A cloud platform-based manufacturing resource intelligent matching method and system

Abstract

The embodiment of the application discloses a kind of manufacturing resource intelligent matching method and system based on cloud platform, it is related to intelligent manufacturing technical field, can guarantee the matching speed and precision of manufacturing resource.The present application designs new matching algorithm, at least including 3 dimensions of matching, wherein the semantic similarity algorithm is used to calculate the semantic similarity of the first type of key features and the second type of key features, and the obtained semantic similarity is matched;The matching of the previous generation and the next generation type is calculated;After the sorting of the divided gradient is completed, for each gradient, the resource comprehensive function is used for matching, thereby improving the matching accuracy of manufacturing resources.And in the process of matching, the storage and query framework of the combination of relational database and non-relational database is used, and the query efficiency of resources is improved.Therefore, the embodiment can guarantee the matching speed and precision of manufacturing resource.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent manufacturing, and particularly relates to a manufacturing resource intelligent matching method and system based on a cloud platform. BACKGROUND

[0002] With the advent of the information age, a new networked manufacturing mode, i.e. cloud manufacturing mode, has been developed. The cloud manufacturing mode refers to virtualizing manufacturing resources and manufacturing capabilities scattered throughout the country to form a cloud manufacturing resource pool, so that the scattered manufacturing resources are integrated on a unified cloud platform, and then the resources in the cloud manufacturing resource pool are allocated according to the needs of users to realize the sharing of manufacturing resources. Due to the large amount of resources involved in the cloud manufacturing environment, the types and forms of the resources are various, and how to quickly and accurately match the service demand side with the service provider is a problem that needs to be solved at present. However, the traditional keyword matching cannot meet the personalized needs of users and the massive manufacturing resource data. Moreover, the resource service matching cannot be quantitatively calculated, and the user needs are in a dynamic and variable network environment, which seriously affects the efficiency and accuracy of the manufacturing resource service matching, and the industry also needs to improve the resource matching scheme in cloud manufacturing.

[0003] In some current schemes, the supply and demand relationship is used for matching, the keyword matching method is used to obtain a basic information preliminary selection set of resources, then the reasoning rules are used for matching to obtain a preselected set of resource function information, and finally the fuzzy comprehensive evaluation method is used for matching to obtain the optimal service resource. The design idea of this kind of resource matching scheme is actually an improvement based on the traditional data matching technology of the cloud platform, and is commonly used in online distribution of marketing resources, logistics resources and inventory resources, and is mainly applied in the e-commerce field. However, if applied to the actual field of intelligent manufacturing, there is a problem that the service needs of customers are inconsistent with the resource models in the database, resulting in the problems of missed search or low search accuracy of production resources.

[0004] There are also schemes that use user demand word vectors and manufacturing resource word vectors as inputs of a convolutional neural network for convolution training, and a result evaluation module judges the cosine similarity of user demand items and manufacturing resource items to allocate the optimal scheme for the user. However, this matching method has the problem that the manufacturing resource and user demand word vectors are completely independent in the modeling process and have no interaction until the abstract vector representation is generated, which makes the sentence lose a lot of semantic details in the abstract modeling process and also loses the opportunity for semantic interaction calculation between sentences, which also causes the problem of low matching accuracy.

[0005] Some schemes are set by a user or a third party to set a series of keyword tables, that is, the keyword table is set by manual, so as to clearly indicate the user demand. The keyword table setting has subjectivity, which affects the authenticity in the actual matching process, and too many human factors, whether the personnel are experienced or not will affect the setting of the keyword table and finally affect the matching precision. Moreover, due to too many human factors, the matching process is complicated, time-consuming and poor in real-time performance in actual execution.

[0006] Therefore, in the cloud manufacturing scenario, how to guarantee the matching speed and precision of the manufacturing resource becomes a problem to be studied. SUMMARY

[0007] Embodiments of the present application provide a cloud platform-based manufacturing resource intelligent matching method and system, which can guarantee the matching speed and precision of the manufacturing resource.

[0008] To achieve the above object, embodiments of the present application adopt the following technical solutions:

[0009] In a first aspect, the method provided by the embodiments of the present application comprises:

[0010] S1, the cloud platform receives a service request R sent by the user terminal, decomposes the information of the service request by a recursive decomposition algorithm and obtains n sub-tasks, wherein R=(R1, R2,..., Rn), n represents the number of sub-tasks, 1≤j≤n, Rj represents the manufacturing capability meeting the manufacturing demand of the jth sub-task j ,…,R n ), n represents the number of sub-tasks, 1≤j≤n, Rj represents the manufacturing capability meeting the manufacturing demand of the jth sub-task j

[0011] S2, according to the QoS index and weight specified by the user, an expression f(QoS) is constructed;

[0012] S3, the key features are extracted from the expression f(QoS), and the word groups representing the key features are determined;

[0013] S4, for each sub-task, according to the key features of the manufacturing demand of the sub-task, the manufacturing resource with the same key features is queried in the non-relational database, and the mapping between the key features of the manufacturing demand of the sub-task and the key features of the queried manufacturing resource is established;

[0014] S5, the matching value is calculated according to the established mapping relationship, and the manufacturing resource with the matching value greater than the threshold value is added to the candidate manufacturing resource set of the corresponding sub-task;

[0015] S6, the candidate manufacturing resource set is sent to the user terminal.

[0016] ​In a second aspect, embodiments of the present application provide a system, comprising: a cloud platform, a user terminal, a non-relational database (NoSQL) and a relational database (SQL); the cloud platform comprises a service demand analysis module, a function matching module and an optimization recommendation module;

[0017] The service demand analysis module is configured to receive a service request sent by the user terminal, and decompose the service request into n sub-tasks through a special representation method, where n represents the number of sub-tasks; then, according to the QoS indicators and weights specified by the user, an expression f(QoS) is constructed;

[0018] The service demand analysis module is configured to extract key features from the expression f(QoS), convert word groups representing the key features from a first format into a second format, where the first format is a format used to describe manufacturing capabilities of manufacturing demands, and the second format is a format used to describe manufacturing capabilities of manufacturing resources; and decompose manufacturing capabilities that meet the manufacturing demands of the sub-tasks into the key features; then, for each sub-task, query manufacturing resources with the same key features in the non-relational database according to the key features of the manufacturing demands of the sub-tasks, and establish a mapping between the key features of the manufacturing demands of the sub-tasks and the key features of the queried manufacturing resources; finally, calculate a matching value according to the established mapping relationship.

[0019] The optimization recommendation module is configured to add manufacturing resources with a matching value greater than a threshold to a candidate manufacturing resource set of the corresponding sub-task; and send the candidate manufacturing resource set to the user terminal.

[0020] The manufacturing resource intelligent matching method and system based on a cloud platform provided by the embodiments of the present application improve the matching precision of manufacturing resources by designing a new matching method, which includes at least three dimensions of matching links, wherein the semantic similarity algorithm is used to calculate the semantic similarity of the first type of key features and the second type of key features, and the obtained semantic similarity is used for matching; the matching of the previous generation and the next generation is calculated; after the sorting of the divided gradient is completed, the resource comprehensive function is used for matching for each gradient, thereby improving the matching precision of the manufacturing resources. In the matching process, the storage and query framework combining the relational database and the non-relational database is used, thereby improving the query efficiency of the resources. Therefore, the embodiments can guarantee the matching speed and precision of the manufacturing resources. BRIEF DESCRIPTION OF DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0022] Figure 1 A cloud manufacturing resource storage model schematic diagram provided for an embodiment of the present application.

[0023] Figure 2 A supply-demand matching strategy flow chart provided for an embodiment of the present application.

[0024] Figure 3 A manufacturing demand schematic diagram (Rj) provided for an embodiment of the present application.

[0025] Figure 4 A resource capability mapping relationship schematic diagram (Ai) provided for an embodiment of the present application.

[0026] Figure 5 A method flow schematic diagram provided for an embodiment of the present application. DETAILED DESCRIPTION

[0027] In order for those skilled in the art to better understand the technical solutions of the present application, the present application will be further described in detail below in combination with the drawings and specific embodiments. In the following, embodiments of the present application will be described in detail, and examples of the embodiments are shown in the drawings, in which the same or similar reference numerals represent the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by reference to the drawings are exemplary and are only used to explain the present application, and cannot be interpreted as a limitation on the present application. Those skilled in the art can understand that, unless specifically stated, the singular forms "a", "an" and "the" used herein can also include the plural forms. It should be further understood that the phrase "comprising" used in the specification of the present application means that the features, integers, steps, operations, elements and / or components exist, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof. It should be understood that when we say that an element is "connected" or "coupled" to another element, it can be directly connected or coupled to the other element, or there can be intermediate elements. In addition, "connected" or "coupled" used herein can include wireless connection or coupling. The phrase "and / or" used herein includes any one of the associated listed items and all combinations thereof. Those skilled in the art can understand that, unless otherwise defined, all terms (including technical terms and scientific terms) used herein have the same meaning as that generally understood by those skilled in the art in the field to which the present application belongs. It should also be understood that terms such as those defined in a general dictionary should be understood to have meanings consistent with those in the context of the prior art, and should not be interpreted with idealized or overly formal meanings unless defined as such.

[0028] In order to provide more accurate queries and supply-demand matching for users, as Figure 2As shown, the embodiment of the application adopts a comprehensive matching strategy for both supply and demand to meet the constraints of both supply and demand. The matching strategy process is divided into three parts: service demand analysis, function matching (or functional matching), and service recommendation (pushing the candidate manufacturing resource set to the user terminal). The embodiment of the application provides a cloud platform-based intelligent manufacturing resource matching method, as shown in Figure 5 The method of the embodiment includes the following steps:

[0029] S1, the cloud platform receives the service request R sent by the user terminal, and decomposes the information of the service request by recursive decomposition algorithm and obtains n sub-tasks.

[0030] Among them, the service request submitted by the customer (denoted as R) can be analyzed by a special representation method, and decomposed into multiple sub-tasks by the special representation method, denoted as R=(R1, R2, …, Rn). j ,…,R n n represents the number of sub-tasks, 1≤j≤n, R j represents the manufacturing capacity that meets the manufacturing demand of the jth sub-task, the information of the service request includes the type of processing manufacturing task, the number of products that need to be processed, the delivery time of the processing manufacturing task, etc. The sub-task attribute related information obtained by decomposing the information of the service request by recursive decomposition algorithm includes the number of processing manufacturing master task, sub-task number, sub-task processing process requirement, sub-task product quantity, sub-task delivery time, etc. Among them, the sub-task attribute related information is a further decomposition and refinement on the information of the service request, for example: the information of the service request includes “the type of processing manufacturing task”, and “the type of processing manufacturing task” records that “electroplating processing” is needed for two surfaces of metal parts respectively, at least two sub-tasks can be obtained by decomposition, namely sub-task 1 and sub-task 2, wherein the attribute related information of sub-task 1 obtained by decomposing the information of the service request includes the number of processing manufacturing master task “electroplating 1”, the sub-task number “electroplating 1-surface 1”, and the sub-task processing process requirement “electroplating time at least 20 minutes”.

[0031] S2, according to the QoS index and weight specified by the user, an expression f(QoS) is constructed;

[0032] Among them, the QoS index and weight specified by the user are also submitted to the cloud platform by the user operating the user terminal, and the specific information of the QoS index and weight specified by the user can be submitted together in the service request R or separately.

[0033] The integrated QoS indicator parameters in the f(QoS) expression include at least service time T, service cost C, processing capacity P, information feedback capacity I, delivery accuracy A, and after-sales service capacity S. These QoS indicator parameters can be used as the key features to be extracted in S3. In the f(QoS) expression, each QoS indicator parameter is multiplied by a corresponding weight and added together, for example: f(QoS) = service time T * weight T' + service cost C * weight C' + processing capacity P * weight P' + information feedback capacity I * weight I' + delivery accuracy A * weight A' + after-sales service capacity S * weight S'.

[0034] S3, extracting key features from the f(QoS) expression and determining a phrase representing the key features. In the process of determining the phrase representing the key features, format conversion can be performed, such as converting from a first format to a second format, the first format being a format for describing the manufacturing capacity of the manufacturing demand, and the second format being a format for describing the manufacturing capacity of the manufacturing resource. Generally, the key features extracted from the f(QoS) expression are the original words, which are in the first format and are also used to describe the manufacturing capacity of the manufacturing demand. The phrase representing the key features can also be referred to as "keywords", which can be extracted from the key features using the TextRank algorithm.

[0035] After that, the information field of the manufacturing capacity that meets the manufacturing demand of the subtask can be described and vector extracted through word vector modeling, so as to complete the description of the key features of the manufacturing capacity that meets the manufacturing demand of the subtask. The manufacturing demand of the subtask can be further decomposed into a plurality of key features of the manufacturing capacity, which is represented as R j =(r j1 ,r j2 ,…r ji ,…,r jm ), where m represents the number of key features of the manufacturing capacity of the subtask R j . R j is the manufacturing capacity for meeting the manufacturing demand, represented as R j =(r j1 ,r j2 ,…r ji ,…,r jm ), where m represents the number of key features required to meet the manufacturing demand. For example, when m = 4, R j =(r j1 ,r j2 ,r j3 ,r j4 ).

[0036] S4. For each subtask, based on the key features of the manufacturing requirements of the subtask, query the non-relational database for manufacturing resources with the same key features, and establish a mapping between the key features of the manufacturing requirements of the subtask and the key features of the queried manufacturing resources.

[0037] One approach is to use functional semantic mapping, specifically by employing a semantic approximation method to semantically map the manufacturing capabilities of each key feature to different manufacturing resources on different views in a non-relational database.

[0038] S5. Calculate the matching value based on the established mapping relationship, and add the manufacturing resources with matching values ​​greater than the threshold to the candidate manufacturing resource set of the corresponding subtask;

[0039] Specifically, based on the mapping relationship and the granularity level of the manufacturing resource, manufacturing resources with a functional matching value greater than a threshold are placed in the corresponding candidate service queue. It should be noted that there are at least three types of "matching values" appearing in this embodiment. This embodiment designs multiple matching methods, specifically including: 1. Calculating the semantic similarity between the first type of key features and the second type of key features using a semantic similarity algorithm, and using the obtained semantic similarity as the matching value, where the obtained semantic similarity matching value is denoted as S. j =(s j1 ,s j2 ,…s ji ,…,s jm ); 2. By calculating the previous generation and p jx The matching value of the descendant type is represented as PMV(p jk ,a iv 3. After sorting the divided gradients, for each gradient, calculate the matching value using the resource synthesis function, and then apply the matching value FMV(R) calculated by the resource synthesis function. j A i ).

[0040] S6. Send the candidate manufacturing resource set to the user terminal.

[0041] In this embodiment, as Figure 1 As shown, the information of manufacturing resources recorded in the cloud platform is stored in the view of the non-relational database;

[0042] The information stored in the relational database at least includes resource basic information, supplier information, input and output information and resource function information; wherein, the resource basic information includes resource name, resource number, equipment type, equipment model, etc.; the supplier information includes manufacturing and processing capacity, personnel configuration, technical support, reliability, price, etc.; the input is user demand, and the output is demand and resource matching result; the resource function information includes state attribute, equipment parameter, processing performance, technical performance, etc.

[0043] The information stored in the non-relational database at least includes Qos information and identity information, and can further include function information. Wherein, the Qos information includes information for indicating service time T, information for indicating service cost C, information for indicating processing capacity P, information for indicating information feedback capacity I, information for indicating delivery accuracy A, and information for indicating after-sales service capacity S; these Qos information can be displayed in plain code, such as the information for indicating service time T is the string of service time T itself, or can be displayed in encrypted code, such as the information for indicating service time T is an encrypted string after encrypting the string of service time T, which needs to be decrypted and restored to plain code in actual use.

[0044] The information of the manufacturing resource entered into the cloud platform includes basic information, state attribute (machine tool idle, full load, etc.), equipment parameter (equipment size, spindle speed, equipment maximum load, etc.), processing performance (processing precision, positioning accuracy, processable material, etc.), technical performance (numerical control system, operation mode, communication mode), performance evaluation (service time, credit, service quality, etc.), and these resource-related information can be stored in the non-relational database.

[0045] In order to make the query accurate and comprehensive, the resource mapping based on semantic information is carried out in the embodiment. According to the mapping result, the function matching value (FMV) is calculated through the hierarchical granularity difference between the function attribute provided by the customer and the function attribute of the manufacturing resource in different views. The FMV is sorted in descending order, and then the manufacturing resource with FMV greater than the threshold value is added to the candidate service set. Specifically, in S4, the mapping between the key features of the manufacturing demand of the subtask and the key features of the manufacturing resource obtained by query is established, including:

[0046] S41, querying the view in which the required key characteristics of R j are located in the non-relational database;

[0047] S42, for the view obtained by query in S41, when there are two key features with the same parent node in the non-relational database, it is determined that the two key features belong to the same view, and N v (Rj ) views, wherein N v (R j ) represents the total number of views in which the required key characteristics of R j

[0048] S43, the mapping is established in the following way: establishing a set of manufacturing resource candidates U j that meet the matching conditions of R i and initializing, wherein U i = u i1 ∪u i2 ∪u i3 …∪u im , u im represents a set of resources that meet the mth view search condition, after initialization u i1 , u i2 , u i3 , … u im is empty.

[0049] Specifically, in S5, the matching value is calculated according to the established mapping relationship, including:

[0050] S51, the semantic similarity between the first type of key features and the second type of key features is calculated by a semantic similarity algorithm, and the obtained semantic similarity is taken as the matching value, wherein the first type of key features includes key features decomposed from R j , and the second type of key features includes key features decomposed from the manufacturing capability of the manufacturing resource. Specifically, the manufacturing resource and the user demand can be described by feature and vector extraction on the cloud platform through word vector modeling, and the convolutional neural network (CNN) is used to map the manufacturing resource and the user demand word vector to a common space with vector matching basis, so as to construct a target function by matching the distance between the two groups of word vectors, and calculate the semantic similarity.

[0051] S52, the key features decomposed from R j are represented as elements in R j = (r j1 , r j2 , … r ji , …, r jm ), and m represents the number of key features that meet the manufacturing demand of the subtask, 1≤i≤m.

[0052] S53, the key features decomposed from the manufacturing capability of the manufacturing resource are represented as elements in P j = (p j1 , p j2 , … p ji , …, p jm ), and P j ​manufacturing capability of the manufacturing resource corresponding to the jth subtask, p jm the mth feature representing the manufacturing capability of the manufacturing resource of the jth subtask, the matching value of the obtained semantic similarity is represented as S j = (s j1 ,s j2 ,…s ji ,…,s jm ), wherein s jm represents the matching value of r jm and p jm .

[0053] Further, in S5, the manufacturing resource with the matching value greater than the threshold value is added to the candidate service queue, including: reading the matching threshold a of the semantic similarity, and comparing S j and the threshold value a, wherein if s jx >a, the manufacturing resource individual of the type p jx is added to u ix , and the matching value of the previous generation and the descendant type of p jx is calculated, wherein s jx ∈S j , 1≤x≤m, a∈(0,1). The manufacturing resources are sorted and added to the candidate service queue by using the matching value of the previous generation and the descendant type of p jx .

[0054] For example: in this embodiment, the matching threshold a of the semantic similarity is defined, a∈(0,1). The value of a is inversely proportional to the number of matched manufacturing resources, the greater the value of a, the fewer the matched manufacturing resources, and the higher the accuracy of the mapping result. The smaller the value of a, the more the matched manufacturing resources, and the higher the recall rate of the mapping result. Then s j and the threshold value a can be compared. If s j1 >a, all manufacturing resource individuals of the type p j1 are added to u i1 . At the same time, the matching value of the previous generation and the descendant type of p j1 is calculated. If the matching value of the type is greater than the threshold value a, all manufacturing resources under the type are added to u i1 . Specifically, the matching value of the previous generation and the descendant type of p j1 is calculated, including: wherein PMV(p jk ,a iv ) represents the matching value of the type of p jk and the type of a iv , a iv represents the previous or next type of the type of p jk , p jkN represents the k-th feature of the manufacturing capability of the manufacturing resources for the j-th subtask. H (p jk ,a iv ) represents p jk Type and a iv The granularity difference between types. For example... Figure 3 As shown, N H (p j1 ,a i1 ) = 2, N H (p j3 ,a i3 ) = 0, N H (p j4 ,a i5 =3.

[0055] The ranking of manufacturing resources includes two parts: establishing a gradient and ranking resources within the gradient. Specifically, using the established mapping, manufacturing resources are divided into gradients based on the number of key features matching them. If a key feature matching a manufacturing resource appears in all elements of Ui, the manufacturing resource is ranked in the first gradient; otherwise, it is ranked in the second gradient. i In the my elements, the manufacturing resource is ranked in the (y+1)th gradient, 0≤y≤m-1; the sorting order of the divided gradients is in ascending order of gradient number, and the sorting result is added to the candidate manufacturing resource set of the subtask.

[0056] After sorting the identified gradients, for each gradient, a matching value is calculated using the resource synthesis function, and then the matching value FMV(R) is calculated according to the resource synthesis function. j A i ), then according to FMV(R j A i Sort the resources in descending order and add the sorted results to the candidate manufacturing resource set of the subtask; where, A i =(a i1 a i2 a i3 ,…,a im ) represents the resource candidate set sorted in descending order.

[0057] For example, in this embodiment, manufacturing resources are divided into different gradients based on the number of key features that match the resources. For instance, if a manufacturing resource appears in u i1 u i2 u i3 and u i4 In the middle, these resources are ranked in the first gradient. If the resource is in u i1 ui2 , u i3 and u i4 Any three of which occur, the resource is ranked in the second gradient, and so on. After that, the gradients are sorted in ascending order. Within the gradient, the matching values calculated by the resource comprehensive function are sorted in descending order. The calculation formula is as follows:

[0058]

[0059] Wherein, A i = (a i1 , a i2 , a i3 , …, a im ) represents the manufacturing capacity of the manufacturing resource. And R j = (r j1 , r j2 , … r ji , …, r jm ) represents the manufacturing capacity of the manufacturing demand. Finally, the result of the comprehensive sorting is added to the candidate manufacturing resource set of the subtask for combined optimization and recommendation to the customer.

[0060] The embodiment also provides a cloud platform-based intelligent manufacturing resource matching system, comprising a cloud platform, a user terminal, a non-relational database (NoSQL) and a relational database (SQL); the cloud platform comprises a service demand analysis module, a function matching module and an optimization recommendation module;

[0061] The service demand analysis module is configured to receive a service request sent by the user terminal, and decompose the service request into n subtasks through a special representation method, where n represents the number of subtasks; then, according to the QoS indicators and weights specified by the user, an expression f(QoS) is constructed;

[0062] The service demand analysis module is configured to extract key features from the expression f(QoS), convert a phrase representing the key features from a first format into a second format, where the first format is a format used to describe the manufacturing capacity of the manufacturing demand, and the second format is a format used to describe the manufacturing capacity of the manufacturing resource; and decompose the manufacturing capacity of the manufacturing demand of the subtask into the key features; then, for each subtask, according to the key features of the manufacturing demand of the subtask, query the non-relational database for manufacturing resources with the same key features, and establish a mapping between the key features of the manufacturing demand of the subtask and the key features of the queried manufacturing resources; finally, calculate a matching value according to the established mapping relationship;

[0063] The optimization recommendation module is configured to add the manufacturing resources with the matching value greater than a threshold to a candidate manufacturing resource set of the corresponding subtask; and send the candidate manufacturing resource set to the user terminal.

[0064] Specifically, the optimization recommendation module is specifically configured to divide the manufacturing resources into gradients according to the number of key features matched with the manufacturing resources by using the established mapping, then calculate the matching values by the resource comprehensive function for each gradient, and sort the matching values calculated by the resource comprehensive function in descending order, and finally add the sorting results to the candidate manufacturing resource set of the subtask.

[0065] The embodiment proposes a manufacturing information storage architecture combining a relational database and a non-relational database, and improves the query efficiency of resources. The storage and query framework combining the relational database and the non-relational database is used, so that the management and maintenance of information and the query efficiency are good. The query efficiency of resources is improved. In view of the problem of low query accuracy caused by the inconsistency between the service demand of a customer and the resource model in the database, a resource mapping matching algorithm based on view and granularity is proposed. A new matching algorithm is proposed based on the storage model, and the accuracy and recall rate of matching query are improved.

[0066] The manufacturing resource intelligent matching method and system based on the cloud platform provided by the embodiment of the application place the manufacturing resources with a function matching value greater than a threshold in a corresponding candidate service queue according to a mapping relationship and a granularity layer where the manufacturing resources are located. A new matching algorithm is designed in the scheme, which includes at least three dimensions of matching. The semantic similarity between the first type of key features and the second type of key features is calculated by a semantic similarity algorithm, and the obtained semantic similarity is matched. The matching between the previous generation and the next generation is calculated. After the sorting of the divided gradients is completed, the resource comprehensive function is used for matching for each gradient, so that the matching accuracy of the manufacturing resources is improved. In the matching process, the storage and query framework combining the relational database and the non-relational database is used, so that the query efficiency of resources is improved. Therefore, the embodiment can guarantee the matching speed and accuracy of manufacturing resources.

[0067] Each of the embodiments in the specification is described in a progressive manner, and the same or similar parts of each of the embodiments can be referred to each other. Each of the embodiments focuses on the difference from other embodiments. Especially, the device embodiment is described relatively simply because it is basically similar to the method embodiment, and the relevant parts can be referred to the part of the description of the method embodiment. The above is merely a specific implementation of the application, but the protection scope of the application is not limited thereto. Any skilled person in the art can easily think of changes or replacements within the technical range disclosed by the application, which should be covered in the protection scope of the application. Therefore, the protection scope of the application should be subject to the protection scope of the claims.

Claims

1. A cloud-based intelligent matching method for manufacturing resources, characterized in that, include: S1. The cloud platform receives a service request R from a user terminal, and decomposes the service request information into n subtasks using a recursive decomposition algorithm, where R = (R1, R2, ..., Rn). j ,…,R n ), where n represents the number of subtasks, 1≤j≤n, R j This represents the manufacturing capacity required to meet the manufacturing needs of the j-th subtask. S2. Construct the f(QoS) expression based on the QoS metrics and weights specified by the user; S3. Extract key features from the f(QoS) expression and determine the word groups representing the key features; S4. For each subtask, based on the key features of the manufacturing requirements of the subtask, query the non-relational database for manufacturing resources with the same key features, and establish a mapping between the key features of the manufacturing requirements of the subtask and the key features of the queried manufacturing resources. S5. Calculate the matching value based on the established mapping relationship, and add the manufacturing resources with matching values ​​greater than the threshold to the candidate manufacturing resource set of the corresponding subtask; S6. Send the candidate manufacturing resource set to the user terminal; Information about manufacturing resources entered into the cloud platform is stored in a view of the non-relational database; The information stored in a relational database includes at least: basic resource information, supplier information, input / output information, and resource function information; The information stored in the non-relational database includes at least QoS information and identity information; In S4, a mapping is established between the key features of the manufacturing requirements of the subtask and the key features of the queried manufacturing resources, including: S41. Query the non-relational database to find the condition R. j The view containing the key features required; S42. For the view obtained in S41, if two key features have the same parent node in the non-relational database, then these two key features are determined to belong to the same view, and N is ultimately obtained. v (R j ) views, where N v (R j ) indicates that R is satisfied j The total number of views containing the required key features; S43. Methods for establishing mappings include: establishing mappings that satisfy R. j Candidate set U of manufacturing resources matching the conditions i And initialize it, where U i =u i1 ∪u i2 ∪u i3 …∪u im u im This represents the set of resources that satisfy the search criteria for the m-th view; In S5, the matching value is calculated based on the established mapping relationship, including: A semantic similarity algorithm is used to calculate the semantic similarity between a first type of key feature and a second type of key feature, and the obtained semantic similarity is used as the matching value. The first type of key feature includes: from R... j The key features decomposed from the manufacturing resources, the second category of key features includes: key features decomposed from the manufacturing capabilities of the manufacturing resources; From R j The key features extracted from the decomposition are represented as: R j =(r j1 ,r j2 ,…r ji ,…,r jm In the subtask, m represents the number of key features that satisfy the manufacturing requirements of the subtask, 1≤i≤m; The key feature derived from the manufacturing capacity of the manufacturing resources is represented as: P j = (p j1 ,p j2 ,…p ji ,…,p jm The elements in ) P j The manufacturing capacity of the manufacturing resources corresponding to the j-th subtask. The obtained semantic similarity matching value is represented as S. j =(s j1 ,s j2 ,…s ji ,…,s jm ), where s jm Indicates r jm With p jm The matching value.

2. The method according to claim 1, characterized in that, In S5, adding manufacturing resources with matching values ​​greater than a threshold to the candidate service queue includes: Read the semantic similarity matching threshold 'a' and compare it with S. j and threshold a, where if s jx >a, then p jx Type of manufacturing resource individual join u ix In, among which, s jx ∈S j , 1≤x≤m, a∈(0,1).

3. The method according to claim 2, characterized in that, Sort the manufacturing resources, including: Using the established mapping, manufacturing resources are divided into gradients based on the number of key features that match them. If a key feature that matches a manufacturing resource appears in all elements of Ui, then the manufacturing resource is ranked in the first gradient.

4. The method according to claim 2, characterized in that, Also includes: If the key characteristics that match manufacturing resources appear in U i In the my elements, the manufacturing resources are ranked in the (y+1)th gradient, 0≤y≤m-1; The gradients are sorted in ascending order of gradient number, and the sorting results are added to the candidate manufacturing resource set of the subtask.

5. The method according to claim 3, characterized in that, The sorting of the manufacturing resources also includes: After sorting the identified gradients, for each gradient, a matching value is calculated using the resource synthesis function, and then the matching value is determined according to the resource synthesis function. Then according to Sort the resources in descending order and add the sorted results to the candidate manufacturing resource set of the subtask; in, A i =( a i1 a i2 a i3 ,…,a im ) represents the resource candidate set sorted in descending order.

6. A cloud-based intelligent matching system for manufacturing resources, employing the intelligent matching method described in claim 1, characterized in that, include: The cloud platform comprises a cloud platform, a user terminal, a NoSQL database, and a SQL database; the cloud platform includes a service requirement analysis module, a function matching module, and an optimization recommendation module. The service demand analysis module is used to receive service requests sent by the user terminal, and decompose the service requests into n sub-tasks using a special representation method, where n represents the number of sub-tasks; then, it constructs an expression f(QoS) based on the QoS indicators and weights specified by the user. The service requirement analysis module is used to extract key features from the f(QoS) expression, convert the phrases representing key features from a first format to a second format, wherein the first format is a format for describing manufacturing capabilities for manufacturing requirements, and the second format is a format for describing manufacturing capabilities for manufacturing resources; and decompose the manufacturing capabilities that meet the manufacturing requirements of sub-tasks into key features. Then, for each subtask, based on the key features of the manufacturing requirements of the subtask, the manufacturing resources with the same key features are queried in the non-relational database, and a mapping is established between the key features of the manufacturing requirements of the subtask and the key features of the queried manufacturing resources; finally, the matching value is calculated based on the established mapping relationship. The optimization recommendation module is used to add manufacturing resources with matching values ​​greater than a threshold to the candidate manufacturing resource set of the corresponding sub-task; and send the candidate manufacturing resource set to the user terminal.

7. The system according to claim 6, characterized in that, The optimization recommendation module is specifically used to divide manufacturing resources into gradients based on the number of key features that match the manufacturing resources, using the established mapping. Then, for each gradient, the matching value is calculated using the resource synthesis function, and the matching values ​​are sorted in descending order according to the resource synthesis function; finally, the sorting results are added to the candidate manufacturing resource set of the subtask.

Images