decreases. For example, an exponential decay function can be used to calculate the weights, giving higher weights to order data more recent to the current time. By assigning time decay weights, the interference of outdated data on the analysis results can be effectively avoided, improving the timeliness of data analysis. In this way, the optimization of cold chain warehouse shelf layout can more accurately reflect current market demand and the actual storage trends of goods.

[0064] The Apriori algorithm is used to mine the weighted order dataset layer by layer to identify frequently occurring item combinations. First, frequent itemsets are determined based on a set minimum support threshold; frequent itemsets are item combinations that appear more than the set threshold in the weighted order dataset, and these combinations have a high correlation.

[0065] Next, based on these frequent itemsets, a candidate set of association rules is generated. After obtaining the candidate set of association rules, it is filtered using multiple indicators such as support, confidence, and lift. By setting thresholds for support, confidence, and lift, a preliminary set of rules that meet the criteria is selected. An association rule represents the relationship between certain items (usually goods, behaviors, or events) under specific conditions. For example, in order data analysis, association rules can reveal the pattern that a customer who has purchased certain goods (cola) is also likely to purchase other goods (ice cream). The basic form of an association rule is: A→B, where A and B are sets of goods or itemsets. A→B means "If goods A are purchased, then it is very likely that goods B will be purchased." An association rule consists of two parts: the antecedent and the consequent. The antecedent is the left-hand side of the rule, representing the condition or event that has already occurred, such as the customer purchasing goods A. The consequent is the right-hand side of the rule, representing the possible event given the antecedent, such as the customer being very likely to purchase goods B. The goal of association rules is to discover valuable relationships hidden in large datasets, helping to identify potential product pairings or behavioral patterns, thereby achieving rational allocation of resources.

[0066] Association rules have three main metrics: support, confidence, and lift.

[0067] Support: This indicates the frequency with which a combination of goods included in an association rule appears in total orders; a rule with high support means that the combination is more common in actual orders. Support(A→B) = Frequency of orders containing both A and B / Total number of orders.

[0068] Confidence: Represents the probability that the consequent will occur if the antecedent occurs in an association rule; for example, confidence represents the proportion of orders that purchase product A but also purchase product B. Confidence(A→B) = Number of orders supporting both A and B / Number of orders supporting A.

[0069] Lift: Represents the strength of the association between the antecedent and consequent in an association rule, measuring their independence. A lift greater than 1 indicates a positive correlation, equal to 1 indicates no correlation, and less than 1 indicates a negative correlation. Lift(A→B) = Confidence(A→B) / Support(B).

[0070] Multi-indicator screening can extract high-quality association rules from numerous candidate rules, ensuring that subsequent optimization decisions are more accurate and effective. By comprehensively considering support, confidence, and lift, it can be ensured that the final selected association rules have strong business relevance and practical application value.

[0071] After initial association rule screening, density clustering analysis (such as the DBSCAN algorithm or K-means algorithm) is used to cluster the initially screened rules. For example, based on the DBSCAN algorithm, density clustering analysis is performed on rules that cluster core points with a confidence level greater than or equal to 40% and are classified as strong association rules; rules that cluster marginal points with a confidence level greater than or equal to 30% but less than 40% and are classified as weak association rules; and rules with a lift less than 1 (noisy points) are classified as negative association rules.

[0072] Density clustering analysis groups rules with similar purchasing patterns and association characteristics together based on their similarity. This method can uncover strongly correlated rule sets from a large pool of initial rules, reflecting the high degree of correlation between products in purchasing behavior. Density clustering analysis effectively identifies strong association patterns hidden in complex data, avoiding redundant and low-quality rules that may exist in traditional rule selection methods. The strongly correlated rules resulting from clustering analysis are more in line with actual needs and can better guide the optimization of cold chain automated storage and retrieval system (AS / RS) racking layout.

[0073] Finally, specific purchase-related product pairs are extracted from the set of strong association rules, forming a set of association features. Extracting these specific product pair association features from the set of strong association rules provides ample data support for precise optimization of shelf layout. The set of association features provides warehouse managers with clear references, making shelf location planning more scientific, avoiding unnecessary frequent relocation of goods, and improving overall operational efficiency.

[0074] Furthermore, the historical order dataset is subjected to a product purchase correlation analysis to obtain a set of correlation features for purchase-related product pairs. This also includes calculating the correlation strength of each product pair in the set of strong correlation rules; marking product pairs with correlation strength greater than or equal to a preset correlation strength as purchase-related product pairs; in the digital twin model, the correlation rules of the purchase-related product pairs are mapped to dynamic connection lines between virtual shelves; wherein the line width of the connection line is positively correlated with the correlation strength.

[0075] In specific application scenarios, the reliability and effectiveness of strong association rules can be measured by calculating the association strength of each strong association rule in the set of strong association rules. This can effectively filter out product pairs with high-strength associations. Lift can be chosen as the association strength indicator, where lift (A→B) = Confidence (A→B) / Support (B), where A and B are product pairs; Confidence (A→B) represents the probability of purchasing product B if product A has been purchased; and Support (B) represents the frequency of product B appearing in all orders.

[0076] After calculating the association strength of each item pair, item pairs with an association strength greater than or equal to the preset association strength are marked as purchase-related item pairs. The preset association strength is set based on business needs and actual warehouse operations. Its purpose is to ensure that only item pairs with strong purchase association can enter the subsequent shelf layout optimization stage. This avoids introducing weak association rules and ensures the accuracy and practicality of the optimization results.

[0077] In a digital twin model, the association rules for purchasing related product pairs are mapped to dynamic connection lines between virtual shelves. Each pair of related product rules establishes a virtual link with the warehouse shelf location via these connection lines. These lines represent the relationships between the products and are dynamically reflected in the digital twin warehouse model. By mapping association rules to connection lines between virtual shelves, the relationships and dependencies between products can be visually displayed in the digital twin model, facilitating warehouse managers' observation of product layout and optimization effects.

[0078] The width of each connecting line is proportional to the strength of the association. That is, the stronger the association, the wider the connecting line, indicating a stronger correlation between the goods and a closer relative distance between their shelf locations. In this way, the digital twin model can visually show which pairs of goods are strongly correlated and which pairs need to be placed on adjacent shelves. This visual representation helps warehouse managers better understand the storage requirements of the goods.

[0079] Reference Figure 3As shown, the spatial adjacency coefficient of the purchase-related goods pair is calculated based on the set of associated features, including the support, confidence, and lift of each associated feature in the set of associated features; the spatial adjacency coefficient is obtained based on the support, confidence, and lift.

[0080] The spatial adjacency coefficient is calculated and determined in the following manner:

[0081] ;

[0082] C and D represent the prerequisite and concluding goods for purchasing the related goods pair, respectively; Sc C→D S represents the spatial adjacency coefficient of a product to a CD; C→D Indicates the product's compatibility with CDs; C C→D Indicates the confidence level of the product regarding CD; L C→D Indicates the degree to which the goods enhance the CD; These represent the weighting coefficients for support, confidence, and lift, respectively.

[0083] In specific application scenarios, the spatial adjacency coefficient is used to quantify the priority of goods pairs stored in adjacent shelf locations, reflecting the strength of their association. The calculation of the spatial adjacency coefficient is based on three indicators: support, confidence, and lift. Assigning values ​​based on business experience; for example, for promotional items, prioritizing high confidence levels and tolerating low support levels. The values ​​are 0.3, 0.6, and 0.1 respectively; for everyday goods, priority is given to those with high support. The values ​​are set to 0.2, 0.2, and 0.6 respectively; for high-value goods, high relevance is prioritized, while low-frequency rules are weakened. Take values ​​of 0.5, 0.3, and 0.2 respectively.

[0084] In addition, refer to Figure 3 As shown, calculating the spatial adjacency coefficient of the purchased related goods pair further includes determining whether the purchased related goods pair has bidirectional association characteristics. If so, the bidirectional spatial adjacency coefficients of the purchased related goods pair are merged to obtain a merged spatial adjacency coefficient; the merged spatial adjacency coefficient is used as the actual spatial adjacency coefficient; wherein, the merged spatial adjacency coefficient is obtained based on the following method:

[0085] ;

[0086] Among them, W XY ' represents the spatial adjacency coefficient of the purchase of related goods pair XY; X and Y represent the goods purchased in the related goods pair; W XY W represents the spatial adjacency coefficient of X as the premise product and Y as the conclusion product. YXThis represents the spatial adjacency coefficient of Y as the premise item and X as the conclusion item; This represents the attenuation coefficient.

[0087] In specific application scenarios, when purchasing related product pairs, if the purchasing behavior of product X and product Y exhibits a bidirectional association, it indicates a strong mutual influence between the two products. For example, if a customer typically purchases product X and also purchases product Y, and conversely, purchases of product Y are also likely to result in the purchase of product X, such a product pair is labeled as "bidirectionally associated."

[0088] When there are bidirectional association rules (X→Y and Y→X) for purchasing related goods, the weight of the primary direction and the attenuated weight of the secondary direction are summed to avoid directly adding the weights, which would lead to inflated weights. A weight exceeding 1.0 would result in a misjudgment as "forced adjacency". Through the separation of primary and secondary weights and a dynamic attenuation mechanism (for example, the attenuation coefficient can be dynamically adjusted based on different scenarios, with a default value of 0.2 to avoid repeated accumulation), the dominance of the strong association direction is preserved while partially absorbing the additional influence of the reverse rule. This balances computational efficiency and business flexibility while retaining the directional semantics of the association rule.

[0089] Furthermore, referring to Figure 4 As shown, in the digital twin model, the shelf location is planned according to the spatial adjacency coefficient, including determining whether the storage methods of the purchased related goods pair are the same: if the storage methods are the same, the shelves of the purchased related goods pair are adjusted to storage units in the same logical partition according to the spatial adjacency coefficient; if the storage methods are different, the shelves of the purchased related goods pair are adjusted to storage units in different logical partitions according to the spatial adjacency coefficient.

[0090] In practical applications, the warehouse storage units are first logically partitioned. These partitions are based on the storage method of the goods. Storage units within the same logical partition share the same storage method (e.g., refrigerated, frozen, ambient temperature). In the digital twin model, this represents a set of shelves sharing the same set of storage constraints. Storage units in different logical partitions have different storage methods, and in the digital twin model, these represent a set of shelves with mutually exclusive storage constraints. Goods with different storage methods should be placed in separate areas to avoid improper storage due to temperature differences. For example, frozen foods and ambient temperature foods need to be placed in the frozen and ambient temperature zones respectively, even if they are closely related; the shelf positions need to be adjusted according to their storage methods.

[0091] Furthermore, referring to Figure 4As shown, adjusting the shelves of the purchased related product pairs to storage units in the same logical partition according to the spatial adjacency coefficient includes classifying the purchased related product pairs into strongly related, moderately related, and weakly related product pairs according to the spatial adjacency coefficient; in the digital twin model, adjusting the shelves of the strongly related product pairs to adjacent storage units; adjusting the shelves of the moderately related product pairs to non-adjacent storage units; and filling vacant spaces with shelves for the weakly related product pairs; wherein, the adjacent storage units are defined as nodes directly connected in the virtual shelf topology; and the non-adjacent storage units are defined as nodes connected through at least one intermediate node in the virtual shelf topology.

[0092] In specific application scenarios, the purchased related product pairs are classified into strongly related, moderately related, and weakly related product pairs based on the spatial adjacency coefficient. For example, pairs with a spatial adjacency coefficient greater than or equal to 0.8 are classified as strongly related product pairs; pairs with a spatial adjacency coefficient less than 0.8 but greater than or equal to 0.5 are classified as moderately related product pairs; and pairs with a spatial adjacency coefficient less than 0.5 are classified as weakly related product pairs. In the digital twin model, the warehouse shelf layout can be viewed as a virtual topology, where nodes represent storage units and edges represent connections between storage units. By setting the topological relationships of shelf positions, shelf positions can be adjusted according to different product associations.

[0093] Strongly related product pairs are moved to adjacent storage units to ensure they are close to each other in the physical warehouse and reduce picking paths; moderately related product pairs are moved to non-adjacent storage units to ensure they are in the same logical partition for easy and quick access; weakly related product pairs are adjusted by filling empty storage units to optimize space utilization.

[0094] Specifically, in the digital twin model, referring to Figure 5 As shown, the rationality of the multiple shelf location adjustment schemes is evaluated based on path planning to obtain the optimal shelf location layout. This includes loading multiple shelf location adjustment schemes into the digital twin model; for each shelf location adjustment scheme, AGVs are simulated in the digital twin model to perform picking of multiple order data, and the picking evaluation results of the multiple shelf location adjustment methods are obtained; the evaluation indicators of the picking evaluation results include picking distance, temperature control stability, and AGV energy consumption; the shelf location adjustment scheme with the best picking evaluation results is extracted as the optimal shelf location layout.

[0095] In specific application scenarios, the digital twin model first loads multiple shelf location adjustment schemes. These schemes originate from the aforementioned spatial adjacency coefficient calculation and correlation analysis steps, and are different shelf location layout schemes after spatial allocation and logical partitioning planning. Each scheme may include different shelf arrangements, product locations, and partition divisions. In the digital twin model, different historical order data or simulated order data are used to execute picking tasks. Each order includes multiple items. AGVs pick items according to the optimal path plan. The AGVs execute picking based on the preset shelf layout and picking path planning, recording key indicators for each picking process, including picking distance, temperature control stability, and AGV energy consumption. The picking distance is the total distance the AGV travels during the picking process for each order. A shorter picking distance indicates optimized shelf layout, reducing AGV travel time and material handling time. Consideration is given to whether the shelf layout affects the temperature control requirements of cold chain products. A reasonable shelf layout can reduce mutual interference between different temperature control zones and maintain temperature control stability. The total energy consumed by AGVs when performing picking tasks is measured. Energy consumption is affected by AGV path length, path complexity, and shelf location. Optimizing the path reduces AGV energy consumption and helps improve overall warehouse operational efficiency. For each shelf location adjustment plan, various evaluation indicators are calculated and comprehensively considered.

[0096] Furthermore, the AGV performs picking according to the effective path, referring to... Figure 6 As shown, determining the effective path includes: generating AGV picking paths based on a genetic algorithm in the digital twin model to obtain a candidate path set; introducing a path planning cost function and calculating the average path cost and standard deviation of the candidate paths; filtering effective paths from the candidate path set based on the average path cost and standard deviation to obtain a filtered path set; sorting the path costs of the filtered path set in descending order and extracting the first path as the effective path.

[0097] In specific application scenarios, within the digital twin model, a Genetic Algorithm (GA) is used to generate AGV picking routes. A GA is an optimization algorithm that simulates natural selection and genetic mechanisms, effectively exploring the solution space and finding near-optimal paths. First, multiple initial paths are generated as the initial population for the GA. Each path represents a possible picking route for the AGV in the warehouse. After several generations, a set of candidate paths is generated, representing multiple possible routes for the AGV from its starting position to its target position. To evaluate the quality of each path, a path planning cost function is introduced, which comprehensively considers three indicators: path length, number of turns, and temperature control stability (whether the path effectively avoids temperature-unstable areas and ensures temperature stability in the cold chain environment). The average and standard deviation of the path costs for all candidate paths are calculated. The standard deviation measures the dispersion of the path costs; a larger standard deviation indicates greater path variation. Effective paths are selected from the candidate path set based on the average path cost and standard deviation. The selection criteria include average path cost and standard deviation constraints. Firstly, paths with average path costs within a certain range and small standard deviations are considered to ensure the stability of path planning. Finally, a set of paths meeting the path cost requirements and having small standard deviations are selected as candidate effective paths. The selected effective paths are then sorted in descending order of path cost. From the sorted set of effective paths, the path with the lowest cost is extracted as the selected effective path, which is the optimal path for the AGV to perform picking tasks.

[0098] AGV path optimization based on genetic algorithms, combined with a multi-objective cost function, can provide the optimal picking route scheme in cold chain warehousing systems, improve warehouse operation efficiency, reduce energy consumption, ensure temperature control stability, and enhance the system's adaptability and flexibility.

[0099] Example 2: Based on the same inventive concept, this embodiment of the invention provides a cold chain shelf location optimization system based on digital twins, referring to... Figure 7 As shown, the system includes,

[0100] The data extraction module extracts historical order data and obtains a historical order dataset, which includes order time, product category, quantity, and storage method.

[0101] The correlation analysis module performs product purchase correlation analysis on the historical order dataset to obtain a set of correlation features for purchase-related product pairs;

[0102] The adjacency coefficient calculation module calculates the spatial adjacency coefficient of the purchased related goods pair based on the set of associated features;

[0103] The digital twin modeling module is used to build digital twin models of cold chain automated warehouses;

[0104] The intelligent planning module interacts with the adjacency coefficient calculation module and the digital twin modeling module, and is configured to generate multiple shelf location adjustment schemes in the digital twin model based on the spatial adjacency coefficient.

[0105] The scheme evaluation module is configured to evaluate the multiple shelf location adjustment schemes based on path planning in the digital twin model, and determine the optimal shelf location layout based on multi-objective optimization.

[0106] The drive execution module is configured to convert the optimal shelf location layout into warehouse equipment control commands, and drive the cold chain automated storage and retrieval system shelves to adjust according to the optimal shelf location layout.

[0107] The cold chain shelf location optimization system based on digital twins described in this embodiment of the invention is used to execute the cold chain shelf location optimization method of Embodiment 1, and has the same technical effect, which will not be repeated here.

[0108] In summary, the cold chain shelving location optimization method and system based on digital twins described in this invention provides optimized solutions for shelving location layout based on actual needs through purchase correlation analysis and spatial adjacency coefficient calculation of historical order data. This significantly reduces order picking distance and time, and lowers operating costs. Furthermore, simulation verification of different shelving layout schemes using digital twin models reduces implementation risks while ensuring the scientific validity and operability of the final solution.

[0109] By exploring the correlation between goods purchases and quantifying spatial adjacency coefficients, the shelves are adjusted to store frequently occurring goods nearby, significantly reducing picking paths, reducing the travel distance and operation time of stacker cranes or AGV equipment, and reducing the empty running rate of equipment and the risk of temperature fluctuations in the cold chain environment.

[0110] By dynamically adjusting the shelf layout based on historical order data characteristics, the rigidity of static zoning strategies can be avoided, allowing for flexible responses to seasonal order fluctuations or changes in product category demand, thereby improving the utilization rate of warehouse resources.

[0111] By using digital twin models to perform virtual simulation and path planning evaluation of multiple shelving adjustment schemes, the optimal layout scheme can be quickly selected before physical implementation, reducing the number of actual handling operations and trial and error costs, while also reducing cold energy loss caused by frequent inbound and outbound operations in cold chain warehousing.

[0112] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0113] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0114] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0115] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0116] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A method for optimizing the location of cold chain shelving based on digital twins, characterized in that: include, Extract historical order data to obtain a historical order dataset; the historical order data includes order time, product category, quantity, and storage method; Perform a product purchase correlation analysis on the historical order dataset to obtain a set of correlation features for purchase-related product pairs; Calculate the spatial adjacency coefficient of the purchased related goods pair based on the set of associated features; A digital twin model of a cold chain automated warehouse is established, and the shelf positions are planned in the digital twin model according to the spatial adjacency coefficient to obtain multiple shelf position adjustment schemes; In the digital twin model, multiple shelf location adjustment schemes are evaluated based on path planning, and the optimal shelf location layout is obtained based on the evaluation results; The optimal shelf location layout is executed in the physical domain cold chain automated warehouse. The calculation of the spatial adjacency coefficient of the purchase-related product pair based on the set of associated features includes: each associated feature in the set of associated features includes the support, confidence, and lift of the purchase-related product pair; and the spatial adjacency coefficient is calculated based on the support, confidence, and lift. In the digital twin model, the shelf location is planned according to the spatial adjacency coefficient, including: determining whether the storage methods of the purchased related goods pair are the same; if the storage methods are the same, the shelves of the purchased related goods pair are adjusted to storage units in the same logical partition according to the spatial adjacency coefficient; if the storage methods are different, the shelves of the purchased related goods pair are adjusted to storage units in different logical partitions according to the spatial adjacency coefficient. Adjusting the shelves of the purchased related goods pairs to storage units in the same logical partition according to the spatial adjacency coefficient includes: classifying the purchased related goods pairs into strongly related, moderately related, and weakly related goods pairs according to the spatial adjacency coefficient; in the digital twin model, adjusting the shelves of the strongly related goods pairs to adjacent storage units; adjusting the shelves of the moderately related goods pairs to non-adjacent storage units; and filling vacant spaces with shelves for the weakly related goods pairs; wherein, the adjacent storage units are defined as nodes directly connected in the virtual shelf topology; and the non-adjacent storage units are defined as nodes connected through at least one intermediate node in the virtual shelf topology.

2. The method for optimizing the location of cold chain shelves based on digital twins according to claim 1, characterized in that: Perform a product purchase correlation analysis on the historical order dataset to obtain a set of correlation features for purchase-related product pairs, including: Each historical order data entry is assigned a time decay weight to obtain a weighted order dataset; Based on a priori algorithm, frequent itemsets of product combinations in the weighted order dataset are mined layer by layer to obtain a set of candidate association rules; The candidate association rule set is filtered using multiple indicators to obtain a preliminary filtered rule set; the filtering indicators include support, confidence, and lift. Density clustering analysis is performed on the preliminary set of screening rules to obtain a set of strongly associated rules; Extract the set of association features of the purchase-related product pairs from the set of strong association rules.

3. The method for optimizing the location of cold chain shelves based on digital twins according to claim 2, characterized in that: Perform product purchase correlation analysis on the historical order dataset to obtain a set of correlation features for purchase-related product pairs, and also include: Calculate the association strength of each item pair in the set of strong association rules; Mark the product pairs with an association strength greater than or equal to the preset association strength as purchase related product pairs; In the digital twin model, the association rules of the purchased related product pairs are mapped as dynamic connection lines between virtual shelves; wherein, the line width of the connection line is positively correlated with the association strength.

4. The method for optimizing the location of cold chain shelves based on digital twins according to claim 1, characterized in that: Calculating the spatial adjacency coefficient of the purchased related goods pair also includes, If the purchase of related goods pairs has bidirectional association characteristics, then the bidirectional spatial adjacency coefficients of the purchase of related goods pairs are merged to obtain the merged spatial adjacency coefficients. The merged spatial adjacency coefficient is used as the actual spatial adjacency coefficient; The adjacency coefficient of the merged space is obtained based on the following method: ； Among them, W XY ' represents the spatial adjacency coefficient of the purchase of related goods pair XY; X and Y represent the goods purchased in the related goods pair; W XY W represents the spatial adjacency coefficient of X as the premise product and Y as the conclusion product. YX This represents the spatial adjacency coefficient of Y as the premise item and X as the conclusion item; This represents the attenuation coefficient.

5. The method for optimizing the location of cold chain shelves based on digital twins according to claim 1, characterized in that: In the digital twin model, the rationality of the multiple shelf location adjustment schemes is evaluated based on path planning to obtain the optimal shelf location layout, including: Multiple shelf location adjustment schemes are loaded into the digital twin model; For each of the aforementioned shelf location adjustment schemes, the AGV is simulated to pick multiple order data in the digital twin model to obtain the picking evaluation results of the multiple shelf location adjustment schemes; the evaluation indicators of the picking evaluation results include picking distance, temperature control stability, and AGV energy consumption. The optimal shelf location layout is the shelf location adjustment scheme obtained from the picking evaluation results.

6. The method for optimizing the location of cold chain shelves based on digital twins according to claim 5, characterized in that: The AGV performs picking based on a valid path, which includes the following: In the digital twin model, AGV picking routes are generated based on a genetic algorithm to obtain a set of candidate routes; A path planning cost function is introduced, and the average path cost and standard deviation of the candidate paths are calculated. Valid paths are selected from the candidate path set based on the average path cost and standard deviation to obtain a selected path set; Arrange the path costs of the filtered path set in descending order, and extract the first path as the valid path.

7. A cold chain shelving location optimization system based on digital twins, characterized in that: include, The data extraction module extracts historical order data and obtains a historical order dataset, which includes order time, product category, quantity, and storage method. The correlation analysis module performs product purchase correlation analysis on the historical order dataset to obtain a set of correlation features for purchase-related product pairs; The adjacency coefficient calculation module calculates the spatial adjacency coefficient of the purchased related goods pair based on the set of associated features; The digital twin modeling module is used to build digital twin models of cold chain automated warehouses; The intelligent planning module interacts with the adjacency coefficient calculation module and the digital twin modeling module, and is configured to generate multiple shelf location adjustment schemes in the digital twin model based on the spatial adjacency coefficient. The scheme evaluation module is configured to evaluate the multiple shelf location adjustment schemes based on path planning in the digital twin model, and determine the optimal shelf location layout based on multi-objective optimization. The drive execution module is configured to convert the optimal shelf location layout into warehouse equipment control commands, and drive the cold chain automated storage and retrieval system shelves to adjust according to the optimal shelf location layout. The calculation of the spatial adjacency coefficient of the purchase-related product pair based on the set of associated features includes: each associated feature in the set of associated features includes the support, confidence, and lift of the purchase-related product pair; and the spatial adjacency coefficient is calculated based on the support, confidence, and lift. In the digital twin model, the shelf location is planned according to the spatial adjacency coefficient, including: determining whether the storage methods of the purchased related goods pair are the same; if the storage methods are the same, the shelves of the purchased related goods pair are adjusted to storage units in the same logical partition according to the spatial adjacency coefficient; if the storage methods are different, the shelves of the purchased related goods pair are adjusted to storage units in different logical partitions according to the spatial adjacency coefficient. Adjusting the shelves of the purchased related goods pairs to storage units in the same logical partition according to the spatial adjacency coefficient includes: classifying the purchased related goods pairs into strongly related, moderately related, and weakly related goods pairs according to the spatial adjacency coefficient; in the digital twin model, adjusting the shelves of the strongly related goods pairs to adjacent storage units; adjusting the shelves of the moderately related goods pairs to non-adjacent storage units; and filling vacant spaces with shelves for the weakly related goods pairs; wherein, the adjacent storage units are defined as nodes directly connected in the virtual shelf topology; and the non-adjacent storage units are defined as nodes connected through at least one intermediate node in the virtual shelf topology.

Images