Supplier cooperation scheme creation method and device based on industrial ecology
By using the SGHAN model in the industrial ecosystem, combining LSTM, Softmax functions and GATs, dynamic interest encoding and attention encoder are constructed, which solves the problem that target users are difficult to meet, and the accurate and efficient creation of supplier collaboration solutions is achieved.
Patent Information
- Application Number
- CN202410980444.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-22
- Publication Date
- 2024-11-08
AI Technical Summary
The existing technology is difficult to meet the dynamic needs of target users, and it is difficult to accurately establish a model of the dynamic needs of target users, and the process of dynamically creating supplier collaboration solutions is complicated.
Using a supplier collaborative solution creation method based on the industrial ecosystem, using the pre-constructed SGHAN model, combined with LSTM, Softmax functions and GATs, a dynamic interest encoding, attention impact encoder and graph attention network are constructed to generate supply suitable for target users. Partnership collaboration solutions.
By learning the similar behavior of target users in selecting suppliers and helping partners, we can accurately and dynamically generate target users' supplier collaboration solutions to meet dynamic needs and simplify the creation process of supplier collaboration solutions.
Smart Images

Figure CN118916544A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of computer technology, and in particular to a method and device for creating a supplier collaboration solution based on an industrial ecosystem. Background Art
[0002] The concept of industrial ecology comes from the ecosystem, which can automatically solve complex dynamic problems and integrate the ecosystem. However, with the application of social attributes and functions on more and more industrial information platforms, many studies have revealed that an enterprise is easily influenced by its partners when choosing upstream and downstream suppliers to provide products or services, which in turn provides new possibilities for the platform to recommend suppliers to the enterprise users or generate an overall service solution (referred to as supplier collaboration solution) that is collaborative and adapted to the interests of the enterprise users and includes multiple suppliers.
[0003] In the related technology, the performance of the module instance can be scored according to the module type of the smart product service system, and then the upper and lower optimization configuration models can be established with the module instance and the supplier of the module instance as the decision variables. The two optimization configuration models can be solved to obtain the optimal supplier configuration plan; it is also possible to solve the resource allocation problem by building a computing offloading model in a blockchain-based mobile edge computing system.
[0004] However, the related technologies cannot meet the dynamic needs of target users. It is difficult to accurately establish a model of the dynamic needs of target users. The process of dynamically creating supplier collaboration solutions is relatively complicated and urgently needs to be improved. Summary of the invention
[0005] The present application provides a method and device for creating a supplier collaboration solution based on an industrial ecosystem to solve problems in related technologies such as the inability to meet the dynamic needs of target users, the difficulty in accurately establishing a model of the dynamic needs of target users, and the complexity of the process of dynamically creating supplier collaboration solutions.
[0006] The first aspect of the present application provides a method for creating a supplier collaboration plan based on an industrial ecosystem, comprising the following steps: obtaining target user data of a target user in the industrial ecosystem and partner data of at least one partner corresponding to the target user; inputting the target user data and the partner data into a pre-built SGHAN (Social-powered Graph Hierarchical Attention Network) model to obtain candidate suppliers for the target user; and generating a supplier collaboration plan suitable for the target user based on the candidate suppliers.
[0007] Optionally, in one embodiment of the present application, before inputting the target user data and the partner data into the pre-built SGHAN model, it also includes: establishing a first RNNs (Recurrent Neural Networks) based on LSTM (Long-Short Term Memory), and using the first RNNs to construct a dynamic interest code; establishing a second RNNs based on the LSTM and Softmax function, and using the second RNNs to construct an attention influence encoder; using GATs (Graph Attention Networks) and the Softmax function to construct a graph attention network; constructing the SGHAN model based on the dynamic interest code, the attention influence encoder and the graph attention network.
[0008] Optionally, in one embodiment of the present application, the inputting the target user data and the partner data into the pre-built SGHAN model includes: based on the target user data, using the dynamic interest encoding to obtain the initial interest representation of the target user; based on the partner data and the initial interest representation, using the attention influence encoder to obtain at least one initial preference representation of the target user under the influence of the at least one partner; according to the initial interest representation, using the graph attention network to assign weights to the at least one initial preference representation, so as to obtain the preference representation of the target user under the influence of the at least one partner based on at least one initial preference representation with different weights; and predicting the initial candidate suppliers of the target user based on the initial interest representation and the preference representation.
[0009] Optionally, in one embodiment of the present application, before obtaining the candidate suppliers of the target user, it also includes: based on the initial candidate suppliers, solving the candidate probability of the initial candidate suppliers using a preset probability formula; based on the candidate probability, judging whether the initial candidate suppliers meet the preset candidate conditions using a loss function; if the initial candidate suppliers meet the preset candidate conditions, using the initial candidate suppliers as the candidate suppliers of the target user.
[0010] Optionally, in one embodiment of the present application, the expression of the loss function may be, but is not limited to,:
[0011]
[0012] Where w represents the number of suppliers in the supplier collaboration solution.
[0013] The second aspect of the present application provides a supplier collaboration solution creation device based on an industrial ecosystem, including: an acquisition module, used to acquire target user data of a target user in the industrial ecosystem and partner data of at least one partner corresponding to the target user; an input module, used to input the target user data and the partner data into a pre-built SGHAN model to obtain candidate suppliers for the target user; and a first generation module, used to generate a supplier collaboration solution suitable for the target user based on the candidate suppliers.
[0014] Optionally, in one embodiment of the present application, it also includes: a first building module, used to establish a first RNNs based on LSTM before inputting the target user data and the partner data into the pre-built SGHAN model, and use the first RNNs to construct dynamic interest coding; a second building module, used to establish a second RNNs based on the LSTM and Softmax function, and use the second RNNs to construct an attention influence encoder; a third building module, used to construct a graph attention network using GATs and the Softmax function; a fourth building module, used to construct the SGHAN model based on the dynamic interest coding, the attention influence encoder and the graph attention network.
[0015] Optionally, in one embodiment of the present application, the input module includes: a first generation unit, used to obtain the initial interest representation of the target user based on the target user data using the dynamic interest encoding; a second generation unit, used to obtain at least one initial preference representation of the target user under the influence of the at least one partner based on the partner data and the initial interest representation using the attention influence encoder; a third generation unit, used to assign weights to the at least one initial preference representation based on the initial interest representation using the graph attention network, so as to obtain the preference representation of the target user under the influence of the at least one partner based on at least one initial preference representation with different weights; and a prediction unit, used to predict the initial candidate suppliers of the target user based on the initial interest representation and the preference representation.
[0016] Optionally, in one embodiment of the present application, it also includes: a solution module, which is used to solve the candidate probability of the initial candidate supplier based on the initial candidate supplier using a preset probability formula before obtaining the candidate supplier of the target user; a judgment module, which is used to judge whether the initial candidate supplier meets the preset candidate condition based on the candidate probability using a loss function; and a second generation module, which is used to use the initial candidate supplier as the candidate supplier of the target user when the initial candidate supplier meets the preset candidate condition.
[0017] Optionally, in one embodiment of the present application, the expression of the loss function may be, but is not limited to,:
[0018]
[0019] Where w represents the number of suppliers in the supplier collaboration solution.
[0020] The third aspect of the present application provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for creating a supplier collaboration solution based on industrial ecology as described in the above embodiment.
[0021] The fourth aspect of the present application provides a computer-readable storage medium, which stores a computer program. When the program is executed by a processor, it implements the above-mentioned method for creating a supplier collaboration solution based on industrial ecology.
[0022] The fifth aspect of the present application provides a computer program product, including a computer program, which, when executed, implements the above-mentioned method for creating a supplier collaboration solution based on industrial ecology.
[0023] The embodiment of the present application can dynamically generate a supplier collaboration plan suitable for the target user based on the pre-built SGHAN model. As a deep learning model, the SGHAN model can learn similar behaviors from partners that are helpful to the target user in selecting suppliers during the supplier collaboration plan creation process, so as to meet the dynamic needs of the target user and dynamically generate the supplier collaboration plan for the target user more accurately. Thus, the problems in the related art that the dynamic needs of the target user cannot be met, it is difficult to accurately establish a model of the dynamic needs of the target user, and the process of dynamically creating a supplier collaboration plan is more complicated are solved.
[0024] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS
[0025] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:
[0026] Figure 1 A flowchart of a method for creating a supplier collaboration solution based on an industrial ecosystem provided according to an embodiment of the present application;
[0027] Figure 2 A schematic diagram of the structure of the SGHAN model provided according to an embodiment of the present application;
[0028] Figure 3 A block diagram of an SGHAN model execution process provided according to an embodiment of the present application;
[0029] Figure 4 A flowchart of the working principle of a method for creating a supplier collaboration solution based on an industrial ecosystem provided according to an embodiment of the present application;
[0030] Figure 5 A schematic block diagram of a device for creating a supplier collaboration solution based on an industrial ecosystem according to an embodiment of the present application;
[0031] Figure 6 It is a schematic diagram of the structure of an electronic device provided according to an embodiment of the present application. DETAILED DESCRIPTION
[0032] Embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.
[0033] The following describes the supplier collaboration solution creation method and device based on the industrial ecology of the embodiment of the present application with reference to the accompanying drawings. In view of the problems mentioned in the above background technology that the dynamic needs of target users cannot be met, it is difficult to accurately establish a model of the dynamic needs of target users, and the process of dynamically creating supplier collaboration solutions is relatively complicated, the present application provides a supplier collaboration solution creation method based on the industrial ecology. In this method, a supplier collaboration solution suitable for the target user can be dynamically generated based on the pre-built SGHAN model. As a deep learning model, the SGHAN model can learn similar behaviors from partners who are helpful to the target user in selecting suppliers during the supplier collaboration solution creation process, so as to meet the dynamic needs of the target user and more accurately dynamically generate the supplier collaboration solution for the target user. As a result, the problems in the related technology that the dynamic needs of the target user cannot be met, it is difficult to accurately establish a model of the dynamic needs of the target user, and the process of dynamically creating supplier collaboration solutions is relatively complicated are solved.
[0034] Specifically, Figure 1 A flowchart of a method for creating a supplier collaboration solution based on an industrial ecosystem provided according to an embodiment of the present application.
[0035] like Figure 1 As shown, the method for creating a supplier collaboration solution based on industrial ecology includes the following steps:
[0036] In step S101, target user data of a target user in the industrial ecology and partner data of at least one partner corresponding to the target user are obtained.
[0037] It can be understood that in the embodiments of the present application, the target user data may include but is not limited to a target user set, a corresponding supplier set, and a supplier collaboration solution set, etc., and the present application does not impose any specific limitations.
[0038] The target user set of the embodiment of the present application can be represented by U={u1,u2,...,u M}, the supplier set can be represented by S = {s1,s2,...,s N} represents that G = (U, V) represents a static relationship graph, where It is the edge in the graph, which represents the ecological relationship between the target users of the industrial ecology.
[0039] Furthermore, in the embodiment of the present application, the behavior history of the target user u∈U is divided into a series of supplier collaboration schemes in chronological order: The t-th supplier collaboration plan for target user u Represents a chronological list of suppliers in, represents the pth supplier included in the tth supplier collaboration scheme of target user u, and n is the number of suppliers included in the tth supplier collaboration scheme of target user u.
[0040] The goal of creating a supplier collaboration solution in the embodiment of the present application is to provide a supplier collaboration solution for a given target user u. Then recommend a set of suppliers that the target user u may be interested in at the next time step n+1 from S, that is, These suppliers can be used together with the suppliers that have been used previously as a new supplier collaboration solution to jointly meet the needs of target users.
[0041] In addition, the partner data in the embodiment of the present application may include, but is not limited to, a partner set and a corresponding supplier collaboration solution set, etc., and the present application does not impose any specific limitations.
[0042] As a possible implementation method, the embodiment of the present application can obtain target user data of a target user in the industrial ecosystem and partner data of at least one corresponding partner.
[0043] Optionally, in one embodiment of the present application, before inputting the target user data and partner data into the pre-built SGHAN model, it also includes: establishing a first RNNs based on LSTM, and using the first RNNs to construct a dynamic interest encoding; establishing a second RNNs based on LSTM and Softmax function, and using the second RNNs to construct an attention influence encoder; using GATs and Softmax functions to construct a graph attention network; and constructing an SGHAN model based on dynamic interest encoding, attention influence encoder and graph attention network.
[0044] It can be understood by those skilled in the art that the SGHAN model in the embodiment of the present application includes three core modules, such as Figure 2 As shown, it can be but not limited to: dynamic interest coding, attention influence encoder and graph attention network, etc., and this application does not make specific restrictions, so as to achieve the creation of accurate dynamic supplier collaboration solutions for target users.
[0045] Among them, in some embodiments, the dynamic interest coding of the embodiment of the present application utilizes the first RNNs to simulate the supplier selection process of the target user in the target user data, such as the current supplier collaboration scheme, so as to extract the initial interest of the target user. Among them, the initial interest of the target user can be defined as the preference reflected by the target user in the current supplier collaboration scheme, that is, the current preference after selecting a series of suppliers, which plays a key role in the next supplier selected by him.
[0046] Furthermore, in order to better process long sequence data, the dynamic interest encoding in the embodiment of the present application can adopt the standard LSTM as the basis of the first RNNs.
[0047] The main function of the dynamic interest coding in the embodiment of the present application is to capture and express the dynamic initial interests of the target user by learning and understanding the behavior patterns of the target user.
[0048] In some other embodiments, the attention influence encoder of the embodiment of the present application improves the attention of the supplier level based on LSTM, and uses the second RNNs with LSTM as the basic component, which can be expressed as but not limited to:
[0049]
[0050] Among them, f(.) is the standard unit of LSTM, represents the hidden state of the nth supplier included in the latest supplier collaboration plan of partner k.
[0051] Further, in the embodiment of the present application, if the supplier in the supplier collaboration scheme is more relevant to the initial interest of the target user, the target user will pay more attention to this supplier, so it is necessary to calculate the similarity between the initial interest of the target user and the hidden state of the supplier in the partner supplier collaboration scheme as the attention score. The calculation formula of the attention score can be but is not limited to:
[0052]
[0053] in, is the hidden state of the ith supplier of partner k in the latest partner-supplier collaboration scenario, and W1 and W2 are learnable matrices used to project the above representations into the shared latent space.
[0054] In addition, the embodiment of the present application can also use the Softmax function to normalize the attention score, so that the attention influence encoder has the ability of probabilistic interpretation, wherein the process of using the Softmax function for normalization in the embodiment of the present application can be expressed as but not limited to:
[0055]
[0056] The main function of the attention influence encoder in the embodiment of the present application is to determine, based on the initial interest of the target user, which suppliers should be emphasized or ignored in the latent representation of the supplier collaboration scheme that the partner has used, and to learn and grasp the influence of the partner on the target user through a deep understanding of the supplier collaboration scheme used by the partner.
[0057] In other embodiments, the attention network in the embodiment of the present application is inspired by GATs, and the attention mechanism is applied to analyze the impact of ecological relationships on supplier selection. The inner product followed by a Softmax function is used to represent the similarity between the dynamic initial interest of the target user and the influence of the corresponding partner, which can be expressed as but not limited to:
[0058]
[0059] in, is the representation of the target user u at the lth convolutional layer, It is the attention weight of partner k’s preference influence on the target user u at the ecological relationship level of the l-th convolutional layer.
[0060] Furthermore, after obtaining the attention weight at the ecological relationship level, the embodiment of the present application combines the features on the target user influence graph, and then performs a nonlinear transformation to obtain the node representation of the next convolutional layer, which can be expressed as but not limited to:
[0061]
[0062]
[0063] Among them, W (l) and b (l) are the learnable and shared parameters in the lth convolutional layer, Represents a self-connection item, which is used to retain the dynamic initial interest of the target user. It represents the final explicit encoding of high-order preference influence, which can be obtained by stacking graph attention convolution layers L times and aggregating the information of L-order neighbors of the target user vertex in the influence graph.
[0064] The attention network in the embodiment of the present application can assign different weights to the target user's partners according to the target user's initial interests, and then aggregate the partners' representations according to different weights to obtain the target user's supplier collaboration solution representation, and then improve the personalization and accuracy of recommendations by establishing and analyzing the target user's relationships in the industrial ecosystem and simulating the propagation of supplier preferences during enterprise interactions.
[0065] In step S102, the target user data and the partner data are input into the pre-built SGHAN model to obtain candidate suppliers of the target user.
[0066] As a possible implementation method, the embodiment of the present application can input the target user data and the partner data into the pre-built SGHAN model to obtain the candidate suppliers of the target user.
[0067] Optionally, in one embodiment of the present application, the target user data and the partner data are input into a pre-built SGHAN model, including: based on the target user data, using dynamic interest coding to obtain the target user's initial interest representation; based on the partner data and the initial interest representation, using an attention influence encoder to obtain at least one initial preference representation of the target user under the influence of at least one partner; according to the initial interest representation, using a graph attention network to assign weights to at least one initial preference representation, so as to obtain the preference representation of the target user under the influence of at least one partner based on at least one initial preference representation with different weights; and predicting the target user's initial candidate suppliers based on the initial interest representation and the preference representation.
[0068] In the actual implementation process, the embodiment of the present application will input the target user data and partner data into the pre-built SGHAN model, and its execution process is as follows: Figure 3 As shown, it can be:
[0069] Step S301: Dynamic interest coding.
[0070] It is understandable that the embodiments of the present application can obtain the initial interest representation of the target user based on the target user data by using dynamic interest coding.
[0071] Specifically, the embodiment of the present application can regard the process of selecting each supplier as a point in a time series. Then RNNs can infer the current state of the target user by combining the previous state and the most recently selected supplier, and use the hidden state in RNNs To represent the target user’s initial interest in the current supplier collaboration solution, denoted as p u , in order to facilitate subsequent use, the embodiment of the present application can use it as the initial interest representation
[0072] Step S302: Attention affects encoder.
[0073] It can be understood that the embodiments of the present application can use the attention influence encoder to obtain at least one initial preference representation of the target user under the influence of at least one partner based on the partner data and the initial interest representation.
[0074] Specifically, the attention impact encoder of the embodiment of the present application can find out the individual suppliers that may have the greatest impact on the target users from the supplier collaboration scheme used by at least one partner, and at the same time, it is necessary to consider the time dependency when selecting these suppliers.
[0075] After obtaining the weight factor of each supplier in the supplier collaboration solution, the preference influence g of partner k on the target user is k It can be expressed as, but not limited to, the following formula. For the convenience of subsequent use, the embodiment of the present application can use it as the initial preference representation:
[0076]
[0077] Step S303: Graph attention network.
[0078] It can be understood that the embodiments of the present application can use the graph attention network to assign weights to at least one initial preference representation, thereby obtaining the preference representation of the target user under the influence of at least one partner, and predicting the initial candidate suppliers of the target user.
[0079] Specifically, the embodiment of the present application can construct a corresponding influence graph for each target user, wherein in the constructed influence graph, each node corresponds to the target user and its corresponding partner, and the edge represents the industrial ecological relationship between the target user and the partner. For the node of the target user, the dynamic initial interest p obtained in step S301 is used u Initial interest representation as an influence graph node For the partners of the target user node, the preference influence g obtained in step S302 is used. k Initial Preference Representation as an Influence Diagram
[0080] Furthermore, the embodiment of the present application can use GATs and Softmax function to represent the similarity between the dynamic initial interest of the target user and the influence of the corresponding partner, and the calculation formula is shown above.
[0081] Furthermore, after obtaining the attention weight at the ecological relationship level, the embodiment of the present application can combine the features on the target user influence graph and perform nonlinear transformation to obtain the node preference representation of the next convolutional layer. In order to avoid the problem of gradient disappearance, the embodiment of the present application adopts a skip connection to convert the user's dynamic initial interest p u and his partner's preference representation Taken together, the initial candidate suppliers of the target user are predicted, which can be expressed as but not limited to:
[0082]
[0083] in, It is the target user’s preference representation for candidate suppliers in the current supplier collaboration solution.
[0084] Optionally, in one embodiment of the present application, before obtaining the candidate suppliers of the target user, the method further includes: based on the initial candidate suppliers, solving the candidate probability of the initial candidate suppliers using a preset probability formula; based on the candidate probability, using a loss function to determine whether the initial candidate suppliers meet the preset candidate conditions; if the initial candidate suppliers meet the preset candidate conditions, the initial candidate suppliers are used as the candidate suppliers of the target user. The expression of the loss function can be, but is not limited to,:
[0085]
[0086] Where w represents the number of suppliers in the supplier collaboration solution.
[0087] As a possible implementation method, the embodiment of the present application can use a certain probability formula to solve the candidate probability of the initial candidate supplier based on the initial candidate supplier, and use the loss function to determine whether the initial candidate supplier meets certain candidate conditions, and determine the candidate supplier of the target user when the initial candidate supplier meets certain candidate conditions. Among them, the certain candidate conditions can be set by technicians in this field according to actual conditions, and this application does not make specific restrictions.
[0088] In this embodiment of the present application, the probability of supplier s becoming a candidate supplier in the current supplier collaboration scheme of the target user can be obtained by using a certain probability formula through a Softmax function, and the entire supplier set S is calculated, wherein the expression of the certain probability formula can be but is not limited to:
[0089]
[0090] in, represents the pth supplier included in the t+1th supplier collaboration scheme of user u, and n represents the number of suppliers in the supplier collaboration scheme. represents the final representation of the target user’s preference for the next supplier in the current supplier collaboration scheme, z s represents the embedding of candidate supplier s.
[0091] In addition, the embodiment of the present application selects negative log likelihood as the loss function, which can well measure the gap between the predicted probability of the model and the actual result. The expression of the loss function can be but is not limited to:
[0092]
[0093] Where w represents the number of suppliers in the supplier collaboration solution.
[0094] Among them, the goal of the embodiment of the present application is to minimize the function value of the loss function by adjusting the parameters of the model.
[0095] In step S103, a supplier collaboration plan suitable for the target user is generated based on the candidate suppliers.
[0096] That is to say, when the function value of the loss function is minimum, the embodiment of the present application can generate a supplier collaboration plan suitable for the target user based on the candidate suppliers.
[0097] The working principle of the method for creating a supplier collaboration solution based on industrial ecology proposed in the embodiment of the present application is introduced in detail below with reference to a specific embodiment.
[0098] in, Figure 4A flowchart of the working principle of a method for creating a supplier collaboration solution based on industrial ecology provided according to an embodiment of the present application.
[0099] Step S401: Problem definition.
[0100] It is understandable that the embodiments of the present application can define target user data and partner data to obtain relevant data.
[0101] Step S402: SGHAN model framework.
[0102] It is understandable that the SGHAN model framework of the present application embodiment is as follows Figure 2 As shown, it can be but not limited to: dynamic interest coding, attention influence encoder and graph attention network, etc., and this application does not make specific limitations.
[0103] Step S403: Dynamic interest coding.
[0104] It can be understood that the dynamic interest encoding in the embodiment of the present application utilizes the first RNNs to simulate the supplier selection process of the target user in the current supplier collaboration scheme, thereby extracting the initial interest of the target user and using the hidden state in the RNNs To represent the target user’s initial interest in the current supplier collaboration solution, denoted as p u .
[0105] Step S404: Attention affects encoder.
[0106] It can be understood that the attention influence encoder of the embodiment of the present application can find out the individual suppliers that may have the greatest impact on the target user from the supplier collaboration scheme used by at least one partner, and at the same time, it is necessary to consider the time dependency when selecting these suppliers, so as to obtain the preference influence g of the partner k on the target user. k .
[0107] Step S405: Graph attention network.
[0108] It can be understood that the attention network in the embodiment of the present application can assign weights to at least one initial preference representation, thereby obtaining a preference representation of the target user under the influence of at least one partner.
[0109] Step S406: predicting and recommending candidate suppliers.
[0110] It can be understood that the embodiments of the present application can predict the initial candidate suppliers of the target user by adopting a skip connection to add the dynamic initial interests of the user and the preference representations of his partners.
[0111] Furthermore, the embodiment of the present application can obtain the probability of supplier s becoming a candidate supplier in the target user's current supplier collaboration plan by a certain probability formula, and calculate the entire supplier set S.
[0112] Step S407: Loss function calculation.
[0113] It can be understood that the embodiment of the present application selects negative log-likelihood as the loss function, and minimizes the function value of the loss function by adjusting the parameters of the model, thereby generating a supplier collaboration plan suitable for the target user.
[0114] According to the supplier collaboration solution creation method based on industrial ecology proposed in the embodiment of the present application, a supplier collaboration solution suitable for the target user can be dynamically generated based on the pre-built SGHAN model. As a deep learning model, the SGHAN model can learn similar behaviors from partners who are helpful to the target user in selecting suppliers during the supplier collaboration solution creation process, so as to meet the dynamic needs of the target user and dynamically generate the supplier collaboration solution for the target user more accurately. Thus, the problems in the related technology that the dynamic needs of the target user cannot be met, it is difficult to accurately establish a model of the dynamic needs of the target user, and the process of dynamically creating a supplier collaboration solution is more complicated are solved.
[0115] Next, a device for creating a supplier collaboration solution based on industrial ecology proposed in accordance with an embodiment of the present application will be described with reference to the accompanying drawings.
[0116] Figure 5 A block diagram of a device for creating a supplier collaboration solution based on industrial ecology provided according to an embodiment of the present application.
[0117] like Figure 5 As shown, the supplier collaboration solution creation device 10 based on industrial ecology includes: an acquisition module 100, an input module 200 and a first generation module 300.
[0118] The acquisition module 100 is used to acquire target user data of a target user in the industrial ecology and partner data of at least one partner corresponding to the target user.
[0119] The input module 200 is used to input the target user data and the partner data into the pre-built SGHAN model to obtain candidate suppliers of the target user.
[0120] The first generating module 300 is used to generate a supplier collaboration solution suitable for a target user based on candidate suppliers.
[0121] Optionally, in one embodiment of the present application, it further includes: a first building module, a second building module, a third building module and a fourth building module.
[0122] Among them, the first construction module is used to establish the first RNNs based on LSTM before inputting the target user data and partner data into the pre-built SGHAN model, and use the first RNNs to construct dynamic interest coding.
[0123] The second building module is used to establish a second RNNs based on LSTM and Softmax functions, and use the second RNNs to build an attention influence encoder.
[0124] The third building module is used to build a graph attention network using GATs and Softmax function.
[0125] The fourth building module is used to build the SGHAN model based on dynamic interest encoding, attention influence encoder and graph attention network.
[0126] Optionally, in one embodiment of the present application, the input module 200 includes: a first generation unit, a second generation unit, a third generation unit and a prediction unit.
[0127] The first generating unit is used to obtain the initial interest representation of the target user by using dynamic interest coding based on the target user data.
[0128] The second generating unit is used to obtain at least one initial preference representation of the target user under the influence of at least one partner based on the partner data and the initial interest representation using an attention influence encoder.
[0129] The third generating unit is used to assign a weight to at least one initial preference representation according to the initial interest representation by using a graph attention network, so as to obtain a preference representation of the target user under the influence of at least one partner based on at least one initial preference representation with different weights.
[0130] The prediction unit is used to predict the initial candidate suppliers of the target user based on the initial interest representation and the preference representation.
[0131] Optionally, in one embodiment of the present application, it further includes: a solution module, a judgment module and a second generation module.
[0132] The solution module is used to solve the candidate probability of the initial candidate suppliers by using a preset probability formula based on the initial candidate suppliers before obtaining the candidate suppliers of the target user.
[0133] The judgment module is used to judge whether the initial candidate supplier meets the preset candidate conditions based on the candidate probability and using the loss function.
[0134] The second generating module is used to use the initial candidate supplier as a candidate supplier of the target user when the initial candidate supplier meets the preset candidate condition.
[0135] Optionally, in one embodiment of the present application, the expression of the loss function may be, but is not limited to,:
[0136]
[0137] Where w represents the number of suppliers in the supplier collaboration solution.
[0138] It should be noted that the aforementioned explanation of the embodiment of the method for creating a supplier collaboration solution based on industrial ecology is also applicable to the device for creating a supplier collaboration solution based on industrial ecology in this embodiment, and will not be repeated here.
[0139] According to the supplier collaboration solution creation device based on industrial ecology proposed in the embodiment of the present application, a supplier collaboration solution suitable for the target user can be dynamically generated based on the pre-built SGHAN model. As a deep learning model, the SGHAN model can learn similar behaviors from partners who are helpful to the target user in selecting suppliers during the supplier collaboration solution creation process, so as to meet the dynamic needs of the target user and dynamically generate the supplier collaboration solution for the target user more accurately. Thus, the problems in the related technology that the dynamic needs of the target user cannot be met, it is difficult to accurately establish a model of the dynamic needs of the target user, and the process of dynamically creating a supplier collaboration solution is more complicated are solved.
[0140] Figure 6 This is a schematic diagram of the structure of an electronic device provided according to an embodiment of the present application. The electronic device may include:
[0141] A memory 601 , a processor 602 , and a computer program stored in the memory 601 and executable on the processor 602 .
[0142] When the processor 602 executes the program, the method for creating a supplier collaboration solution based on industrial ecology provided in the above embodiment is implemented.
[0143] Furthermore, the electronic device further comprises:
[0144] The communication interface 603 is used for communication between the memory 601 and the processor 602 .
[0145] The memory 601 is used to store computer programs that can be executed on the processor 602 .
[0146] The memory 601 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.
[0147] If the memory 601, the processor 602 and the communication interface 603 are implemented independently, the communication interface 603, the memory 601 and the processor 602 can be connected to each other through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 6 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.
[0148] Optionally, in a specific implementation, if the memory 601, the processor 602 and the communication interface 603 are integrated on a chip, the memory 601, the processor 602 and the communication interface 603 can communicate with each other through an internal interface.
[0149] The processor 602 may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.
[0150] An embodiment of the present application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned method for creating a supplier collaboration solution based on industrial ecology.
[0151] An embodiment of the present application also provides a computer program product, including a computer program, which, when executed, implements the above-mentioned method for creating a supplier collaboration solution based on industrial ecology.
[0152] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.
[0153] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise clearly and specifically defined.
[0154] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or N executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present application belong.
[0155] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purpose of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or N wirings (electronic devices), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways as necessary and then storing it in a computer memory.
[0156] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above embodiment, the N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. If implemented by hardware, as in another embodiment, it can be implemented by any one or a combination of multiple of the following technologies known in the art: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.
[0157] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.
[0158] In addition, each functional unit in each embodiment of the present application may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.
[0159] The storage medium mentioned above may be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present application. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present application.
Claims
1. A method for creating a supplier collaboration solution based on industrial ecology, characterized in that: The following steps are involved: Acquire target user data of a target user in the industrial ecology and partner data of at least one partner corresponding to the target user; Input the target user data and the partner data into a pre-built social hierarchical graph attention network (SGHAN) model to obtain candidate suppliers for the target user; A supplier collaboration plan suitable for the target user is generated based on the candidate suppliers.
2. The method for creating a supplier collaboration solution based on industrial ecology according to claim 1 is characterized in that: Before inputting the target user data and the partner data into the pre-built SGHAN model, the method further includes: Establishing a first recurrent neural network RNNs based on long short-term memory (LSTM), and using the first RNNs to construct dynamic interest coding; Establishing a second RNNs based on the LSTM and Softmax functions, and using the second RNNs to construct an attention influence encoder; Construct a graph attention network using graph attention network GATs and the Softmax function; The SGHAN model is constructed based on the dynamic interest encoding, the attention influence encoder and the graph attention network.
3. The method for creating a supplier collaboration solution based on industrial ecology according to claim 1 is characterized in that: The step of inputting the target user data and the partner data into a pre-built SGHAN model includes: Based on the target user data, obtaining the initial interest representation of the target user by using the dynamic interest coding; Based on the partner data and the initial interest representation, using the attention influence encoder to obtain at least one initial preference representation of the target user under the influence of the at least one partner; According to the initial interest representation, using the graph attention network to assign a weight to the at least one initial preference representation, so as to obtain a preference representation of the target user under the influence of the at least one partner based on the at least one initial preference representation with different weights; Predicting initial candidate suppliers for the target user based on the initial interest representation and the preference representation.
4. The method for creating a supplier collaboration solution based on industrial ecology according to claim 1 is characterized in that: Before obtaining candidate suppliers of the target user, the method further includes: Based on the initial candidate suppliers, solving the candidate probability of the initial candidate suppliers using a preset probability formula; Based on the candidate probability, using a loss function to determine whether the initial candidate supplier meets the preset candidate condition; If the initial candidate supplier meets the preset candidate condition, the initial candidate supplier is used as a candidate supplier for the target user.
5. The method for creating a supplier collaboration solution based on industrial ecology according to claim 4 is characterized in that: The expression of the loss function is: Where w represents the number of suppliers in the supplier collaboration solution.
6. A supplier collaboration solution creation device based on industrial ecology, characterized in that: include: An acquisition module, used to acquire target user data of a target user in the industrial ecology and partner data of at least one partner corresponding to the target user; An input module, used for inputting the target user data and the partner data into a pre-built SGHAN model to obtain candidate suppliers of the target user; The first generating module is used to generate a supplier collaboration plan suitable for the target user based on the candidate suppliers.
7. The device for creating a supplier collaboration solution based on industrial ecology according to claim 6, characterized in that: Also includes: A first construction module, used for establishing first RNNs based on LSTM before inputting the target user data and the partner data into the pre-constructed SGHAN model, and constructing dynamic interest coding using the first RNNs; A second building module is used to establish a second RNNs based on the LSTM and Softmax functions, and use the second RNNs to build an attention influence encoder; The third building module is used to build a graph attention network using the graph attention network GATs and the Softmax function; The fourth building module is used to build the SGHAN model based on the dynamic interest encoding, the attention influence encoder and the graph attention network.
8. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for creating a supplier collaboration solution based on industrial ecology as described in any one of claims 1 to 5.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement the method for creating a supplier collaboration solution based on industrial ecology as described in any one of claims 1-5.
10. A computer program product, characterized in that It includes a computer program which, when executed, is used to implement the method for creating a supplier collaboration solution based on industrial ecology as described in any one of claims 1-5.
Citation Information
Patent Citations
Selection model for enterprise informatization cooperation partner and implementation method thereof
CN105354688A
Method and device for determining associated provider
CN110517099A
Industrial chain analysis-based park investment attraction recommendation method and system, terminal and medium
CN116383511A
Developing and structuring business ecosystems
US20080300930A1
Method and system for supplier relationship management
WO2001099018A1
Cited By
Enterprise service scheme generation method and device, enterprise service scheme output method and device, equipment and medium
CN121094572A
