Digital and efficient storage method for warehouse goods based on NLP technology
By building a warehouse database and analyzing historical instructions and cargo attributes, the warehousing inaccuracy problem caused by natural language ambiguity in NLP technology is solved, the efficiency and reliability of warehousing cargo storage are improved, and an accurate cargo storage solution is generated.
Patent Information
- Application Number
- CN202411117248.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-15
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2044-08-15
AI Technical Summary
Due to the ambiguity of natural language, the system cannot accurately understand the daily terms of warehousing management personnel, resulting in low accuracy and reliability of cargo storage suggestions and low efficiency.
By building a warehouse database, using historical oral instructions and cargo description information, analyzing warehouse storage rules and real-time status, extracting similar historical instructions and cargo attributes, calculating the availability and environmental similarity of instructions and operations, generating accurate cargo storage solutions, and reducing ambiguity The possibility of error storage caused.
Improve the efficiency, stability and reliability of storage of warehousing goods, reduce erroneous storage caused by ambiguity in verbal command identification, and achieve more accurate and reliable cargo storage suggestions.
Smart Images

Figure CN118628025B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of intelligent warehouse management, and in particular to a digital and efficient storage method for warehoused goods based on NLP technology. Background Art
[0002] With the rapid development of the global economy and the rise of e-commerce, the logistics industry has become increasingly complex and critical. The introduction of artificial intelligence has brought new opportunities to the logistics industry. By improving efficiency, reducing costs and optimizing user experience, AI is completely changing the operation mode of the logistics industry. NLP technology (Natural Language Processing) mainly studies how to understand, interpret and generate human language through computers to achieve more natural and efficient human-computer interaction. It is mainly used in machine translation, speech recognition, information retrieval, chatbots and other scenarios. In the process of cargo warehousing in the logistics industry, managers need to record and store the information of each piece of cargo, such as cargo type, quantity, storage time, warehousing status, etc., so they usually face a huge amount of data processing. At the same time, when extracting warehoused goods, it usually takes a lot of time to determine the specific goods, cargo storage location, status, etc. The huge workload leads to low processing efficiency. Therefore, more and more NLP technologies are used in the process of cargo warehousing in the logistics industry.
[0003] In the prior art, NLP technology is applied to warehouse management. By extracting and analyzing the cargo information and warehouse storage rules, the characteristics of each cargo and its storage requirements are analyzed, and the storage strategy is generated in combination with the layout of the warehouse space, such as which area and shelf the cargo should be stored in. At the same time, the system dynamically updates the storage suggestions for the cargo as the warehouse status changes in real time, such as the entry of new cargo or the exit of cargo. However, when using NLP technology to generate automated storage suggestions, due to the ambiguity of natural language, the system's mechanical arms and intelligent robots cannot accurately compare the daily language of the warehouse manager with the operating instructions, such as "store in the corner area", "store in area A", etc. The system cannot determine the exact cargo placement area, that is, the operating instructions issued by the warehouse manager are ambiguous due to the lack of specific details or context, resulting in the system lacking sufficient accuracy and reliability when learning and generating storage suggestions for cargo. Summary of the invention
[0004] In order to solve the technical problem that the digitized storage method of warehouse goods based on NLP technology in the prior art is prone to ambiguity in voice recognition, which leads to erroneous storage, resulting in low stability, reliability and efficiency of storage based on NLP technology, the purpose of the present invention is to provide a digital and efficient storage method for warehouse goods based on NLP technology, and the technical solution adopted is as follows:
[0005] A digital and efficient storage method for warehouse goods based on NLP technology, the method comprising:
[0006] Step S1: Building a warehouse database based on the warehouse management personnel's historical verbal instructions and feedback on the corresponding goods storage location, goods description information, warehouse storage rules, and real-time warehouse status;
[0007] Step S2: according to the similarity between the latest oral instruction and the historical oral instruction of the warehouse manager, the corresponding warehouse operations of the similar historical oral instructions in the warehouse database are extracted to form a possible warehouse operation set of the latest oral instruction;
[0008] Step S3: obtaining similarity of the cargo attributes according to the cargo attributes extracted by the latest verbal instruction and the cargo attributes corresponding to the storage operations in the possible storage operation set;
[0009] Step S4: obtaining the availability of the storage operation according to the similarity between the latest verbal instruction and the historical verbal instructions, the similarity of the attributes of the goods, and all available storage operations in the possible storage operation set of the latest verbal instruction;
[0010] Step S5: Obtain the environmental similarity and environmental qualification of different vacant locations according to the environmental parameters of the historical storage locations of the goods corresponding to the latest verbal instructions in the warehouse database, the optimal storage environment, and the environmental parameters of all vacant locations in the possible storage operation set;
[0011] Step S6: According to the environmental similarity, environmental qualification and availability of storage operations of different available locations, the storage feasibility of the goods corresponding to the latest verbal instruction at different available locations is obtained;
[0012] Step S7: When the storage feasibility is greater than the first threshold, the corresponding vacant locations can be used for goods storage.
[0013] Furthermore, in step S1:
[0014] The cargo description information includes cargo name, cargo type, cargo size, delivery period, and order batch;
[0015] The warehouse storage rules include safety specifications, cargo classification rules, and warehouse space layout;
[0016] The real-time warehouse status includes real-time temperature and humidity information of each area of the warehouse, real-time inventory status, and real-time cargo flow information.
[0017] Furthermore, the similarity between the latest oral instruction of the warehouse manager and the historical oral instruction in step S2 is specifically:
[0018] Decompose the historical oral instructions of warehouse managers into individual words or phrases to form a set of instruction words;
[0019] Construct a bag-of-words model for the instruction word set and vectorize the bag-of-words model;
[0020] The cosine similarity between the text vector of the latest verbal instruction and the text vector of the historical verbal instruction is calculated to obtain the similarity between the latest verbal instruction and the historical verbal instruction of the warehouse manager.
[0021] Furthermore, the corresponding storage operations of similar historical oral instructions extracted from the warehouse database in step S2 are specifically:
[0022] When the cosine similarity between the text vector of the latest oral instruction and the text vector of the historical oral instruction is greater than a second threshold, the historical oral instruction is considered to be similar to the latest oral instruction, and the corresponding storage operation of the historical oral instruction is extracted.
[0023] Furthermore, in step S3, the similarity of the goods attributes obtained according to the goods attributes extracted from the latest oral instruction and the goods attributes corresponding to the storage operation in the possible storage operation set is specifically:
[0024] Extract the type of goods according to the instruction word information in the latest verbal instruction to form the basic attributes of the goods in the latest verbal instruction;
[0025] The cosine similarity between the basic attributes of the goods in the latest verbal instruction and the attributes of the goods corresponding to the storage operation is calculated to obtain the similarity of the attributes of the goods.
[0026] Furthermore, in step S4, the greater the similarity between the latest oral instruction and the historical oral instruction, and the greater the similarity between the attributes of the goods, the greater the availability of the warehousing operation.
[0027] Furthermore, in step S5, the smaller the absolute value of the difference between the environmental parameters of the historical storage location of the goods corresponding to the latest oral instruction in the warehouse database and the environmental parameters of all the vacant locations in the possible storage operation set, the higher the environmental similarity of the corresponding vacant location.
[0028] Furthermore, in step S5, the smaller the absolute value of the difference between the optimal storage environment of the goods corresponding to the latest verbal instruction in the warehouse database and the environmental parameters of all vacant locations in the possible storage operation set, the higher the environmental qualification of the corresponding vacant location.
[0029] Furthermore, in step S6, when the environmental qualification of different vacant locations is higher, the environmental similarity of different vacant locations is higher, and the availability of storage operations is greater, the storage feasibility of the goods corresponding to the latest verbal instructions in different vacant locations is higher.
[0030] Furthermore, after step S7, the method further includes ranking the corresponding vacant locations whose storage feasibility is greater than the first threshold value from large to small according to the storage feasibility, and providing the ranking results to warehouse management personnel, who decide the storage location of the goods based on the storage feasibility ranking of the vacant locations and the actual warehouse storage situation.
[0031] The present invention has the following beneficial effects:
[0032] First, it can reduce the possibility of incorrect storage of goods due to ambiguity in the recognition of verbal instructions of warehouse managers. By building a warehouse database based on the historical verbal instructions of warehouse managers and the feedback of the corresponding goods storage location, goods description information, warehouse storage rules, and real-time warehouse status; extracting the corresponding storage operations of similar historical verbal instructions in the warehouse database based on the similarity between the latest verbal instructions of warehouse managers and historical verbal instructions, forming a possible storage operation set of the latest verbal instructions; obtaining the similarity of goods attributes based on the goods attributes extracted from the latest verbal instructions and the goods attributes corresponding to the storage operations in the possible storage operation set; obtaining the availability of storage operations based on the similarity between the latest verbal instructions and historical verbal instructions, the similarity of goods attributes, and all available storage operations in the possible storage operation set of the latest verbal instructions. In this way, by converting and analyzing the verbal instructions of warehouse managers, the originally ambiguous verbal instructions are converted into natural language that the system can understand, reducing the possibility of incorrect storage due to ambiguity.
[0033] Second, it can improve the efficiency, stability and reliability of warehouse goods storage. Based on the above conversion and analysis of the warehouse management personnel's verbal instructions, the environmental similarity of different vacant locations is obtained according to the environmental parameters of the historical storage locations of the goods corresponding to the latest verbal instructions in the warehouse database and the environmental parameters of all vacant locations in the possible storage operation set; the environmental qualification of different vacant locations is obtained according to the best storage environment of the goods corresponding to the latest verbal instructions in the warehouse database and the environmental parameters of all vacant locations in the possible storage operation set; the storage feasibility of the goods corresponding to the latest verbal instructions in different vacant locations is obtained according to the environmental qualification of different vacant locations, the environmental similarity of different vacant locations and the availability of storage operations; multiple reasonable storage plans are output to improve the efficiency, stability and reliability of obtaining storage suggestions. BRIEF DESCRIPTION OF THE DRAWINGS
[0034] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0035] Figure 1 A flow chart of a method for efficient digital storage of warehouse goods based on NLP technology provided by one embodiment of the present invention. DETAILED DESCRIPTION
[0036] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation method, structure, characteristics and effects of the digital and efficient storage method of warehouse goods based on NLP technology proposed by the present invention are described in detail as follows in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by technicians in the technical field of the present invention.
[0037] The specific scheme of the digital and efficient storage method for warehouse goods based on NLP technology provided by the present invention is described in detail below with reference to the accompanying drawings.
[0038] The specific scenario targeted by the embodiments of the present invention is: when using NLP technology to generate automated warehousing recommendations, due to the ambiguity of natural language, the system may not be able to accurately understand the daily language used by warehouse managers, such as "store in the corner area", "store casually in area A", etc. The intelligent system robotic arms and intelligent robots cannot accurately compare the daily language of warehouse managers with operating instructions. That is, the instructions issued by warehouse managers are ambiguous with the system operation quality due to the lack of specific details or context, resulting in the NLP system lacking sufficient accuracy and reliability when learning and generating warehousing recommendations for goods, thereby generating erroneous or unreasonable warehousing recommendations.
[0039] See also Figure 1 , which shows a flow chart of a method for digitally storing warehouse goods based on NLP technology provided by an embodiment of the present invention, the method comprising the following steps:
[0040] Step S1: Build a warehouse database based on the warehouse management personnel's historical verbal instructions and the corresponding feedback on the goods storage location, goods description information, warehouse storage rules, and real-time warehouse status.
[0041] Firstly, NLP technology is used to collect historical oral instructions from warehouse managers, and feedback corresponding to the storage of goods is obtained after the oral instructions are issued. The oral instructions and the feedback corresponding to the storage of goods are matched one by one and stored to build an instruction feedback database.
[0042] In this embodiment, the corresponding relationship between the historical oral instructions of the warehouse management personnel and the feedback of the corresponding cargo storage location is shown in Table 1 below.
[0043] Table 1
[0044]
[0045] Then, NLP technology is used to obtain cargo description information. Specifically, cargo description information includes cargo name, cargo type, cargo size, delivery period, and order batch. These cargo description information can extract key information from texts such as cargo name, order information, and historical storage data, such as cargo type, size, order details (delivery period, order quantity), etc. In this way, cargo description information can be obtained through NLP technology to build a cargo information database.
[0046] In this embodiment, the cargo description information obtained based on NLP technology is shown in Table 2 below.
[0047] Table 2
[0048]
[0049] Furthermore, NLP technology is used to analyze warehouse storage rules. Specifically, warehouse storage rules include safety specifications, cargo classification rules, warehouse space layout requirements, etc. In this way, a warehouse information data sample library can be constructed by analyzing warehouse storage rules.
[0050] Furthermore, NLP technology is used to analyze the real-time warehouse status of the warehouse. Specifically, the real-time warehouse status includes the real-time temperature and humidity information of each area of the warehouse, the real-time inventory status, and the real-time cargo flow information. Sensors are deployed in the warehouse to collect the status data of the warehouse environment and warehouse equipment, and NLP technology is used to analyze the real-time inventory level, cargo flow status, temperature and humidity information (specific cargo storage environment requirements) in the warehouse collected by these sensors to build a real-time warehouse status information library.
[0051] Furthermore, all data in the instruction feedback database, the cargo information database, the warehouse information data sample database, and the real-time warehouse status information database are integrated to construct a warehouse database. The data in the warehouse database will constitute the data source in the digital and efficient storage method for warehouse cargo based on NLP technology in this embodiment.
[0052] Step S2: According to the similarity between the latest oral instruction and the historical oral instruction of the warehouse manager, the corresponding warehouse operations of the similar historical oral instructions in the warehouse database are extracted to form a possible warehouse operation set of the latest oral instruction.
[0053] First, the latest verbal instructions of the managers are matched with the warehouse feedback of the goods, so that the more vague natural language can be matched with specific warehouse operations. That is, the latest verbal instructions are used as language labels for warehouse operations. When the system obtains new instructions, it quickly obtains similar warehouse operations according to the language labels to obtain the possible operation set of the latest verbal instructions.
[0054] Among them, the similarity between the latest oral instructions of the warehouse manager and the historical oral instructions in step S2 is specifically as follows: decomposing the historical oral instructions of the warehouse manager into separate words or phrases to form an instruction word set; constructing a bag-of-words model for the instruction word set and vectorizing the bag-of-words model; calculating the cosine similarity between the text vector of the latest oral instruction and the text vector of the historical oral instruction, and obtaining the similarity between the latest oral instructions of the warehouse manager and the historical oral instructions.
[0055] Specifically, speech recognition technology and NLP technology are first used to decompose the oral instructions of warehouse managers into separate words or phrases, which are then grouped to obtain an instruction word set. A bag-of-words model is constructed based on the instruction word set, and the bag-of-words model is vectorized. For example, text 1 is "store fresh goods on the fifth row of shelves in the freezer area", and text 2 is "store the goods on the farthest row of shelves". Words in text 1 and text 2 are extracted through preprocessing, text 1 is "fresh goods, storage, freezer area, fifth row, shelf", and text 2 is "goods, storage, farthest, shelf". Then a bag-of-words model is constructed to achieve vectorization, all words are summarized to obtain a total vocabulary, and converted into specific vectors, text 1 [1,1,1,1,1,0,0,], text 2 [0,1,0,0,1,1,1].
[0056] Furthermore, the cosine similarity between the text vector of the latest verbal instruction and the text vector of the historical verbal instruction is calculated to obtain the similarity between the text vector of the latest verbal instruction and the text vector of the historical verbal instruction. (in Indicates goods).
[0057] It should be noted that the specific process of extracting the corresponding warehousing operations of similar historical oral instructions in the warehouse database in step S2 is as follows: when the cosine similarity between the text vector of the latest oral instruction and the text vector of the historical oral instruction is greater than a second threshold, the historical oral instruction is considered to be similar to the latest oral instruction, and the corresponding warehousing operations of the historical oral instruction are extracted.
[0058] Specifically, the second threshold can be a constant of 0.6. That is, when the similarity When the value is 0.6, the corresponding historical oral instruction is considered to be similar in content to the latest oral instruction.
[0059] Furthermore, the corresponding storage operations of all similar historical oral instructions are extracted to construct a possible storage operation set of the current latest oral instruction. , the possible storage operation set Various warehousing operations ,In this way, the analysis of the latest verbal instructions of the warehouse manager is completed.
[0060] In actual smart warehousing, similar verbal instructions may result in different situations when executing warehousing operations. The main reason for this is that the properties of the goods being stored are different. For example, the first row of shelves in a certain area may be used to store fresh goods that need to be refrigerated and placed in the freezer, and time-sensitive goods may be placed in the area closest to the shipping port. In other words, different goods properties and storage conditions have different effects on warehousing operations. Therefore, it is necessary to group possible warehousing operations into different categories. The type of goods in each operation is compared and analyzed with the attributes of the goods that need to be stored, so as to obtain the storage operations of similar goods and reduce the possibility of goods storage errors.
[0061] Step S3: obtaining similarity of the cargo attributes according to the cargo attributes extracted by the latest oral instruction and the cargo attributes corresponding to the storage operations in the possible storage operation set.
[0062] Among them, in step S3, the similarity of the goods attributes extracted according to the latest oral instruction and the goods attributes corresponding to the storage operation in the possible storage operation set is obtained as follows:
[0063] Extract the type of goods according to the instruction word information in the latest verbal instruction to form the basic attributes of the goods in the latest verbal instruction;
[0064] The cosine similarity between the basic attributes of the goods in the latest verbal instruction and the attributes of the goods corresponding to the storage operation is calculated to obtain the similarity of the attributes of the goods.
[0065] Specifically, first, according to the latest verbal instructions corresponding to the goods and storage operation set The corresponding goods attributes of the warehousing operation in the warehousing operation are constructed to construct the goods attribute vector; then, the goods and warehousing operation corresponding to the latest verbal instruction are calculated. The cosine similarity of the corresponding goods attributes is calculated to obtain the similarity of the goods attributes. .
[0066] It should be further explained here that the cargo description information includes relatively complete cargo information such as cargo name, cargo type, cargo size, delivery period, order batch, order quantity, etc. in Table 2 above; the basic attributes of cargo are cargo information extracted based on verbal instructions, for example, the basic attributes of cargo only include the cargo name; in addition, the cargo attributes corresponding to the warehousing operation in the warehousing operation set refer to the cargo attributes including the cargo name and supplemented with key information such as cargo type, so as to facilitate the correspondence between the system and specific warehousing operations, for example, "shelves" are supplemented with "frozen area shelves".
[0067] Furthermore, in step S3, the similarity of the goods attributes When the value is greater than the third threshold, the two goods are considered to have similar properties, and the same storage operation can be used to store the goods corresponding to the latest verbal instruction. Specifically, the third threshold can be a constant of 0.6. That is, when the similarity of the goods attributes is When the latest verbal instruction corresponds to the goods and storage operation set The properties of the goods corresponding to the corresponding warehousing operations are similar.
[0068] In the above analysis of the latest verbal instructions of warehouse managers, the similarity between the text vector of the latest verbal instruction and the text vector of the known instruction is Calculation is performed by comparing the difference between the two vectors in different dimensions. However, oral instructions usually have a certain degree of generalization. For example, different managers have different habits of issuing instructions, language simplification, etc. This will cause some words and phrases in the bag-of-words model to have the same meaning. For example, "goods" may refer to goods with different attributes in different situations, and warehouse managers will only explain which part of the goods should be placed in which area when issuing instructions. In summary, when analyzing the similarity between text vectors of oral instructions, it is necessary to judge the current scenario and combine the context information to supplement the information of the words in the instructions, thereby improving the accuracy of the instruction similarity calculation.
[0069] Specifically, firstly, the speech recognition technology and NLP technology are used to obtain the goods information corresponding to the latest oral instructions of the warehouse management personnel, and the basic attributes of the goods such as the specific type of the current goods and the optimal storage conditions are obtained; then, the historical oral instructions corresponding to the determined storage plan in the current batch of goods are obtained, and the information of the stored goods is recorded at the same time; further, the differences in the types and storage conditions of the goods in the current batch of goods and the stored goods are analyzed. When the goods belong to the same type and the required storage conditions are the same, the current oral instructions are supplemented according to the instruction vocabulary and phrases of the stored goods; when there are no similar goods corresponding to the current oral instructions in the stored goods, the latest oral instructions are supplemented according to the goods' own attributes. For example, when the type of goods belongs to fresh food and needs to be stored at low temperature, "goods" is supplemented as "fresh goods", and "shelves" are supplemented as "frozen area shelves", etc. Finally, the text vector of the latest oral instruction is reconstructed according to the supplemented instruction information, and the cosine similarity between the text vector of the latest oral instruction and the text vector of the known historical oral instructions is calculated, and the availability of the corresponding storage operation in the storage operation set is further obtained.
[0070] Step S4: According to the similarity between the latest verbal instruction and the historical verbal instruction, the similarity of the goods attributes, and all available warehousing operations in the possible warehousing operation set of the latest verbal instruction, the availability of the warehousing operation is obtained. Specifically, the greater the similarity between the latest verbal instruction and the historical verbal instruction and the greater the similarity of the goods attributes, the greater the availability of the warehousing operation.
[0071] Furthermore, the availability calculation formula of the warehousing operation of the warehousing goods digital efficient storage method based on NLP technology proposed in this embodiment is:
[0072]
[0073] in, Indicates goods, Represents a storage operation in the set of possible storage operations. Indicates the goods corresponding to the latest verbal instruction Warehousing Operations The similarity of the properties of the corresponding goods, Indicates goods The text vector of the latest verbal instruction and the storage operation corresponding to the historical verbal instruction The similarity of the text vectors, Indicates goods Corresponding warehousing operations availability.
[0074] In the above-mentioned NLP-based digital efficient storage method for storage operations, the similarity between the attributes of the goods corresponding to the latest verbal instruction and the goods corresponding to the storage operation, and the similarity between the text vector of the latest verbal instruction for the goods and the text vector corresponding to the historical verbal instruction are combined. The two similarities are comprehensively considered and summed to obtain the availability of the storage operation of the NLP-based digital efficient storage method for storage goods. The greater the similarity between the latest verbal instruction and the historical verbal instruction, and the greater the similarity of the goods attributes, the greater the availability of the storage operation. By selecting a larger value for the similarity between the latest verbal instruction and the historical verbal instruction and the similarity of the goods attributes, the storage operation of similar goods can be obtained, reducing the possibility of errors in the storage of goods.
[0075] It should be further explained that, when all operations in the possible storage operation set are traversed and the availability of the corresponding storage operation is greater than the fourth threshold, it is considered that the storage operation meets the storage operation requirements of the goods in the corresponding latest verbal instruction. Specifically, the fourth threshold can be a constant of 1.
[0076] During the storage and delivery process of goods, most goods are usually stored in the warehouse for a certain period of time. This is due to the arrangement of goods delivery and the transportation time of different goods. Usually, more urgent goods are transported with priority. Therefore, in the process of generating new storage operations based on historical storage operations, there may be stranded goods at the target storage location of the goods, which may cause the digital storage process such as robotic arms and intelligent robots to be unable to proceed. At the same time, the environmental parameters of the storage location may also change due to seasonal changes. Therefore, the original storage operation may also be inappropriate due to seasonal changes. Therefore, it is also necessary to make intelligent adjustments to the storage operation according to the specific conditions of the warehouse's environmental parameters.
[0077] Step S5: Obtain the environmental similarity and environmental qualification of different vacant locations based on the environmental parameters of the historical storage locations of the goods corresponding to the latest verbal instructions in the warehouse database, the optimal storage environment, and the environmental parameters of all vacant locations in the possible storage operation set.
[0078] In step S5, the smaller the absolute value of the difference between the environmental parameters of the historical storage location of the goods corresponding to the latest oral instruction in the warehouse database and the environmental parameters of all the vacant locations in the possible storage operation set, the higher the environmental similarity of the corresponding vacant location.
[0079] Specifically, when there are goods at the target storage location of the warehousing operation feedback, the environmental information of the target location during the historical warehousing process of the operation is extracted to obtain the environmental parameters of the historical operation process. For example, all temperature and humidity ranges of the environmental information of the target location during the historical warehousing process of the operation of the goods are taken, and then the average value of the temperature and humidity of the environmental information of the target location during the historical warehousing process is used as the environmental parameter of the historical operation process; then, the environmental parameters of all vacant locations in the current area of the warehouse are combined and compared with the environmental parameters of the historical operation process (the sum of the absolute values of the differences between different environmental parameters) to obtain the environmental similarity of different vacant locations. , the smaller the absolute value of the difference between the environmental parameters of the historical storage location of the goods corresponding to the latest verbal instruction in the warehouse database and the environmental parameters of all the free locations in the possible storage operation set, the higher the environmental similarity of the corresponding free location. In this embodiment, the opposite of the sum of the absolute values of the differences is used as the exponent of an exponential function with the natural constant e as the base, thereby obtaining the environmental similarity. In other embodiments, for example, when the absolute value of the difference between the environmental parameters of the historical storage location of the goods corresponding to the latest verbal instruction in the warehouse database and the environmental parameters of all the free locations in the possible storage operation set is 0, the environmental similarity is 100%; for another example, when the temperature difference and humidity difference between the two differ by 5% and 10% respectively, the difference is obtained by weighted summing and averaging the temperature difference and humidity difference, and then the difference is subtracted from 100%, that is, the environmental similarity at this time is 92.5%.
[0080] It should be noted that all the vacant positions in the possible warehousing operation set refer to all the vacant positions in the warehousing operation set whose availability is greater than the fourth threshold.
[0081] Further, in step S5, the smaller the absolute value of the difference between the best storage environment of the goods corresponding to the latest verbal instruction in the warehouse database and the environmental parameters of all vacant locations in the possible storage operation set, the higher the environmental qualification of the corresponding vacant location.
[0082] Specifically, the absolute value of the difference between the goods description information in the warehouse database corresponding to the latest verbal instruction and the environmental parameters of all the free positions in the possible storage operation set required by the warehouse storage rules and the optimal storage environment of the goods is calculated. The smaller the absolute value of the difference, the higher the environmental qualification of the corresponding free position. In this embodiment, the opposite of the sum of the absolute values of the difference is used as the exponent of the exponential function with the natural constant e as the base, so as to obtain the environmental qualification. In other embodiments, for example, when the absolute value of the difference between the optimal storage environment of the goods required by the warehouse storage rules and the environmental parameters of all the free positions in the possible storage operation set is 0, the environmental qualification is 100%; for another example, when the temperature difference and humidity difference of the two differ by 5% and 10% respectively, the temperature difference and humidity difference are weighted and averaged to obtain the difference, and then 100% minus the difference, that is, the environmental qualification at this time is 92.5%.
[0083] Step S6: According to the environmental similarity, environmental qualification and availability of storage operations of different available locations, the storage feasibility of the goods corresponding to the latest verbal instruction in different available locations is obtained.
[0084] In summary, in the process of warehousing goods, if you want to get the best storage plan for the goods corresponding to the latest verbal instructions, you need to comprehensively consider the environmental qualification of different vacant locations, the environmental similarity of different vacant locations, and the availability of storage operations. The higher the environmental qualification of different vacant locations, the higher the environmental similarity of different vacant locations, and the greater the availability of storage operations, the higher the feasibility of storage of goods corresponding to the latest verbal instructions in different vacant locations.
[0085] The calculation formula for the feasibility of storing goods in different free locations corresponding to the latest verbal instructions of the digital and efficient storage method for stored goods based on NLP technology proposed in this embodiment can be expressed as:
[0086]
[0087] in, Indicates goods, Represents a storage operation in the set of possible storage operations. It represents the environmental qualification of cargo i in different available locations. Indicates goods Corresponding warehousing operations Availability, Indicates goods Corresponding warehousing operations The environmental similarity of different free positions corresponding to Indicates normalization of values.
[0088] In the above-mentioned NLP-based digital storage method for warehouse goods, the latest verbal instructions correspond to the storage feasibility calculation formula of goods in different vacant locations. The environmental qualification of goods i in different vacant locations Positively correlated with warehousing feasibility, cargo Corresponding warehousing operations Availability Positively correlated with storage feasibility, environmental similarity of vacant locations It is obtained based on the environment corresponding to the storage operation with the greatest availability of storage operation, so the environmental similarity of the vacant location Availability as a warehousing operation The correction, therefore, It represents the feasibility of storage of goods in different available locations corresponding to the latest verbal instructions. After normalizing the value, the final storage feasibility of goods in different available locations corresponding to the latest verbal instructions of the efficient digital storage method for warehouse goods based on NLP technology is obtained.
[0089] Step S7: When the storage feasibility is greater than the first threshold, the corresponding vacant locations can be used for goods storage.
[0090] Specifically, the threshold is set to determine the storage feasibility of the available locations. Make a judgment when the storage feasibility When , it is considered that the current free location can be used for goods storage. Otherwise, other storage operations are used to obtain the storage feasibility of the new storage location until the threshold condition is met or the storage operation with the largest feasibility is selected.
[0091] Furthermore, after step S7, the method further includes ranking the corresponding vacant locations whose storage feasibility is greater than the first threshold value from large to small according to the storage feasibility, and providing the ranking results to warehouse management personnel, who decide the storage location of the goods based on the storage feasibility ranking of the vacant locations and the actual warehouse storage situation.
[0092] In this way, in order to provide warehouse managers with more abundant storage options, while outputting the most feasible storage plan, multiple feasible second and third storage plans are also output for warehouse managers to decide the storage location of the goods based on the storage feasibility ranking of the vacant locations and the actual warehouse storage situation.
[0093] Through the above steps, the oral instructions of the warehouse manager are converted and analyzed, and finally the system generates a storage plan for the goods that currently need to be stored. The warehouse manager can judge the plan and choose whether to command the robotic arm, intelligent robot, etc. to store the goods according to the operation to complete the storage of the goods.
[0094] When the goods need to be taken out, the specific storage location of the goods is determined based on the initial storage records of the goods, and then the required goods are taken out using robotic arms, intelligent robots, etc. to complete the retrieval of the goods.
[0095] It should be noted that the more detailed the oral instructions of warehouse managers are, the more beneficial it is to the generation of warehouse strategies. Therefore, it is also necessary to standardize the content of oral instructions as much as possible to improve the stability of automatic generation of warehouse strategies.
[0096] The digital and efficient storage method for warehouse goods based on NLP technology of this embodiment builds a warehouse database according to the historical oral instructions of the warehouse management personnel and the feedback of the corresponding goods storage location, goods description information, warehouse storage rules, and real-time warehouse status; according to the similarity between the latest oral instructions of the warehouse management personnel and the historical oral instructions, the corresponding warehouse operations of similar historical oral instructions in the warehouse database are extracted to form a possible warehouse operation set of the latest oral instructions; according to the goods attributes extracted from the latest oral instructions and the goods attributes corresponding to the warehouse operations in the possible warehouse operation set, the similarity of the goods attributes is obtained; according to the similarity between the latest oral instructions and the historical oral instructions, the similarity of the goods attributes, and all available warehouse operations in the possible warehouse operation set of the latest oral instructions, the availability of the warehouse operation is obtained. In this way, by converting and analyzing the oral instructions of the warehouse management personnel, the originally ambiguous oral instructions are converted into natural language that the system can understand, thereby reducing the possibility of erroneous storage caused by ambiguity; in addition, based on the above-mentioned conversion and analysis of the oral instructions of the warehouse management personnel, the environmental similarity of different vacant locations is obtained according to the environmental parameters of the historical storage locations of the goods corresponding to the latest oral instructions in the warehouse database and the environmental parameters of all vacant locations in the possible storage operation set; the environmental qualification of different vacant locations is obtained according to the best storage environment of the goods corresponding to the latest oral instructions in the warehouse database and the environmental parameters of all vacant locations in the possible storage operation set; the storage feasibility of the goods corresponding to the latest oral instructions in different vacant locations is obtained according to the environmental qualification of different vacant locations, the environmental similarity of different vacant locations, and the availability of storage operations; the output of multiple reasonable storage plans is realized to improve the efficiency, stability and reliability of obtaining storage suggestions.
[0097] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0098] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.
Claims
1. A digital and efficient storage method for warehouse goods based on NLP technology, characterized in that: The method comprises: Step S1: Building a warehouse database based on the warehouse management personnel's historical verbal instructions and the corresponding feedback on the goods storage location, goods description information, warehouse storage rules, and real-time warehouse status; Step S2: according to the similarity between the latest oral instruction and the historical oral instruction of the warehouse manager, the corresponding warehouse operations of the similar historical oral instructions in the warehouse database are extracted to form a possible warehouse operation set of the latest oral instruction; Step S3: obtaining similarity of the cargo attributes according to the cargo attributes extracted by the latest verbal instruction and the cargo attributes corresponding to the storage operations in the possible storage operation set; Step S4: obtaining the availability of the storage operation according to the similarity between the latest verbal instruction and the historical verbal instructions, the similarity of the attributes of the goods, and all available storage operations in the possible storage operation set of the latest verbal instruction; Step S5: Obtain the environmental similarity and environmental qualification of different vacant locations according to the environmental parameters of the historical storage locations of the goods corresponding to the latest verbal instructions in the warehouse database, the optimal storage environment, and the environmental parameters of all vacant locations in the possible storage operation set; Step S6: According to the environmental similarity, environmental qualification and availability of storage operations of different available locations, the storage feasibility of the goods corresponding to the latest verbal instruction at different available locations is obtained; Step S7: When the storage feasibility is greater than the first threshold, the corresponding vacant location can be used for storage of goods; The similarity between the latest verbal instruction of the warehouse manager and the historical verbal instruction in step S2 is specifically: Decompose the historical oral instructions of warehouse managers into individual words or phrases to form a set of instruction words; Construct a bag-of-words model for the instruction word set and vectorize the bag-of-words model; The cosine similarity between the text vector of the latest verbal instruction and the text vector of the historical verbal instruction is calculated to obtain the similarity between the latest verbal instruction and the historical verbal instruction of the warehouse manager.
2. The method for digitally storing warehouse goods based on NLP technology according to claim 1 is characterized in that: In step S1: The cargo description information includes cargo name, cargo type, cargo size, delivery period, and order batch; The warehouse storage rules include safety specifications, cargo classification rules, and warehouse space layout requirements; The real-time warehouse status includes real-time temperature and humidity information of each area of the warehouse, real-time inventory status, and real-time cargo flow information.
3. The method for digitally storing warehouse goods based on NLP technology according to claim 1 is characterized in that: The corresponding storage operations of similar historical oral instructions extracted from the warehouse database in step S2 are specifically: When the cosine similarity between the text vector of the latest oral instruction and the text vector of the historical oral instruction is greater than a second threshold, the historical oral instruction is considered to be similar to the latest oral instruction, and the corresponding storage operation of the historical oral instruction is extracted.
4. The method for digitally storing warehouse goods based on NLP technology according to claim 3 is characterized in that: In step S3, the similarity of the goods attributes extracted according to the latest verbal instruction and the goods attributes corresponding to the storage operations in the possible storage operation set is obtained as follows: Extract the type of goods according to the instruction word information in the latest verbal instruction to form the basic attributes of the goods in the latest verbal instruction; The cosine similarity between the basic attributes of the goods in the latest verbal instruction and the attributes of the goods corresponding to the storage operations in the possible storage operation set is calculated to obtain the similarity of the attributes of the goods.
5. The method for digitally storing warehouse goods based on NLP technology according to claim 1 is characterized in that: In step S4, the greater the similarity between the latest oral instruction and the historical oral instruction, and the greater the similarity between the attributes of the goods, the greater the availability of the warehousing operation.
6. The method for digitally storing warehouse goods based on NLP technology according to claim 1 is characterized in that: In step S5, the smaller the absolute value of the difference between the environmental parameters of the historical storage location of the goods corresponding to the latest oral instruction in the warehouse database and the environmental parameters of all the vacant locations in the possible storage operation set, the higher the environmental similarity of the corresponding vacant location.
7. The method for digitally storing warehouse goods based on NLP technology according to claim 1 is characterized in that: In step S5, the smaller the absolute value of the difference between the optimal storage environment of the goods corresponding to the latest verbal instruction in the warehouse database and the environmental parameters of all the vacant locations in the possible storage operation set, the higher the environmental qualification of the corresponding vacant location.
8. The method for digitally storing warehouse goods based on NLP technology according to claim 1 is characterized in that: In step S6, when the environmental qualification of different vacant locations is higher, the environmental similarity of different vacant locations is higher, and the availability of storage operation is greater, the storage feasibility of the goods corresponding to the latest verbal instruction in different vacant locations is higher.
9. The method for digitally storing warehouse goods based on NLP technology according to claim 1 is characterized in that: The step S7 also includes ranking the corresponding vacant locations whose storage feasibility is greater than the first threshold from large to small according to the storage feasibility, and providing the ranking results to the warehouse management personnel, who decide the storage location of the goods based on the storage feasibility ranking of the vacant locations and the actual warehouse storage situation.
