Supply chain sharing cold chain storage service management platform and method

Abstract

The invention relates to the technical field of warehouse management, and discloses a supply chain sharing cold chain warehouse service management platform and method. The method comprises the steps of performing data treatment on historical ex-warehouse data and historical storage data to obtain first complete data and second complete data; calculating a first score of each commodity according to the prediction time period and the first complete data, and calculating a second score of each storage area based on the second complete data; dividing the commodities into hot-sell commodities and non-hot-sell commodities based on the first score, and dividing the storage area into a rapid warehouse-out area and a common warehouse-out area based on the second score; matching the first score with the second score to determine a storage scheme; and obtaining stage ex-warehouse data of each commodity in the warehouse every preset time interval, generating a future ex-warehouse trend based on the stage ex-warehouse data, and updating the storage scheme based on the future ex-warehouse trend. According to the method, the commodity popularity, the ex-warehouse convenience and the commodity storage requirements are comprehensively considered, and finally the most appropriate storage position is selected for the commodity.

Description

Technical Field

[0001] The present application relates to the field of warehouse management technology, and in particular to a supply chain shared cold chain warehousing service management platform and method. Background Art

[0002] Modern warehouse management systems (WMS) are gradually being used in most companies. However, although these systems can improve the efficiency of item storage and retrieval and reduce inventory errors, they still face many challenges in large-scale and complex supply chain environments, especially in terms of resource scheduling, cargo distribution, and dynamic optimization of storage space. The system's response speed and management accuracy still need to be further improved.

[0003] To solve the above problems, the following methods are proposed in the prior art. For example, the Chinese patent document with publication number CN117132196A discloses a warehouse management method, device, equipment and storage medium. The method first classifies the heat of warehouse goods based on historical order data, and predicts the heat type of the goods according to the sales volume; then, the heat frequency value of each storage location is calculated according to the type of storage location and the initial value of the heat frequency; finally, according to the heat frequency value and inventory of the storage location, a suitable storage location is selected to transfer the goods to be transferred to the target storage location. For another example, the Chinese patent document with publication number CN114862294A discloses a recommendation management method for small and medium-sized warehouses based on big data analysis. The method improves the circulation rate of goods by optimizing the warehousing process and the outbound process, and by comparing historical data and product sales, high-sales goods are preferentially allocated to the storage location close to the distribution area, while low-sales goods are allocated to the farther storage location, thereby improving the efficiency of warehouse management.

[0004] However, in cold chain warehouse management, since most of the goods involved are sensitive to environmental conditions such as temperature and humidity, in addition to data such as goods and sales, environmental data such as temperature and humidity also need to be included in the recommendation system for comprehensive consideration to avoid goods being placed in locations with large temperature fluctuations, which may cause damage to the goods. Summary of the invention

[0005] The present application provides a supply chain shared cold chain warehousing service management platform and method to solve the problems existing in the above-mentioned background technology.

[0006] In order to achieve the above-mentioned invention object, the present invention proposes a supply chain shared cold chain warehousing service management method, comprising:

[0007] Collect historical outbound data of commodities and historical storage data of each storage area of ​​the warehouse, and perform data management on the historical outbound data and the historical storage data to obtain first complete data and second complete data;

[0008] Calculate a first score for each commodity based on the forecast time period and the first complete data, and calculate a second score for each storage area based on the second complete data;

[0009] Based on the first score, the commodities are divided into hot-selling commodities and non-hot-selling commodities, and based on the second score, the storage area is divided into a fast delivery area and a normal delivery area;

[0010] Obtaining storage parameters of each storage area in the fast delivery area and storage requirements of the hot-selling products;

[0011] Under the constraints of the storage requirements and the storage parameters, the first score and the second score are matched, and a storage solution for each of the hot-selling products is determined based on the matching result;

[0012] The phase outbound data of each commodity in the warehouse is obtained at preset time intervals, the future outbound trend is generated based on the phase outbound data, and the warehousing plan is updated based on the future outbound trend to continuously optimize the warehousing layout.

[0013] Furthermore, data governance of the historical outbound data includes the following steps:

[0014] Arrange each first sub-data in the historical outbound data into a first time series based on the timestamp of each first sub-data, divide the first time series into multiple sub-sequences according to the timestamp, set a corresponding adjustment weight for each sub-sequence, modify the first sub-data in the sub-sequence based on the adjustment weight to obtain second sub-data, and define the adjusted first time series as an initial sequence;

[0015] The mean value and standard deviation of the initial sequence are calculated, and a first numerical range and a second numerical range are generated based on the mean value and the standard deviation. If the second sub-data is within the first numerical range, the second sub-data is set as a normal value; if the second sub-data is outside the second numerical range, the second sub-data is set as an abnormal value. The abnormal value is corrected, and the corrected initial sequence is defined as a standard sequence.

[0016] Further, correcting the abnormal value comprises the following steps:

[0017] The upper limit and the lower limit of the second numerical range are located, and the abnormal value that is smaller than the lower limit is corrected to the lower limit, and the abnormal value that is larger than the upper limit is corrected to the upper limit.

[0018] Further, calculating the first score includes the following steps:

[0019] Generate a trend component sequence, a seasonal component sequence and a residual component sequence corresponding to the standard sequence for each commodity based on a seasonal decomposition algorithm, intercept the sequence corresponding to the forecast time period from the trend component sequence, the seasonal component sequence and the residual component sequence as the first sequence, the second sequence and the third sequence, and perform weighted summation on the first sequence, the second sequence and the third sequence to calculate an initial score;

[0020] The correlation scores between the commodities are calculated based on the standard sequence of each commodity, and the initial scores are corrected based on the correlation scores to obtain the first score.

[0021] Further, calculating the relevance score comprises the following steps:

[0022] The commodity to be calculated is defined as a target commodity, the target commodity is combined with the standard sequences of other commodities to obtain multiple sequence groups, two of the standard sequences in the sequence groups respectively define a first analysis sequence and a second analysis sequence, the first analysis sequence and the second analysis sequence are analyzed based on linear regression to obtain a correlation coefficient between the target commodity and other commodities, the correlation coefficient with the target commodity that is greater than a judgment threshold is used as an influence coefficient, the commodity corresponding to the influence coefficient is defined as an associated commodity, and the relevance score of the target commodity is calculated based on the influence coefficient and the initial score of the associated commodity.

[0023] Further, calculating the second score includes the following steps:

[0024] Based on the straight-line distance from the storage area to the picking area, a fuzzy theory algorithm is used to calculate the distance convenience score of the storage area. Based on the position of the storage area in the shelf, a hierarchical analysis method is used to calculate the accessibility score of the storage area. Based on the actual transportation path from the storage area to the picking area, a linear regression algorithm is used to calculate the path complexity score of the storage area. Based on the Bayesian model algorithm, the influence weights of all scores on the outbound efficiency are analyzed. Based on the influence weights, a weighted sum of all scores is taken to obtain a comprehensive score, and the comprehensive score is defined as the second score.

[0025] Furthermore, generating future outbound trends includes the following steps:

[0026] Acquire actual shipment data of commodities within the forecast time period, convert the actual shipment data into an actual shipment sequence, divide the actual shipment sequence into multiple interval sequences, define the last interval sequence as a first target sequence, define the interval sequence with the highest similarity to the first target sequence as a second target sequence, and define the interval sequence after the second target sequence as a third target sequence;

[0027] The actual outbound sequence is fitted based on the least square method to obtain a linear regression function, and the third target sequence is corrected based on the linear regression function to obtain the future outbound trend.

[0028] Further, matching the first score and the second score comprises the following steps:

[0029] The best-selling product with the highest first score is obtained, a storage area that meets the storage requirements is screened out in the fast delivery area, and the product with the highest second score is located in the screened storage area as a storage location for the best-selling product.

[0030] Furthermore, the division into the fast warehouse-out area and the ordinary warehouse-out area comprises the following steps:

[0031] A first threshold and a second threshold are set, and the commodities whose first score is greater than the first threshold are classified as the hot-selling commodities, and the storage area whose second score is greater than the second threshold is classified as the fast delivery area.

[0032] Furthermore, the present application provides a supply chain shared cold chain warehousing service management platform, which is used to implement the supply chain shared cold chain warehousing service management method described above, and the platform includes:

[0033] A collection module, used to collect historical outbound data of commodities and historical storage data of each storage area of ​​the warehouse, and perform data management on the historical outbound data and the historical storage data to obtain first complete data and second complete data;

[0034] a division module, configured to calculate a first score for each commodity according to the forecast time period and the first complete data, calculate a second score for each storage area based on the second complete data, divide the commodities into hot-selling commodities and non-hot-selling commodities based on the first score, and divide the storage area into a fast delivery area and a normal delivery area based on the second score;

[0035] A matching module, used to obtain storage parameters of each storage area in the fast delivery area and storage requirements of the hot-selling products, match the first score with the second score under the constraints of the storage requirements and the storage parameters, and determine a storage solution for each of the hot-selling products based on the matching results;

[0036] The update module is used to obtain the stage delivery data of each commodity in the warehouse at preset time intervals, generate future delivery trends based on the stage delivery data, and update the warehousing plan based on the future delivery trends to continuously optimize the warehousing layout.

[0037] The present invention ensures data accuracy through comprehensive management of historical outbound data and historical storage data, providing a reliable foundation for subsequent warehouse optimization. Hot-selling goods and fast outbound areas are quickly divided through the first and second scores, and only hot-selling goods and fast outbound areas are matched, which reduces the amount of data calculation and avoids large-scale adjustments to the current warehouse structure. By matching the storage requirements of goods with the storage parameters of the storage area, the most suitable storage area is selected for each product to avoid damage to the goods. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0039] Figure 1 A schematic diagram of a supply chain shared cold chain warehousing service management method for this application;

[0040] Figure 2 A schematic diagram predicting future outbound trends for this application;

[0041] Figure 3 A schematic diagram of a supply chain shared cold chain warehousing service management platform for this application. DETAILED DESCRIPTION

[0042] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0043] It is understood that the terms "first", "second", etc. used in this application may be used herein to describe various elements, but unless otherwise specified, these elements are not limited by these terms. These terms are only used to distinguish a first element from another element. For example, without departing from the scope of this application, a first xx script may be referred to as a second xx script, and similarly, a second xx script may be referred to as a first xx script.

[0044] like Figure 1 As shown, a supply chain shared cold chain warehousing service management platform includes:

[0045] S1: Collect the historical outbound data of the goods and the historical storage data of each storage area of ​​the warehouse, and perform data governance on the historical outbound data and historical storage data to obtain the first complete data and the second complete data.

[0046] Specifically, historical outbound data and historical storage data can be exported from the warehouse management system (WMS). Historical outbound data includes information such as the outbound time, quantity, customer information, outbound destination, order type, and delivery date of the goods. Historical storage data includes inventory status, time when the goods are put in, time when the goods are taken out, shelf number, physical location, etc. Data governance includes operations such as deduplication, missing value processing, timestamp normalization, and outlier detection. Data accuracy is ensured through data governance. Define the governed historical outbound data as the first complete data, and the historical storage data as the second complete data.

[0047] S2: Calculate a first score for each commodity according to the predicted time period and the first complete data, and calculate a second score for each storage area based on the second complete data.

[0048] Before calculating the score, this embodiment first sets the current time point and the future period to be predicted, that is, the preset time period. Since some commodities are not hot-selling throughout the year but are seasonally hot-selling, it is necessary to determine the preset time period. For example, if the warehouse layout is to be optimized in the next month, the popularity of the commodities in the next month is calculated based on historical records.

[0049] During the calculation process, if the outbound situation of a commodity is not affected by other commodities, the first score is calculated in combination with its own outbound frequency in the past period. If the outbound situation of a commodity is affected by other commodities, for example, commodity A is often taken out with commodity B when it is outbound, the first score of commodity A is calculated based on the outbound frequency of commodities A and B. For the second score, this embodiment calculates the second score in combination with the distance between the storage area and the picking area, the convenience of picking up goods, the convenience of transportation, etc. The specific method of calculating the first score and the second score will be introduced later. The higher the first score, the higher the outbound popularity of the commodity, and the higher the second score, the higher the outbound speed and convenience of the storage area.

[0050] S3: The products are divided into hot-selling products and non-hot-selling products based on the first score, and the storage area is divided into a fast delivery area and a normal delivery area based on the second score.

[0051] Specifically in this embodiment, the division into the fast warehouse-out area and the ordinary warehouse-out area includes the following steps:

[0052] A first threshold and a second threshold are set, and commodities with a first score greater than the first threshold are classified as hot-selling commodities, and a storage area with a second score greater than the second threshold is classified as a fast-export area.

[0053] By setting the first threshold and the second threshold, when the first score is greater than the first threshold, the product is classified as a hot-selling product, otherwise it is classified as a non-hot-selling product. When the second score is greater than the second threshold, the storage area is classified as a fast delivery area, otherwise it is classified as a normal delivery area. The closer the storage area is to the picking area, the more convenient it is, and the more likely it is to be classified as a fast delivery area. For example, the first threshold is 70 and the second threshold is 60.

[0054] S4: Obtain the storage parameters of each storage area in the fast delivery area and the storage requirements of hot-selling products.

[0055] S5: Under the constraints of storage requirements and storage parameters, the first score and the second score are matched, and a storage solution for each hot-selling product is determined based on the matching result.

[0056] Matching the first score and the second score includes the following steps:

[0057] The hot-selling product with the highest first score is obtained, a storage area that meets its storage requirements is selected in the fast delivery area, and the storage location of the hot-selling product with the highest second score is located in the selected storage area.

[0058] Storage parameters include the temperature changes in each storage area. The temperature fluctuation range and stability of the storage are determined by the temperature changes. Storage requirements include the storage temperature range of the goods. The closer to the center of the entire warehouse, the higher the corresponding ambient temperature. Therefore, it is necessary to match a reasonable storage location according to the product storage requirements.

[0059] Before generating a storage plan, first divide the goods into hot-selling goods and non-hot-selling goods, and divide the storage area into fast-delivery area and ordinary-delivery area. In the subsequent matching, only the hot-selling goods and the fast-delivery area are matched, which reduces the matching complexity, improves the matching speed, and also avoids large-scale layout adjustments of the warehouse. When matching hot-selling goods, first obtain the goods with the highest first score, and then select the appropriate storage area from the fast-delivery area according to the storage requirements of the hot-selling products, and locate the second highest-scoring storage area in the selected storage area as the storage location of the goods. Then match the subsequent lower first scores. If there are goods that are not matched to the storage area at the end, the goods will be stored in the ordinary-delivery area.

[0060] S6: Obtain the phased outbound data of each commodity in the warehouse at preset intervals, generate future outbound trends based on the phased outbound data, and update the warehousing plan based on the future outbound trends to continuously optimize the warehousing layout.

[0061] The preset time is, for example, 15 days. After the warehousing plan is generated, the product outbound data will continue to be obtained in the future. After 15 days, the future outbound trend of the product will be determined based on the outbound data within 15 days. If it is determined based on the future outbound trend that the product is no longer hot-selling, for example, it is predicted that the sales volume of the product will start to decline, the product will be transferred to the general outbound area, thereby continuously optimizing the warehousing layout and ensuring the efficient operation of the warehouse.

[0062] The present invention ensures data accuracy through comprehensive management of historical outbound data and historical storage data, providing a reliable foundation for subsequent warehouse optimization. Hot-selling goods and fast outbound areas are quickly divided through the first and second scores, and only hot-selling goods and fast outbound areas are matched, which reduces the amount of data calculation and avoids large-scale adjustments to the current warehouse structure. By matching the storage requirements of goods with the storage parameters of the storage area, the most suitable storage area is selected for each product to avoid damage to the goods.

[0063] It is particularly noteworthy that the present invention comprehensively considers the popularity of the goods, the convenience of outbound delivery and the storage requirements of the goods, and ultimately selects the most suitable storage location for the goods.

[0064] In this embodiment, data governance of historical outbound data includes the following steps:

[0065] Based on the timestamp of each first sub-data in the historical outbound data, it is sorted into a first time series, and the first time series is divided into multiple sub-sequences according to the timestamp, and a corresponding adjustment weight is set for each sub-sequence. Based on the adjustment weight, the first sub-data in the sub-sequence is corrected to obtain the second sub-data, and the adjusted first time series is defined as the initial sequence.

[0066] In this embodiment, the abnormal values ​​in the historical outbound data are located and corrected based on the following steps. The historical outbound data includes the first sub-data corresponding to different types of goods. The first sub-data has a timestamp, such as outbound time: 15:00 on January 1, type: frozen food A. Then, according to the order of the timestamps, the first sub-data corresponding to the same type of goods are sorted into a first time series, and then the first time series is divided into multiple sub-sequences. The division method can be manually specified, and the value of the adjustment weight is between 0 and 1. For example, the sequence within 3 months before the current time point is divided and used as sub-sequence 1, and the sequence from the first 3 months to the first 6 months is used as sub-sequence 2. The adjustment weight set for sub-sequence 1 is 0.9, and the adjustment weight set for sub-sequence 2 is 0.8. Each first sub-data in sub-sequence 1 is multiplied by 0.9 to obtain the second sub-data, and each first sub-data in sub-sequence 2 is multiplied by 0.8 to obtain the second sub-data. Finally, the adjusted first time series is defined as the initial sequence.

[0067] The further away the generation time of the historical outbound data is from the current time point, the lower its reference significance is. Therefore, through this step, the historical data that is far away from the current time point can be reduced to avoid affecting the calculation of the subsequent first score.

[0068] The mean value and standard deviation of the initial sequence are calculated, and a first numerical range and a second numerical range are generated based on the mean value and the standard deviation. If the second sub-data is within the first numerical range, the second sub-data is set as a normal value. If the second sub-data is outside the second numerical range, the second sub-data is set as an abnormal value. The abnormal value is corrected, and the corrected initial sequence is defined as a standard sequence.

[0069] In this embodiment, the lower limit of the first numerical range is μ-g 1 σ, the upper limit of the first numerical range is μ+g 1 σ, where μ is the mean value of the initial sequence, σ is the standard deviation of the initial sequence, and g 1 is the first correction coefficient. The value of the first correction coefficient is set within a range less than 1 according to actual needs. For example, for the initial sequence: 3, 20, 18, 21, 25, 18, 26, 23, 22, 40, the average value is 21.6 and the standard deviation is 8.64. If the first correction coefficient is 1, the first value range is 12.93 to 30.24, and the data within this range are normal data. For the second value range, the corresponding lower limit is μ-g 2 σ, the upper limit of the first numerical range is μ+g 2 σ, where g 2 is the second correction coefficient. The value of the second correction coefficient is set within a numerical range greater than 1 according to actual needs. Similarly, for the above initial sequence, the second correction coefficient is set to 2, then the second numerical range is 4.32 to 38.88, and the second sub-data outside the second numerical range is defined as an abnormal value.

[0070] Specifically, the present invention corrects the abnormal value including the following steps:

[0071] The upper limit and lower limit of the second numerical range are located, and abnormal values ​​less than the lower limit are corrected to the lower limit, and abnormal values ​​greater than the upper limit are corrected to the upper limit.

[0072] In the above example, the second sub-data with values ​​of 3 and 40 are outliers. Since 3 is less than the lower limit of the second numerical range, the value 3 is corrected to 4.32. Since 40 is greater than the upper limit of the second numerical range, the value 40 is corrected to 38.88. The above method can quickly locate the outliers in the initial sequence and correct them, thereby ensuring the accuracy of subsequent data analysis.

[0073] Calculating the first score includes the following steps:

[0074] Based on the seasonal decomposition algorithm, the trend component sequence, seasonal component sequence and residual component sequence are generated for the standard sequence corresponding to each commodity. The sequences corresponding to the forecast time period are intercepted from the trend component sequence, seasonal component sequence and residual component sequence as the first sequence, second sequence and third sequence. The initial score is calculated by weighted summing of the first sequence, the second sequence and the third sequence.

[0075] The correlation scores between the commodities are calculated based on the standard sequence of each commodity, and the initial scores are corrected based on the correlation scores to obtain a first score.

[0076] Specifically, here we use the additive model to perform seasonal decomposition. First, we apply a sliding window to the standard sequence to calculate the trend component, obtain the trend value of the data corresponding to each time point in the standard sequence, and the size of the sliding window is 3 months. The centralized moving average algorithm is used to calculate the trend component of the data corresponding to each time point, and then the seasonal component and residual component are calculated based on the trend component. The specific calculation method is the existing technology. The larger the trend component, the larger the product shipment volume without being affected by the season. The larger the seasonal component, the greater the impact of the season on the product shipment volume. The larger the residual component, the greater the volatility of the product shipment, that is, the greater the randomness.

[0077] Then, the trend sequence, seasonal sequence and residual sequence are constructed according to the trend component, seasonal component and residual component respectively. The data of the same period as the forecast time period are intercepted in the trend sequence, seasonal sequence and residual sequence. For example, if the forecast time period is March, the sequence data corresponding to March in the trend sequence, seasonal sequence and residual sequence are intercepted as the first sequence, the second sequence and the third sequence respectively. The sum of all the values ​​in the first sequence is taken as the first value. Similarly, the sum of the values ​​in the second sequence and the third sequence is taken as the second value and the third value. The weighted sum of the first value, the second value and the third value is obtained to obtain the initial score. In particular, the weight of the first value is 0.5, the weight of the second value is 0.2, and the weight of the third value is 0.2. After obtaining the initial score, the correlation score between the products is calculated, and the correlation score is added to the initial score to obtain the result as the first score.

[0078] In this embodiment, calculating the relevance score includes the following steps:

[0079] The commodity to be calculated is defined as the target commodity. The target commodity is combined with the standard sequences of other commodities to obtain multiple sequence groups. The two standard sequences in the sequence groups respectively define the first analysis sequence and the second analysis sequence. The first analysis sequence and the second analysis sequence are analyzed based on linear regression to obtain the correlation coefficient between the target commodity and other commodities. The correlation coefficient with the target commodity that is greater than the judgment threshold is taken as the influence coefficient. The commodity corresponding to the influence coefficient is defined as an associated commodity. The relevance score of the target commodity is calculated based on the influence coefficient and the initial score of the associated commodity.

[0080] Assuming that the relevance score of the target product is to be calculated, the standard sequence of the target product and the standard sequence of the remaining products are obtained, the target product is time-aligned with the standard sequence of each of the remaining products, and the correlation analysis is performed based on the aligned data to obtain the Pearson correlation coefficient R between the standard sequence of the target product and the standard sequence of the remaining products. The correlation coefficient R is between -1 and 1. The closer the correlation coefficient R is to 1, the positive correlation between the two time series is indicated, that is, the increase of one time series data will cause the increase of another time series data at the same time. After that, the initial score of each product is obtained by using the seasonal decomposition algorithm, the initial score of each product is multiplied by the influence coefficient to obtain the first value, the first values ​​of all related products are added and summed to obtain the second value, and the second value is used as the relevance score of the target product. The higher the relevance score of the target product, the easier it is for the target product to be affected by other products, and the result of the influence will lead to an increase in the product's outbound volume.

[0081] In this embodiment, calculating the second score includes the following steps:

[0082] Based on the straight-line distance from the storage area to the picking area, the fuzzy theory algorithm is used to calculate the distance convenience score of the storage area. Based on the location of the storage area in the shelf, the hierarchical analysis method is used to calculate the accessibility score of the storage area. Based on the actual transportation path from the storage area to the picking area, the linear regression algorithm is used to calculate the path complexity score of the storage area. Based on the Bayesian model algorithm, the influence weights of all scores on the outbound efficiency are analyzed. Based on the influence weights, the weighted sum of all scores is used to obtain the comprehensive score, and the comprehensive score is defined as the second score.

[0083] For the convenience score, three fuzzy sets are first defined, namely, the close distance set, the medium distance set, and the long distance set. Each distance set defines a corresponding membership function, which includes trigonometric functions, trapezoidal functions, and Gaussian functions. This embodiment uses trigonometric functions as the membership function of the close distance set. The trigonometric function includes three sub-functions, namely, u=1, (d≤d 1 ), u=(d 2 -d) / (d 2 -d 1 ), (d1 <d≤d 2 ), u=0, (d>d 2 ), where u is the degree of membership, d is the straight-line distance from the storage area to the picking area, d1 and d2 are preset critical distances, respectively. The value of d1 is 10 meters, and the value of d2 is 50 meters. For example, if the straight-line distance from the storage area to the picking area is 30, then its degree of membership with the short-distance set is 0.4. No further examples are given for the membership functions of the medium-distance set and the long-distance set, and those skilled in the art can set them according to actual conditions. Finally, the convenience score is calculated by taking the weighted sum of the degree of membership of each type. The three types have memberships of 0.4, 0.4, and 0, respectively, and the corresponding weights are 0.3, 0.3, and 0.4, respectively. The convenience score is 0.4*0.3+0.4*0.3+0*0.4=0.24.

[0084] When using the analytic hierarchy process to calculate the accessibility score of a storage area, first determine all factors that affect shelf accessibility, such as shelf location, whether specific equipment is required, whether it is easy to operate, whether there are obstacles, etc., and whether it is easy for operators to access. The above factors are scored using values ​​1-9, and then a judgment matrix is ​​constructed based on the values ​​of each factor. The construction method will not be repeated; then the judgment matrix is ​​normalized, and finally the eigenvalue and eigenvector of the normalized matrix are calculated by eigenvalue decomposition. The eigenvector contains the weight of each factor. Then, each storage area is scored according to the above criteria. For example, a storage area has a shelf location score of 9, a device compatibility score (whether specific equipment is required, whether it is easy to operate) of 6, a difficulty of access (whether there are obstacles, etc.) score of 8, and an operability score (whether it is easy for operators to access) of 9. The weights of the corresponding factors are 0.4, 0.3, 0.2, and 0.1, respectively, so the accessibility score of the storage area is 7.4.

[0085] For the path complexity score, first set the number of turns from the storage area to the picking area and the total length of the path as two influencing factors, and then construct a linear regression function to calculate the path complexity score. The linear regression function is, for example, S = N·ω 1 +L·ω 2 , where S is the path complexity score, N is the number of turns, L is the total path length, and ω1 and ω2 are the weights of the number of turns and the total path length, respectively.

[0086] Finally, the transportation time of goods from the shelves to the picking area in the storage area is obtained from the historical storage data. Then, based on these transportation times, the conditional probability of each score is calculated in combination with the Bayesian algorithm to obtain the impact weight on transportation efficiency. Finally, based on the impact weight, the distance convenience score, accessibility score, and path complexity score calculated above are weighted and summed, and the final comprehensive score is used as the first score of each storage area.

[0087] In this embodiment, generating the future delivery trend includes the following steps:

[0088] Obtain the actual outbound data of the goods within the preset time, convert the actual outbound data into the actual outbound time series, fit the actual outbound time series based on the least squares method, obtain the linear regression function, and generate the ideal time series and growth trend based on the linear regression function;

[0089] The actual outbound sequence is divided into multiple subsequences, and the last subsequence is defined as the first target sequence. The subsequence with the highest similarity to the first target sequence is defined as the second target sequence. The subsequence after the second target sequence is defined as the third target sequence. The third target sequence is modified based on the growth trend and defined as the future outbound trend.

[0090] In this embodiment, generating the future delivery trend includes the following steps:

[0091] The actual shipment data of the goods within the forecast time period is obtained, the actual shipment data is converted into an actual shipment sequence, the actual shipment sequence is divided into multiple interval sequences, the last interval sequence is defined as the first target sequence, the interval sequence with the highest similarity to the first target sequence is defined as the second target sequence, and the interval sequence after the second target sequence is defined as the third target sequence.

[0092] The actual outbound sequence is fitted based on the least squares method to obtain a linear regression function, and the third target sequence is corrected based on the linear regression function to obtain the future outbound trend.

[0093] For example, if the forecast period is 5 days, after adjusting the warehouse layout for 30 days, the actual outbound sequence is constructed based on the actual outbound data of the goods every day within 30 days, and then the actual outbound sequence is divided into multiple interval sequences, such as data within 1-5 days is divided into an interval sequence, and data within 5-10 days is divided into an interval sequence. Then, the least squares method is used to fit the actual outbound sequence to obtain a linear regression function in the form of y=kx+b, where y is the outbound quantity of goods in a day, k and b are the fitted coefficients and intercepts, and x is time. For example, the linear regression function is y=8.3x+105.

[0094] The last interval sequence is located as the first target sequence, that is, the interval sequence corresponding to 25-30 days. Figure 2 The prediction process is schematically explained. First, locate Figure 2 Q1 in , obtain the interval sequence that is most similar to this interval, that is, the second target sequence, such as Figure 2 In Q2, the Pearson correlation coefficient can be used as the similarity, and the interval sequence after the second target sequence is the third target sequence, such as Figure 2 If the k value in the linear regression function is positive, it means that the actual outbound sequence is increasing. Therefore, the third target sequence is appropriately increased to obtain the future outbound trend. Figure 2 Take the correction process as an example: substitute the last time point t of the actual outbound sequence into the linear regression function to obtain the third value. After moving the third target sequence to interval Q4, take the third value as the starting point and keep its fluctuation characteristics unchanged. The specific numerical calculation process will not be introduced again.

[0095] like Figure 3 As shown, the present application provides a supply chain shared cold chain warehousing service management platform, which is used to implement the above-mentioned supply chain shared cold chain warehousing service management method, and the platform includes:

[0096] The collection module is used to collect the historical outbound data of the goods and the historical storage data of each storage area of ​​the warehouse, and perform data management on the historical outbound data and the historical storage data to obtain the first complete data and the second complete data.

[0097] The division module is used to calculate a first score for each commodity according to the predicted time period and the first complete data, calculate a second score for each storage area based on the second complete data, divide the commodities into hot-selling commodities and non-hot-selling commodities based on the first score, and divide the storage area into a fast delivery area and a normal delivery area based on the second score.

[0098] The matching module is used to obtain the storage parameters of each storage area in the fast delivery area and the storage requirements of the hot-selling goods. Under the constraints of the storage requirements and storage parameters, the first score and the second score are matched, and the storage plan for each hot-selling product is determined based on the matching results.

[0099] The update module is used to obtain the stage delivery data of each commodity in the warehouse at preset intervals, generate future delivery trends based on the stage delivery data, and update the warehousing plan based on the future delivery trends to continuously optimize the warehousing layout.

[0100] It should be understood that the various technical features of the above-mentioned embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the various technical features in the above-mentioned embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0101] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A supply chain shared cold chain warehousing service management method, characterized in that: Collect historical outbound data of commodities and historical storage data of each storage area of ​​the warehouse, and perform data management on the historical outbound data and the historical storage data to obtain first complete data and second complete data; Calculate a first score for each commodity based on the forecast time period and the first complete data, and calculate a second score for each storage area based on the second complete data; Based on the first score, the commodities are divided into hot-selling commodities and non-hot-selling commodities, and based on the second score, the storage area is divided into a fast delivery area and a normal delivery area; Obtaining storage parameters of each storage area in the fast delivery area and storage requirements of the hot-selling products; Under the constraints of the storage requirements and the storage parameters, the first score and the second score are matched, and a storage solution for each of the hot-selling products is determined based on the matching result; The phase outbound data of each commodity in the warehouse is obtained at preset time intervals, the future outbound trend is generated based on the phase outbound data, and the warehousing plan is updated based on the future outbound trend to continuously optimize the warehousing layout.

2. The method according to claim 1, characterized in that Data governance of the historical outbound data includes the following steps: Arrange each first sub-data in the historical outbound data into a first time series based on the timestamp of each first sub-data, divide the first time series into multiple sub-sequences according to the timestamp, set a corresponding adjustment weight for each sub-sequence, modify the first sub-data in the sub-sequence based on the adjustment weight to obtain second sub-data, and define the adjusted first time series as an initial sequence; The mean value and standard deviation of the initial sequence are calculated, and a first numerical range and a second numerical range are generated based on the mean value and the standard deviation. If the second sub-data is within the first numerical range, the second sub-data is set as a normal value; if the second sub-data is outside the second numerical range, the second sub-data is set as an abnormal value. The abnormal value is corrected, and the corrected initial sequence is defined as a standard sequence.

3. The method according to claim 2, characterized in that Correcting the outlier includes the following steps: The upper limit and the lower limit of the second numerical range are located, and the abnormal value that is smaller than the lower limit is corrected to the lower limit, and the abnormal value that is larger than the upper limit is corrected to the upper limit.

4. The method according to claim 2, characterized in that: Calculating the first score includes the following steps: Generate a trend component sequence, a seasonal component sequence and a residual component sequence corresponding to the standard sequence for each commodity based on a seasonal decomposition algorithm, intercept the sequence corresponding to the forecast time period from the trend component sequence, the seasonal component sequence and the residual component sequence as the first sequence, the second sequence and the third sequence, and perform weighted summation on the first sequence, the second sequence and the third sequence to calculate an initial score; The correlation scores between the commodities are calculated based on the standard sequence of each commodity, and the initial scores are corrected based on the correlation scores to obtain the first score.

5. The method according to claim 4, characterized in that Calculating the relevance score comprises the following steps: The commodity to be calculated is defined as a target commodity, the target commodity is combined with the standard sequences of other commodities to obtain multiple sequence groups, two of the standard sequences in the sequence groups respectively define a first analysis sequence and a second analysis sequence, the first analysis sequence and the second analysis sequence are analyzed based on linear regression to obtain a correlation coefficient between the target commodity and other commodities, the correlation coefficient with the target commodity that is greater than a judgment threshold is used as an influence coefficient, the commodity corresponding to the influence coefficient is defined as an associated commodity, and the relevance score of the target commodity is calculated based on the influence coefficient and the initial score of the associated commodity.

6. The method according to claim 1, characterized in that Calculating the second score includes the following steps: Based on the straight-line distance from the storage area to the picking area, a fuzzy theory algorithm is used to calculate the distance convenience score of the storage area. Based on the position of the storage area in the shelf, a hierarchical analysis method is used to calculate the accessibility score of the storage area. Based on the actual transportation path from the storage area to the picking area, a linear regression algorithm is used to calculate the path complexity score of the storage area. Based on the Bayesian model algorithm, the influence weights of all scores on the outbound efficiency are analyzed. Based on the influence weights, a weighted sum of all scores is taken to obtain a comprehensive score, and the comprehensive score is defined as the second score.

7. The method according to claim 1, characterized in that Generating future outbound trends involves the following steps: Acquire actual shipment data of commodities within the forecast time period, convert the actual shipment data into an actual shipment sequence, divide the actual shipment sequence into multiple interval sequences, define the last interval sequence as a first target sequence, define the interval sequence with the highest similarity to the first target sequence as a second target sequence, and define the interval sequence after the second target sequence as a third target sequence; The actual outbound sequence is fitted based on the least square method to obtain a linear regression function, and the third target sequence is corrected based on the linear regression function to obtain the future outbound trend.

8. The method according to claim 1, characterized in that Matching the first score and the second score comprises the following steps: The best-selling product with the highest first score is obtained, a storage area that meets the storage requirements is screened out in the fast delivery area, and the product with the highest second score is located in the screened storage area as a storage location for the best-selling product.

9. The method according to claim 1, characterized in that: The division into the fast warehouse-out area and the ordinary warehouse-out area comprises the following steps: A first threshold and a second threshold are set, and the commodities whose first score is greater than the first threshold are classified as the hot-selling commodities, and the storage area whose second score is greater than the second threshold is classified as the fast delivery area.

10. A supply chain shared cold chain warehousing service management platform, used to implement the supply chain shared cold chain warehousing service management method as described in any one of claims 1 to 9, characterized in that: include: A collection module, used to collect historical outbound data of commodities and historical storage data of each storage area of ​​the warehouse, and perform data management on the historical outbound data and the historical storage data to obtain first complete data and second complete data; a division module, configured to calculate a first score for each commodity according to the forecast time period and the first complete data, calculate a second score for each storage area based on the second complete data, divide the commodities into hot-selling commodities and non-hot-selling commodities based on the first score, and divide the storage area into a fast delivery area and a normal delivery area based on the second score; A matching module, used to obtain storage parameters of each storage area in the fast delivery area and storage requirements of the hot-selling products, match the first score with the second score under the constraints of the storage requirements and the storage parameters, and determine a storage solution for each of the hot-selling products based on the matching results; The update module is used to obtain the stage delivery data of each commodity in the warehouse at preset time intervals, generate future delivery trends based on the stage delivery data, and update the warehousing plan based on the future delivery trends to continuously optimize the warehousing layout.

Images