Goods storage allocation matching method and device and warehousing system
The density clustering algorithm classifies and matches the cargo space, solving the problem of low entry efficiency in the traditional cargo entry matching method, and achieving a more efficient cargo entry process.
Patent Information
- Application Number
- CN202411485788.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-23
- Publication Date
- 2025-02-25
AI Technical Summary
The traditional method of cargo entry matching has the problem of low efficiency in cargo entry, and it is necessary to manually judge the type of goods and allocate the cargo space.
The density clustering algorithm is used to classify the incoming goods, and the local density of the cargo characteristic data is calculated through the trained clustering model and the clustering center is determined, so as to match the target cargo position.
This greatly improves the efficiency of cargo storage, achieves higher clustering accuracy and efficiency, and solves the problem of low storage efficiency in traditional methods.
Smart Images
Figure CN119515252A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of warehouse management, and in particular to a method and device for matching goods entry locations and a warehouse entry system. Background Art
[0002] Intelligent warehouse management systems are widely used in our production, especially the warehousing and outbound of goods, which are the two most important links of intelligent warehouse management systems. In order to facilitate management, similar goods are stored together. Therefore, how to quickly classify a large number of goods when goods are put into storage is the key to improving the efficiency of goods entering the warehouse. Most traditional goods entering the warehouse matching methods manually judge the type of goods and assign goods to them, which has the problem of low efficiency of goods entering the warehouse. Summary of the invention
[0003] The present invention provides a cargo entry cargo location matching method, device and warehouse entry system to solve the problem of low cargo entry efficiency in traditional cargo entry matching methods.
[0004] In a first aspect, the present invention provides a method for matching cargo storage locations, comprising:
[0005] Inputting the cargo feature data of each cargo to be stored into the trained clustering model, and obtaining the clustering result of each cargo to be stored through the clustering model;
[0006] Matching the target cargo location of each of the goods to be stored according to the clustering results of each of the goods to be stored;
[0007] The data processing steps of the clustering model include:
[0008] Calculating the Euclidean distance between the cargo characteristic data;
[0009] The local density of each of the cargo feature data is calculated according to the kernel function, and the kernel function is:
[0010]
[0011] Among them, ρ i represents the local density of the characteristic data of the i-th cargo, d ij represents the Euclidean distance between the i-th commodity feature data and the j-th commodity feature data, d c represents the cutoff distance, n represents the total number of cargo characteristic data, and χ(x) represents the step function;
[0012] Determining at least one cluster center data from a plurality of the cargo feature data according to the local density of each of the cargo feature data;
[0013] Each cargo feature data except the cluster center data is assigned to the cluster where the cluster center data closest to itself is located, and the clustering results of each cluster center data are used as the clustering results corresponding to each cargo to be stored.
[0014] In a second aspect, the present invention provides a device for matching cargo storage locations, comprising:
[0015] A cargo clustering module, used to input cargo feature data of each cargo to be stored into a trained clustering model, and obtain clustering results of each cargo to be stored through the clustering model;
[0016] A cargo location matching module, used for matching the target cargo location of each of the cargoes to be stored according to the clustering results of each of the cargoes to be stored;
[0017] The data processing steps of the clustering model include:
[0018] Calculating the Euclidean distance between the cargo characteristic data;
[0019] The local density of each of the cargo feature data is calculated according to the kernel function, and the kernel function is:
[0020]
[0021] Among them, ρ i represents the local density of the characteristic data of the i-th cargo, d ij represents the Euclidean distance between the i-th commodity feature data and the j-th commodity feature data, d c represents the cutoff distance, n represents the total number of cargo characteristic data, and χ(x) represents the step function;
[0022] Determining at least one cluster center data from a plurality of the cargo feature data according to the local density of each of the cargo feature data;
[0023] Each cargo feature data except the cluster center data is assigned to the cluster where the cluster center data closest to itself is located, and the clustering results of each cluster center data are used as the clustering results corresponding to each cargo to be stored.
[0024] In a third aspect, the present invention provides a warehouse warehousing system, comprising a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to run the computer program to execute the goods warehousing location matching method described in the first aspect.
[0025] Compared with the related art, the present invention realizes the classification of goods to be stored by density clustering algorithm, which greatly improves the efficiency of goods storage compared with the traditional goods storage method that uses manual classification of goods. At the same time, the density clustering algorithm adopted by the present invention is not a traditional density clustering algorithm, but an improvement of the traditional density clustering algorithm, which has higher clustering accuracy and clustering efficiency. It solves the problem of low efficiency of goods storage in the traditional goods storage matching method.
[0026] Details of one or more embodiments of the present application are set forth in the following drawings and description to make other features, objects, and advantages of the present application more readily apparent. BRIEF DESCRIPTION OF THE DRAWINGS
[0027] Figure 1 is a flow chart of a method for matching cargo storage locations provided in an embodiment of the present invention;
[0028] Figure 2 It is a flow chart of a method for matching incoming goods to storage locations provided in some embodiments of the present invention. DETAILED DESCRIPTION
[0029] In order to more clearly understand the purpose, technical solutions and advantages of the present application, the present application is described and illustrated below in conjunction with the accompanying drawings and embodiments.
[0030] Unless otherwise defined, the technical terms or scientific terms involved in this application shall have the general meaning understood by people with ordinary skills in the technical field to which this application belongs. The words "one", "a", "a", "the", "these" and the like in this application do not represent quantitative restrictions, and they can be singular or plural. The terms "include", "comprise", "have" and any variants thereof involved in this application are intended to cover non-exclusive inclusions; for example, a process, method and system, product or device comprising a series of steps or modules (units) is not limited to the listed steps or modules (units), but may include unlisted steps or modules (units), or may include other steps or modules (units) inherent to these processes, methods, products or devices. The words "connect", "connected", "coupled" and the like involved in this application are not limited to physical or mechanical connections, but may include electrical connections, whether directly or indirectly. The "multiple" involved in this application refers to two or more. "And / or" describes the association relationship of associated objects, indicating that there can be three relationships. For example, "A and / or B" can mean: A exists alone, A and B exist at the same time, and B exists alone. Usually, the character " / " indicates that the objects associated with each other are in an "or" relationship. The terms "first", "second", "third", etc. involved in this application are only used to distinguish similar objects and do not represent a specific ordering of the objects.
[0031] In an embodiment of the present invention, a method for matching cargo entry locations is provided. Figure 1 is a flow chart of a method for matching cargo storage locations provided in an embodiment of the present invention. Figure 1 As shown, the process includes the following steps:
[0032] Step S110, inputting the cargo feature data of each cargo to be stored into the trained clustering model, and obtaining the clustering result of each cargo to be stored through the clustering model.
[0033] Each cargo characteristic data includes multiple attribute data of the cargo. For example, the multiple attribute data are length, width, height, volume, and weight. That is, the cargo characteristic data of a cargo is composed of the length, width, height, volume, and weight of the cargo.
[0034] In order to facilitate the clustering model to process the cargo characteristic data, the cargo characteristic data also needs to be preprocessed.
[0035] In some of the embodiments, before inputting the cargo feature data of each cargo to be stored into the trained clustering model, the method further includes:
[0036] Step S100, standardizing or normalizing each cargo feature data, and converting each cargo feature data after standardization or normalization into a vector form;
[0037] The standardization formula is:
[0038]
[0039] Among them, X represents the target attribute data before processing in the cargo characteristic data, μ represents the mean of the target attribute data, σ represents the standard deviation of the target attribute data, and Z represents the target attribute data after processing in the cargo characteristic data. The target attribute data refers to the attribute data currently being calculated, and each attribute data in each cargo characteristic data needs to undergo the above-mentioned standardization processing. The mean of the target attribute data refers to the mean of the target attribute data in all cargo characteristic data, and the standard deviation of the target attribute data refers to the standard deviation of the target attribute data in all cargo characteristic data.
[0040] The normalization formula is:
[0041]
[0042] Among them, X represents the target attribute data before processing in the cargo feature data, X max and X minRespectively represent the maximum and minimum values of the target attribute data, and Y represents the processed target attribute data in the cargo characteristic data. Each attribute data in each cargo characteristic data needs to undergo the above normalization processing, and the maximum and minimum values of the target attribute data refer to the maximum and minimum values of the target attribute data in all cargo characteristic data.
[0043] It should be noted that either standardization processing or normalization processing can be selected, that is, in some embodiments, the cargo characteristic data is standardized, and in other embodiments, the cargo characteristic data is normalized, which is mainly used to scale each data value to a certain range to facilitate the subsequent construction of feature vectors.
[0044] For example, the cargo feature data can be converted into a vector form according to the dimensions of <length, width, height, volume, weight>. For example, using the data normalization method, a vector <0.5, 0.3, 0.4, 0.06, 0.7> can be constructed.
[0045] Step S120, matching the target storage location of each cargo to be stored according to the clustering result of each cargo to be stored.
[0046] It is usually more convenient to store similar goods in a centralized location. After the clustering results of each incoming goods are determined, some relatively concentrated adjacent cargo locations can be allocated to the same goods. For example, if the category of some incoming goods is A, and by querying the receipt records and storage records of the warehouse management system, it is found that goods of category A are already stored in the warehouse, then the incoming goods of category A can be allocated to the cargo locations adjacent to the existing goods of the same category. If there are no goods of category A stored in the warehouse, some centralized cargo locations can be allocated to the incoming goods of category A nearby.
[0047] At the same time, when allocating cargo space, attention should be paid to the storage requirements of the cargo. In some embodiments, step S120 specifically includes:
[0048] The target storage location of each incoming cargo is matched according to the clustering results and storage requirements of each incoming cargo.
[0049] Storage requirements include storage temperature and storage humidity, etc. On the premise of meeting the centralized storage of similar goods, it is also necessary to ensure that the storage space allocated for the goods to be stored can meet the storage temperature and storage humidity requirements of the corresponding goods.
[0050] The data processing steps of the clustering model include:
[0051] Step S111, calculating the Euclidean distance between the cargo feature data.
[0052] The preprocessed cargo feature data is in the form of <0.5, 0.3, 0.4, 0.06, 0.7>, and the Euclidean distance between each cargo feature data can be calculated.
[0053] Step S112, calculate the local density of each cargo feature data according to the kernel function, the kernel function is:
[0054]
[0055] Among them, ρ i represents the local density of the characteristic data of the i-th cargo, d ij represents the Euclidean distance between the i-th commodity feature data and the j-th commodity feature data, d c represents the cutoff distance, n represents the total number of cargo characteristic data, and χ(x) represents the step function.
[0056] Essentially, the local density of a commodity feature data refers to the number of other commodity feature data within a circle with the commodity feature data as the center and the cutoff distance as the radius. The higher the local density of a commodity feature data, the more other commodity feature data gather towards this commodity feature data, and the more likely this commodity feature data is to be a cluster center.
[0057] Step S113: determining at least one cluster center data from the plurality of cargo feature data according to the local density of each cargo feature data.
[0058] When selecting cluster center data, it should be ensured that the cluster center data has a sufficiently high local density. Preferably, in order to ensure that there is sufficient difference between the cluster center data, so as to ensure that the different data clusters finally formed have obvious differences, the Euclidean distance between any two cluster center data needs to be far enough.
[0059] Therefore, in some embodiments, step S113 specifically includes:
[0060] Under the premise of meeting the target conditions, as many cluster center data as possible are determined from multiple cargo feature data. The target conditions include: the local density of the cluster center data is greater than a first threshold; the Euclidean distance between any two cluster center data is greater than a second threshold.
[0061] When the selected cluster center data meets the target conditions, it can be ensured that the cluster center data has a sufficiently high local density and the Euclidean distance between each cluster center data is large enough. At the same time, under the premise of meeting the target conditions, determining as many cluster center data as possible can ensure that the final classification of goods is more refined, thereby ensuring that the similarity of similar goods is higher.
[0062] Specifically, the present invention provides two specific means for determining cluster center data.
[0063] In one embodiment, under the premise of satisfying the target condition, as many cluster center data as possible are determined from the multiple cargo feature data, including:
[0064] Traverse different first data combinations from the cargo feature data whose local density is greater than a first threshold; determine the first data combination whose Euclidean distance between any two cargo feature data is greater than a second threshold as the second data combination; and determine the cargo feature data in the second data combination with the largest number of data as the cluster center data.
[0065] Exemplarily, when the cargo feature data with a local density greater than the first threshold is data a, b, c, d, e, the first data combination can be a data combination selected from data a, b, c, d, e in any number, such as a combination of a and b, or a combination of data a, b, c, d, e. Then, in each first data combination, it is calculated whether the Euclidean distance between any two cargo feature data is greater than the second threshold, and the first data combination that meets the condition is determined as the second data combination. For example, if the combination of a and b and the combination of a, c and d meet the condition, the data a, c and d are finally used as the cluster center data because the combination has more data.
[0066] In another embodiment, under the premise of satisfying the target condition, as many cluster center data as possible are determined from the multiple cargo feature data, including:
[0067] From the cargo feature data whose local density is greater than the first threshold, different data combinations are traversed in order of data quantity from large to small, until the Euclidean distance between any two cargo feature data in the current data combination is greater than the second threshold, and the cargo feature data in the current data combination is determined as the cluster center data.
[0068] Exemplarily, when the cargo characteristic data with a local density greater than the first threshold is data a, b, c, d, and e, traversing different data combinations in order of data quantity from large to small means that the combination of data a, b, c, d, and e is determined first, followed by the combination of data a, b, c, d, and the combination of data a, b, c, e, etc. If the combination of data a, b, c, d, and e does not satisfy the condition that the Euclidean distance between any two cargo characteristic data is greater than the second threshold, then continue to determine whether the combination of data a, b, c, and d satisfies the above condition. If it does, directly use data a, b, c, and d as cluster center data without continuing to traverse subsequent combinations with less data quantity. Therefore, compared with the previous embodiment, this embodiment has a higher efficiency in determining cluster center data.
[0069] Furthermore, the clustering model needs to be pre-trained with a large amount of historical data. The clustering model adopts unsupervised training. The parameters optimized during the training process of the clustering model include the cutoff distance, the first threshold, and the second threshold. As can be seen from the above description, different values of the cutoff distance, the first threshold, and the second threshold will produce different clustering effects. The clustering model is trained with a large amount of historical data, and the above parameters are optimized and saved.
[0070] It should be noted that in the data preparation stage of model training, data preprocessing is also required, including standardization or normalization, and conversion into vector form, which is the same as the preprocessing steps described above. Also, there may be some abnormal data in the historical data, so data cleaning is required. On the one hand, missing values need to be removed or filled, that is, if an attribute is missing, the data group is eliminated or filled with the average value of the attribute. On the other hand, outliers need to be identified and processed (such as too large or too small size, weight, etc.), and data groups need to be eliminated.
[0071] Step S114, allocating each cargo characteristic data except the cluster center data to the cluster of the cluster center data closest to itself, and the clustering results of each cluster center data are used as the clustering results corresponding to each cargo to be stored.
[0072] When the data of each cluster center is determined, for each remaining cargo feature data, it is assigned to the cluster where the cluster center data with the smallest Euclidean distance is located, and finally each data cluster can be formed. The cargo corresponding to each data in each data cluster is the same type of cargo.
[0073] In summary, the present invention realizes the classification of goods to be stored by density clustering algorithm, which greatly improves the efficiency of goods storage compared with the traditional goods storage method that uses manual classification of goods. At the same time, the density clustering algorithm used in the present invention is not a traditional density clustering algorithm, but an improvement on the traditional density clustering algorithm, which has higher clustering accuracy and clustering efficiency. It solves the problem of low efficiency of goods storage in traditional goods storage matching methods.
[0074] As described above, at least one embodiment of each part of the cargo entry location matching method provided by the present invention has been described separately. It should be noted that each part of the cargo entry location matching method can be combined using different embodiments to form different embodiments of the cargo entry location matching method.
[0075] Reference Figure 2 In a specific overall embodiment, the method for matching incoming goods to storage locations includes:
[0076] 1. Use density clustering algorithm (model) to classify goods.
[0077] (1) Data preparation.
[0078] Carry out appearance inspection and weighing of goods to obtain physical properties of goods: length, width, height, volume, weight, etc., and obtain storage requirements of goods: temperature, humidity, etc. Temperature, humidity, etc. are physical quantities within a range and cannot be used in calculations, but can only be used as constraints. Group data in the order of attributes, for example, according to the dimensions of <length, width, height, volume, weight>.
[0079] (2) Data cleaning.
[0080] When using density clustering algorithms, data integrity is required. For some abnormal data groups, on the one hand, missing values need to be removed or filled, that is, if an attribute is missing, the data group should be eliminated or filled with the average value of the attribute. On the other hand, outliers (such as too large or too small size, weight, etc.) need to be identified and processed, and data groups need to be eliminated.
[0081] (3) Data standardization or normalization.
[0082] Since the dimensions and ranges of different attributes may vary greatly (for example, length may be in meters, while weight may be in kilograms), direct clustering may cause some attributes to dominate the clustering process. Therefore, the data needs to be standardized or normalized, and you can choose one method.
[0083] Data normalization (Z-score normalization)
[0084] The data is scaled to a target of 0 mean and 1 standard deviation. For each attribute X, the calculation formula for its standardized value Z is:
[0085]
[0086] Among them, μ is the mean of the attribute and σ is the standard deviation of the attribute.
[0087] Data Normalization
[0088] Scale the data to a specific range, usually [0,1]. For each attribute X, its normalized value Y is calculated as:
[0089]
[0090] Among them, X min and X max are the minimum and maximum values of the attribute, respectively.
[0091] (4) Construct vector data.
[0092] For example, according to the dimensions of <length, width, height, volume, weight>, using the data normalization method, a vector group is constructed as <0.5, 0.3, 0.4, 0.06, 0.7>.
[0093] (5) Use improved density clustering algorithm for model training.
[0094] The improved density clustering algorithm is as follows:
[0095] Ⅰ. Calculate the distance between any two points: use Euclidean distance as the metric.
[0096] Ⅱ. Calculate local density: According to the selected kernel function (the truncated kernel function can be used) and the truncation distance d c , calculate the local density of each data point.
[0097] III. Calculate relative distance: Calculate the relative distance of each data point based on the calculation results of local density.
[0098] IV. Determine the cluster center: In the scatter plot of local density-relative distance, select those points with large local density and relative distance as the cluster center.
[0099] V. Assign remaining points: Assign the remaining data points to the cluster where the nearest cluster center is located.
[0100] The truncated kernel function is a simpler kernel function, which has the following form:
[0101]
[0102] Here, χ(x) is a step function defined as:
[0103]
[0104] In the truncated kernel function, data point j is considered to contribute to the local density of data point i only when the distance between data point j and data point i is less than the truncation distance.
[0105] (6) Model clustering results.
[0106] The density clustering algorithm is an unsupervised machine learning algorithm, which means that the data classification results are not known in advance. The model is trained through massive data, and the parameters are continuously adjusted to make the model reach the optimal state, and the parameter values are recorded.
[0107] (7) The data to be tested is input into the model for classification.
[0108] The value to be detected is input into the trained model to obtain the classification result.
[0109] 2. Cargo location matching.
[0110] In a conventional warehouse layout, a warehouse is divided into multiple storage areas, each storage area is divided into multiple shelves, and a shelf is divided into multiple storage locations. The centralized storage of goods of the same classification can facilitate temperature and humidity control and inventory of goods. Therefore, after using the model to obtain the classification of goods to be stored, the receipt records and storage records of the warehouse management system are searched, and the storage location is allocated nearby. In addition, the most suitable storage location is allocated according to the constraints such as temperature and humidity of the goods.
[0111] The method for matching the cargo entry cargo location in the above embodiment first extracts the characteristic data, then normalizes the data, trains the model using the improved density clustering algorithm, and inputs the data to be classified into the model for classification. The cargo classification model is trained using a machine learning algorithm, and the model is classified according to the cargo location characteristic data input, and finally matching is performed according to the cargo location constraint conditions and the cargo historical storage records.
[0112] In summary, the method for matching goods entering the warehouse in the above embodiment has the following technical effects:
[0113] The traditional warehouse management system selects the storage location for incoming goods directly based on constraints such as length, width and height, and judges one by one, which is inefficient, especially when the number of storage locations is particularly large, which affects the warehousing efficiency. The present invention uses a density clustering algorithm, which is a machine learning algorithm, which can improve the efficiency of storage location selection and truly realize the intelligence of the warehouse; according to the classification results of the goods to be stored, combined with the storage records of the same classification in the warehousing system, the storage locations can be quickly matched.
[0114] In the embodiments of the present invention, a device for matching cargo storage locations is also provided, which is used to implement the above embodiments and preferred implementations, and will not be repeated for those that have been described. The terms "module", "unit", "subunit", etc. used below can be a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.
[0115] The device for matching cargo storage locations is characterized by comprising:
[0116] The cargo clustering module is used to input cargo feature data of each cargo to be stored into the trained clustering model, and obtain the clustering results of each cargo to be stored through the clustering model;
[0117] The cargo location matching module is used to match the target cargo location of each cargo to be stored according to the clustering results of each cargo to be stored;
[0118] The data processing steps of the clustering model include:
[0119] Calculate the Euclidean distance between the cargo feature data;
[0120] The local density of each cargo feature data is calculated based on the kernel function, and the kernel function is:
[0121]
[0122] Among them, ρ i represents the local density of the characteristic data of the i-th cargo, d ij represents the Euclidean distance between the i-th commodity feature data and the j-th commodity feature data, d c represents the cutoff distance, n represents the total number of cargo characteristic data, and χ(x) represents the step function;
[0123] Determining at least one cluster center data from the plurality of cargo feature data according to the local density of each cargo feature data;
[0124] Each cargo feature data except the cluster center data is assigned to the cluster of the cluster center data closest to itself, and the clustering results of each cluster center data are used as the clustering results corresponding to each cargo to be stored.
[0125] It should be noted that the above modules can be functional modules or program modules, and can be implemented by software or hardware. For modules implemented by hardware, the above modules can be located in the same processor; or the above modules can be located in different processors in any combination.
[0126] In an embodiment of the present invention, a warehouse entry system is also provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to run the computer program to execute the goods entry location matching method provided by the present invention.
[0127] It should be understood that the specific embodiments described herein are only used to explain the application, rather than to limit it. Based on the embodiments provided in this application, all other embodiments obtained by ordinary technicians in this field without creative work are within the protection scope of this application.
[0128] Obviously, the drawings are only some examples or embodiments of the present application. For ordinary technicians in the field, the present application can also be applied to other similar situations based on these drawings without creative work. In addition, it is understandable that although the work done in this development process may be complicated and lengthy, for ordinary technicians in the field, certain changes in design, manufacturing or production based on the technical content disclosed in this application are only conventional technical means and should not be regarded as insufficient content disclosed in this application.
Claims
1. A method for matching cargo storage locations, characterized in that: include: Inputting the cargo feature data of each cargo to be stored into the trained clustering model, and obtaining the clustering result of each cargo to be stored through the clustering model; Matching the target cargo location of each of the goods to be stored according to the clustering results of each of the goods to be stored; The data processing steps of the clustering model include: Calculating the Euclidean distance between the cargo characteristic data; The local density of each of the cargo feature data is calculated according to the kernel function, and the kernel function is: Among them, ρ i represents the local density of the characteristic data of the i-th cargo, d ij represents the Euclidean distance between the i-th commodity feature data and the j-th commodity feature data, d c represents the cutoff distance, n represents the total number of cargo characteristic data, and χ(x) represents the step function; Determining at least one cluster center data from a plurality of the cargo feature data according to the local density of each of the cargo feature data; Each cargo feature data except the cluster center data is assigned to the cluster where the cluster center data closest to itself is located, and the clustering results of each cluster center data are used as the clustering results corresponding to each cargo to be stored.
2. The method for matching cargo storage locations according to claim 1, characterized in that: Determining at least one cluster center data from a plurality of the cargo feature data according to the local density of each of the cargo feature data comprises: Under the premise of satisfying the target conditions, as many cluster center data as possible are determined from the plurality of cargo feature data, wherein the target conditions include: The local density of the cluster center data is greater than a first threshold; The Euclidean distance between any two of the cluster center data is greater than a second threshold.
3. The method for matching cargo storage locations according to claim 2, characterized in that: Under the premise of meeting the target conditions, as many cluster center data as possible are determined from the plurality of cargo feature data, including: Traversing different first data combinations from the cargo feature data having a local density greater than the first threshold; Determine a first data combination in which the Euclidean distance between any two cargo feature data is greater than the second threshold as a second data combination; The cargo feature data in the second data combination with the largest data quantity is determined as the cluster center data.
4. The method for matching cargo storage locations according to claim 2, characterized in that: Under the premise of meeting the target conditions, as many cluster center data as possible are determined from the plurality of cargo feature data, including: From the cargo feature data whose local density is greater than the first threshold, different data combinations are traversed in order of data quantity from large to small, until the Euclidean distance between any two cargo feature data in the current data combination is greater than the second threshold, and the cargo feature data in the current data combination is determined as the cluster center data.
5. The method for matching cargo storage locations according to claim 2, characterized in that: The clustering model adopts unsupervised training, and the parameters of the clustering model optimized during the training process include the cutoff distance, the first threshold and the second threshold.
6. The method for matching cargo storage locations according to claim 1, characterized in that: Each of the cargo characteristic data includes a plurality of attribute data of the cargo; Before the cargo feature data of each cargo to be stored is input into the trained clustering model, it also includes: Performing standardization or normalization on each of the cargo characteristic data, and converting each of the cargo characteristic data after standardization or normalization into a vector form; The standardization processing formula is: Wherein, X represents the target attribute data before processing in the cargo characteristic data, μ represents the mean of the target attribute data, σ represents the standard deviation of the target attribute data, and Z represents the target attribute data after processing in the cargo characteristic data; The normalization processing formula is: Among them, X represents the target attribute data before processing in the cargo feature data, X max and X min They represent the maximum and minimum values of the target attribute data respectively, and Y represents the target attribute data after processing in the cargo feature data.
7. The method for matching cargo storage locations according to claim 6, characterized in that: The multiple attribute data are length, width, height, volume, and weight.
8. The method for matching cargo storage locations according to claim 6, characterized in that: Matching the target cargo location of each of the goods to be stored according to the clustering results of each of the goods to be stored includes: The target cargo location of each of the goods to be stored is matched according to the clustering result and storage demand of each of the goods to be stored.
9. A device for matching cargo storage locations, characterized in that: include: A cargo clustering module, used to input cargo feature data of each cargo to be stored into a trained clustering model, and obtain clustering results of each cargo to be stored through the clustering model; A cargo location matching module, used for matching the target cargo location of each of the cargoes to be stored according to the clustering results of each of the cargoes to be stored; The data processing steps of the clustering model include: Calculating the Euclidean distance between the cargo characteristic data; The local density of each of the cargo feature data is calculated according to the kernel function, and the kernel function is: Among them, ρ i represents the local density of the characteristic data of the i-th cargo, d ij represents the Euclidean distance between the i-th commodity feature data and the j-th commodity feature data, d c represents the cutoff distance, n represents the total number of cargo characteristic data, and χ(x) represents the step function; Determining at least one cluster center data from a plurality of the cargo feature data according to the local density of each of the cargo feature data; Each cargo feature data except the cluster center data is assigned to the cluster where the cluster center data closest to itself is located, and the clustering results of each cluster center data are used as the clustering results corresponding to each cargo to be stored.
10. A storage warehousing system, comprising a memory and a processor, characterized in that: The memory stores a computer program, and the processor is configured to run the computer program to execute the method for matching incoming goods to storage locations according to any one of claims 1 to 8.