Method and system for generating a solution based on integration of demand-resource matching

Abstract

This application discloses a solution generation method and system based on integrated demand-resource matching, relating to the field of resource scheduling technology. The method includes: acquiring multimodal data generated during on-site customer interaction; extracting features from the multimodal data to obtain a demand element vector; inputting the demand element vector into a preset solution generation model to obtain a target solution; mapping the feature parameters in the target solution to a research and development task flow, and obtaining the skill tag vectors of candidate research and development resources and the demand vector of the research and development task flow; calculating the matching score between the research and development task flow and the candidate research and development resources based on the skill tag vectors and the demand vector; and determining the target research and development resource among the candidate research and development resources based on the matching score, thereby achieving integrated demand-solution-resource matching. This application achieves the technical effect of improving resource scheduling efficiency.

Description

Technical Field

[0001] This application relates to the field of resource scheduling technology, specifically to a solution generation method and system based on integrated demand-resource matching. Background Technology

[0002] In the process of industrial product customization and R&D, three stages are typically involved: requirements gathering, solution design, and resource scheduling. Current technologies often rely on unstructured text records for requirements gathering, lacking standardized data definitions, making it difficult to accurately understand requirement parameters in subsequent design phases. Simultaneously, solution generation usually depends on manual processes or independent computer-aided design software, and the generated solution parameters are often not validated against industry standards or company specifications in real time, easily leading to designs that are impractical. Furthermore, the matching and scheduling of R&D resources often lags behind solution confirmation and relies on the project manager's subjective experience for task breakdown and personnel assignment, lacking a precise matching mechanism based on objective data characteristics. This discreteness in data processing at each stage results in high information loss and low resource scheduling efficiency throughout the entire process from requirements gathering to implementation. Summary of the Invention

[0003] To address the technical issues in related technologies where the matching and scheduling of R&D resources often lags behind solution confirmation, and relies on the subjective experience of project managers for task decomposition and personnel assignment, lacking a precise matching mechanism based on objective data characteristics, resulting in high information loss and low resource scheduling efficiency throughout the entire process from requirement to implementation, this application provides a solution generation method and system based on integrated requirement-resource matching.

[0004] The specific technical solution adopted is as follows: Acquire multimodal data generated during on-site interactions with customers, extract features from the multimodal data, and obtain a demand element vector; Input the demand element vector into the preset solution generation model to obtain the target solution; The feature parameters in the target solution are mapped to the R&D task flow, and the skill tag vectors of the candidate R&D resources and the requirement vectors of the R&D task flow are obtained. Based on skill tag vectors and demand vectors, a matching score is calculated between the R&D task flow and candidate R&D resources. Based on the matching score, the target R&D resource among the candidate R&D resources is determined to achieve integrated demand-solution-resource matching.

[0005] In one possible implementation of this application, feature extraction is performed on multimodal data to obtain a demand element vector, including: Determine the demand element identification template, which is based on the four-dimensional elements of form, function, parameters and standard. By using demand element identification templates, multimodal data is structurally parsed to obtain standardized element vectors; Implicit demand analysis is performed on the standardized element vector to obtain the demand element vector.

[0006] In one possible implementation of this application, implicit demand analysis is performed on the standardized element vector to obtain a demand element vector, including: The standardized element vectors are matched with multiple historical cases in a pre-set multimodal knowledge base to calculate the demand similarity. When the similarity of requirements exceeds a preset threshold, the implicit requirements in the historical cases corresponding to the similarity of requirements are extracted. Implicit requirements include at least one of potential performance upgrade requirements and cost control requirements. By combining implicit needs with the customer's business scenario, the standardized element vector is optimized to obtain the demand element vector.

[0007] In one possible implementation of this application, the demand element vector is input into a preset solution generation model to obtain the target solution, including: Input the demand element vector into the preset solution generation model, process the demand element vector based on the preset solution generation model, and output a preliminary solution that meets the requirements. The preliminary design was adapted to design specifications to obtain a revised design. The revised solution is compared with the solutions of competitors, and the revised solution is updated based on the comparison results to obtain the target solution.

[0008] In one possible implementation of this application, the modified solution is compared with competitor solution information, and the modified solution is updated based on the comparison results to obtain the target solution, including: Extract competitor solution information related to the modified solution from the pre-defined multimodal knowledge base; We analyze the comparative differences between competitor solutions and target solutions from multiple dimensions to enable clients to adjust their solutions based on the comparative differences and determine the adjustment parameters. Receive the customer's adjustment parameters, update the revised plan based on the adjustment parameters, and generate the target plan after customer confirmation.

[0009] In one possible implementation of this application, a matching score is calculated between the R&D task flow and candidate R&D resources based on skill tag vectors and demand vectors, including: Calculate the cosine similarity between the skill tag vector and the demand vector; Cosine similarity is used as the matching score between R&D task flow and candidate R&D resources.

[0010] In one possible implementation of this application, determining the target R&D resource among the candidate R&D resources based on the matching score includes: Determine the maximum value of the matching score; The candidate R&D resource corresponding to the maximum matching score is selected as the target R&D resource.

[0011] In one possible implementation of this application, after determining the target R&D resource among the candidate R&D resources based on the matching score, the method further includes: When it is determined that there is a gap in R&D capabilities for the target R&D resources, R&D resource data is extracted from a pre-set multimodal knowledge base; The R&D resource data corresponding to the R&D capability gap will be added to the target R&D resources.

[0012] In one possible implementation of this application, before performing feature extraction on the multimodal data to obtain the demand element vector, the method further includes: Semantic segmentation is performed on multimodal data to separate statements containing requirement parameters from redundant statements.

[0013] To achieve the above objectives, a solution generation system based on integrated demand-resource matching is also provided. The system includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of the solution generation method based on integrated demand-resource matching.

[0014] This application has, but is not limited to, the following technical effects: By acquiring multimodal data generated during on-site interactions with customers, feature extraction is performed on the multimodal data to obtain a demand element vector. Then, the demand element vector is input into a preset solution generation model to obtain a target solution. The feature parameters in the target solution are then mapped to a research and development task flow. The skill tag vectors of candidate research and development resources and the demand vectors of the research and development task flow are obtained. By matching the skill tag vectors with the demand vectors, a matching score is obtained. The matching score is used to determine the required target research and development resources, realizing the integration of demand-solution-resource matching. This eliminates the reliance on experience and subjective bias in manual scheduling, significantly improves the accuracy and response speed of research and development resource allocation, and solves the problems of high information loss and low resource scheduling efficiency throughout the entire process from demand to implementation. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating the first embodiment of the solution generation method based on integrated demand-resource matching in this application; Figure 2 This is a schematic diagram of the device structure of the hardware operating environment involved in the embodiments of this application. Detailed Implementation

[0016] It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit this application.

[0017] This application provides a solution generation method based on integrated demand-resource matching. In the first embodiment of this solution generation method based on integrated demand-resource matching, refer to... Figure 1 ,include: Step S10: Obtain multimodal data generated during on-site interaction with customers, extract features from the multimodal data, and obtain a demand element vector.

[0018] As an example, the integrated demand-resource matching solution generation method can also be applied to an integrated demand-resource matching solution generation system. This system includes four core modules: a mobile intelligent agent, a cloud-based AI processing engine, a pre-set multimodal knowledge base, and a real-time collaboration platform. Each module achieves real-time data interaction and collaborative work through an encrypted communication protocol, as detailed below: The mobile intelligent agent is primarily a lightweight AI assistant within the mobile devices (such as smartphones and tablets) used by sales personnel when communicating with customers. Serving as the front-end carrier for demand collection and on-site interaction, it receives multimodal data (voice, text, images, etc.) from real-time communication between sales personnel and customers, completing initial local speech-to-text transcription and data preprocessing. It then sends requests to the cloud-based AI processing engine for demand analysis and solution generation, instantly presenting the results, such as renderings and solution documents, to the customer. It supports the input and uploading of on-site customer confirmation actions, including digital signature collection and demand version confirmation. This module has offline caching capabilities, allowing it to temporarily store data in on-site environments with poor network signals, automatically synchronizing to the cloud once the network is restored, ensuring communication continuity.

[0019] The cloud-based AI processing engine is the core computing unit of this system. It is mainly used to analyze the needs of multimodal data, generate solutions based on the analyzed needs, and then match them with the corresponding R&D resources.

[0020] A pre-defined multimodal knowledge base is used to store various types of data. It adopts a three-tier architecture design: product library, standard library, and competitor library. The functions of each layer are as follows: Product Library: Stores data such as parameters, specifications, design drawings, and historical cases of the company's existing products, supporting rapid matching of requirements with existing products and iterative optimization of solutions.

[0021] Standards Library: Integrates industry-related technical standards, design specifications, safety guidelines, and other content to ensure that generated solutions automatically adapt to compliance requirements.

[0022] Competitor Database: This database includes product information, strengths and weaknesses, and market feedback data of mainstream competitors, providing a data basis for competitor comparison analysis and differentiated solution design.

[0023] The real-time collaboration platform is the core hub connecting the three major roles of business, product, and R&D, building a full-process collaborative workflow. Its functions include: real-time synchronization of on-site collected requirements information and customer confirmation results to the product and R&D ends; support for automatic allocation and progress tracking of R&D tasks; real-time notification of requirement changes and sharing of impact assessment results; and enabling online communication and document collaboration among business, product, and R&D personnel, breaking down information barriers and ensuring the accuracy and efficiency of requirement transmission.

[0024] As an example, multimodal data could be voice / text / image data from communication between a salesperson and a customer.

[0025] As an example, feature extraction of multimodal data can be performed by using a pre-defined requirement element template to perform structured parsing of the obtained multimodal data, resulting in a requirement element vector. The requirement element vector includes product form requirements, functional requirements, parameter indicators, compliance standards, and implicit requirement data.

[0026] Before step S10, the following are also included: Semantic segmentation is performed on multimodal data to separate statements containing requirement parameters from redundant statements.

[0027] As an example, the mobile intelligent agent has a high-precision real-time speech transcription function, which can synchronously convert the voice communication between salesperson and customer into text. At the same time, it adopts a Transformer-based semantic segmentation algorithm to automatically divide the communication text into statements containing requirement parameters and redundant statements, extract the core communication context, and provide accurate input for subsequent requirements analysis.

[0028] Step S10 includes: The requirement element identification template is determined based on the four-dimensional elements of form, function, parameters and standard.

[0029] By using demand element identification templates, multimodal data is structured and parsed to obtain standardized element vectors.

[0030] As an example, a four-dimensional requirement element identification template of "form-function-parameter-standard" is preset. The cloud AI processing engine performs structured parsing on the transcribed text data based on this template, automatically extracting key information such as product form requirements (e.g., size, appearance style), functional requirements (e.g., core functions, auxiliary functions), parameter indicators (e.g., performance parameters, material parameters) and compliance standards (e.g., industry standards, safety levels), and generating standardized element vectors.

[0031] Implicit demand analysis is performed on the standardized element vector to obtain the demand element vector.

[0032] As an example, after generating standardized element vectors, customers often have some implicit needs that require the assistance of sales staff to uncover. In this case, implicit needs analysis can be performed on the standardized element vectors to obtain demand element vectors, thereby accurately capturing the user's actual needs.

[0033] The steps involved in performing implicit demand analysis on the standardized element vectors to obtain the demand element vectors include: The standardized element vectors are matched with multiple historical cases in a pre-set multimodal knowledge base to calculate the demand similarity.

[0034] As an example, a latent demand mining model is constructed based on historical case similarity analysis. The currently extracted explicit demand elements are matched with historical cases in a pre-set multimodal knowledge base to calculate demand similarity. Demand similarity can be calculated using cosine similarity.

[0035] When the similarity of requirements exceeds a preset threshold, implicit requirements are extracted from historical cases corresponding to the similarity of requirements. Implicit requirements include at least one of potential performance upgrade requirements and cost control requirements.

[0036] By combining implicit needs with the customer's business scenario, the standardized element vector is optimized to obtain the demand element vector.

[0037] As an example, the preset threshold can be 0.85, 0.9, etc., and there is no specific limitation.

[0038] As an example, for historical cases with high demand similarity (e.g., greater than 0.85), we extract the implicit demands that were not explicitly stated but were eventually implemented (such as potential performance upgrade demands or cost control demands). We then optimize these demands by combining them with information such as the current customer's industry attributes and business scenarios to form a list of recommended implicit demands, which helps sales staff and customers to further confirm these demands.

[0039] Step S20: Input the demand element vector into the preset scheme generation model to obtain the target scheme.

[0040] As an example, the demand element vector is input into the preset solution generation model, and the demand element vector is processed by the preset solution generation model to obtain the required target solution.

[0041] As an example, the preset solution generation model includes a structure guidance layer and a texture generation layer. The system first parses the geometric constraint parameters in the design rule base to construct a three-dimensional spatial constraint body (or depth map) of the product. This constraint body is used as the input of the structure guidance layer to lock the physical boundaries and core component positions of the product. The texture generation layer performs feature mapping and texture rendering within the range defined by the constraint body based on the semantic features in the customer's requirement element vector. After generation, the system extracts the feature parameters of the result and compares them with the design rules. If there is a conflict, the conflict area is locked and locally resampled for correction, and finally the target solution is generated.

[0042] As an example, the target solution can be a 2D plan view, a 3D solid view, or a textual solution, etc., without any specific limitations.

[0043] Among them, step S20, which generates a solution based on the integrated demand-resource matching, also includes steps S21 to S23, including: Step S21: Input the demand element vector into the preset scheme generation model, process the demand element vector based on the preset scheme generation model, and output a preliminary scheme that meets the requirements.

[0044] As an example, the preset solution generation model receives information such as form, function, and parameters from the demand element vector, automatically calls product library data and design specifications in the multimodal knowledge base, and generates a preliminary rendering that meets the requirements, which is the preliminary solution. Salespeople can adjust the parameters based on customer feedback, and the model iterates and updates the rendering in real time, realizing instant interaction between "demand adjustment and effect preview".

[0045] Step S22: Adapt the preliminary scheme to design specifications to obtain the revised scheme.

[0046] As an example, a pre-defined company design specification rule library is used, which includes color system, layout requirements, structural standards, etc. During the rendering process, the engine automatically verifies whether the generated scheme conforms to the design specifications and makes real-time corrections to the non-conforming parts (such as adjusting color matching and optimizing structural proportions) to obtain a corrected scheme. This ensures that the corrected scheme can be directly connected to the R&D stage without further modifications.

[0047] Step S23: Compare the revised solution with the competitor's solution information, and update the revised solution based on the comparison results to obtain the target solution.

[0048] As an example, competitive solutions information from a pre-defined multimodal knowledge base is invoked to conduct comparative analysis in terms of functionality, parameters, cost, advantages and disadvantages, thereby optimizing the modified solution and obtaining the final target solution.

[0049] Step S23 includes: Extract competitor solution information related to the modified solution from the preset multimodal knowledge base.

[0050] As an example, competitor information could include competitor product information, strengths and weaknesses, market feedback, etc.

[0051] We analyze the comparative data between competitor solutions and target solutions from multiple dimensions to enable clients to adjust their solutions based on the comparative data and determine the adjustment parameters.

[0052] As an example, a comparative analysis is conducted with the target solution from multiple dimensions such as functionality, parameters, cost, advantages and disadvantages to generate visualized comparative difference data. The comparative difference data can be a competitor comparison report. Presenting the comparative difference data to the client allows the client to better understand the differentiated advantages of the current solution, providing data support for the client's decision-making and determining the adjustment parameters in the target solution.

[0053] Receive the customer's adjustment parameters, update the revised plan based on the adjustment parameters, and generate the target plan after customer confirmation.

[0054] As an example, the solution is continuously updated by receiving the customer's adjustment parameters until the customer confirms and signs off on it, at which point the target solution is generated.

[0055] Step S30: Map the feature parameters in the target solution to the R&D task flow, and obtain the skill tag vector of the candidate R&D resources and the requirement vector of the R&D task flow.

[0056] As an example, by introducing a parameter-task mapping rule base, specific product feature parameters (such as material, protocol, and size) can be mapped to standardized R&D task flows and their corresponding skill tags.

[0057] As an example, the skill proficiency and project experience of R&D personnel are transformed into numerical values ​​in the same coordinate system to generate skill tag vectors. Candidate R&D resources can be different R&D personnel, as well as other resources that need to be utilized. For each R&D personnel, information is obtained regarding their professional field (e.g., mechanical design, software development), skill level, project experience, and areas of expertise. Their skill tag vector can be calculated based on their comprehensive abilities. For example, a zero-based vector of length N can be constructed. If a candidate R&D personnel possesses the "thermal simulation" skill tag, has a proficiency score of 5 (out of 5), and has 3 related historical projects, then the ability value is entered at the corresponding index position in the vector. The ability value, Value, can be calculated as: Value = Normalized proficiency value × (1 + log(number of projects participated in + ... )), To prevent extremely small positive numbers from being zero, for example, 10 -6 (When the R&D personnel are new employees, the number of projects is 0). After the calculation of each ability value is completed, the skill tag vector of each R&D personnel is obtained. The normalization method can be sigmoid normalization.

[0058] As an example, the skills required for each subtask in the task flow are weighted and aggregated. The weights are determined by the estimated time and difficulty coefficient of the task, generating a sparse demand vector. The demand vector can be a vector obtained by aggregating the skills required for all subtasks. For example, if the task requires "thermal simulation" and the estimated difficulty coefficient is 0.8, with an estimated time of 10 hours, then the weight value is filled in at the corresponding index position of the vector. The suggested weight calculation formula is: W = ∑(difficulty coefficient × estimated time). After calculating the original weight value W, the demand vector is normalized using the L2 norm to transform it into a unit vector, so that the magnitude of the processed vector is 1, and the weight values ​​of each dimension are between 0 and 1, which facilitates subsequent comparisons with resource vectors using various distance metrics.

[0059] Step S40: Based on the skill tag vector and the demand vector, calculate the matching score between the R&D task flow and the candidate R&D resources, and determine the target R&D resource among the candidate R&D resources according to the matching score, so as to achieve the integration of demand-solution-resource matching.

[0060] As an example, by determining the similarity relationship between skill tag vectors and demand vectors, the R&D task flow can be matched with the required target R&D resources, thereby achieving integrated demand-solution-resource matching.

[0061] Step S40 includes: Calculate the cosine similarity between the skill tag vector and the demand vector; Cosine similarity is used as the matching score between R&D task flow and candidate R&D resources.

[0062] Step S40 further includes: Determine the maximum value of the matching score; The candidate R&D resource corresponding to the maximum matching score is selected as the target R&D resource.

[0063] As an example, the candidate R&D resource corresponding to the maximum matching score in the matching score is selected as the target R&D resource. After determining the target R&D resource, a member matching list and division of labor suggestions are formed.

[0064] As an example, a machine learning model is trained based on historical project data. Input parameters such as the current project's requirements, technical difficulty, and required resources are given to automatically assess the project's complexity level. The input features of the machine learning model are [N... func N stack C std This refers to the number of functional points, the number of technology stacks involved, and the standard level coefficient. It also combines the skill level of the matching members in the target R&D resources with historical project efficiency data to predict the project duration and provide accurate basis for project planning.

[0065] After step S40, the following steps are also included: When it is determined that there is a gap in R&D capabilities for the target R&D resources, R&D resource data is extracted from a pre-set multimodal knowledge base; The R&D resource data corresponding to the R&D capability gap will be added to the target R&D resources.

[0066] As an example, after identifying the target R&D resources and matching members, the project requirements are compared with the total skills of the matched members to identify existing capability gaps. Based on the talent resource data or R&D resource data in the preset multimodal knowledge base, internal reserve talents or external cooperative resources with the corresponding skills for R&D capability gaps are recommended to the target R&D resources to ensure the smooth progress of the project.

[0067] As an example, one way to identify a gap in R&D capabilities is: Skill gap = Max(task requirement level) - Max(team member level). If the skill gap > 0, then there is a research and development capability gap.

[0068] Human resource gap = ∑(task working hours) - ∑(remaining available working hours of members). If the human resource gap > 0, then there is a gap in R&D capability.

[0069] This application achieves an instant closed loop of "on-site communication - needs extraction - solution preview - customer confirmation", significantly improving communication efficiency and customer experience. The built-in competitor comparison algorithm can generate differentiated analysis reports simultaneously, providing real-time data support for customers' on-site decision-making.

[0070] This application provides a solution generation method based on integrated demand-resource matching. By acquiring multimodal data generated during on-site interactions with customers, feature extraction is performed on the multimodal data to obtain demand element vectors. Then, the demand element vectors are input into a preset solution generation model to obtain target solutions. The feature parameters in the target solutions are then mapped to R&D task flows, and the skill tag vectors of candidate R&D resources and the demand vectors of the R&D task flows are obtained. By matching the skill tag vectors with the demand vectors, a matching score is obtained. The matching score is used to determine the required target R&D resources, realizing integrated demand-solution-resource matching. This eliminates the experience dependence and subjective bias of manual scheduling, significantly improves the accuracy and response speed of R&D resource allocation, and solves the problems of high information loss and low resource scheduling efficiency throughout the entire process from demand to implementation.

[0071] Reference Figure 2 , Figure 2 This is a schematic diagram of the device structure of the hardware operating environment involved in the embodiments of this application.

[0072] like Figure 2 As shown, the integrated demand-resource matching solution generation device may include: a processor 1001, a memory 1003, and a communication bus 1002. The communication bus 1002 is used to realize the connection and communication between the processor 1001 and the memory 1003.

[0073] Optionally, the integrated demand-resource matching solution generation device may also include a user interface, network interface, camera, RF (Radio Frequency) circuitry, sensors, WiFi module, etc. The user interface may include a display screen and an input submodule such as a keyboard; optional user interfaces may also include standard wired and wireless interfaces. The network interface may include standard wired and wireless interfaces (such as a Wi-Fi interface).

[0074] Those skilled in the art will understand that Figure 2 The structure of the integrated demand-resource matching solution generation device shown in the figure does not constitute a limitation on the integrated demand-resource matching solution generation device. It may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0075] like Figure 2As shown, the memory 1003, serving as a storage medium, may include an operating system, a network communication module, and a solution generation program based on integrated demand-resource matching. The operating system is a program that manages and controls the hardware and software resources of the integrated demand-resource matching solution generation device, supporting the operation of the integrated demand-resource matching solution generation program and other software and / or programs. The network communication module is used to enable communication between the various components within the memory 1003, as well as communication with other hardware and software in the integrated demand-resource matching solution generation system.

[0076] exist Figure 2 In the integrated demand-resource matching solution generation device shown, the processor 1001 is used to execute the integrated demand-resource matching solution generation program stored in the memory 1003 to implement the steps of the integrated demand-resource matching solution generation method described above.

[0077] The specific implementation of the solution generation device based on demand-resource matching integration in this application is basically the same as the embodiments of the solution generation method based on demand-resource matching integration described above, and will not be repeated here.

[0078] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0079] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0080] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0081] The above are merely preferred embodiments of this application and do not limit the scope of this application. Any equivalent structural or procedural transformations made based on the description and drawings of this application, or direct or indirect applications in other related technical fields, are similarly included within the scope of protection of this application.

[0082] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0083] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A solution generation method based on integrated demand-resource matching, characterized in that, The method includes: Acquire multimodal data generated during on-site interactions with customers, extract features from the multimodal data, and obtain a demand element vector; The required element vector is input into a preset solution generation model to obtain the target solution; The feature parameters in the target solution are mapped to the R&D task flow, and the skill tag vectors of the candidate R&D resources and the demand vectors of the R&D task flow are obtained. Based on the skill tag vector and the demand vector, a matching score is calculated between the R&D task flow and the candidate R&D resources. Based on the matching score, the target R&D resource among the candidate R&D resources is determined to achieve integrated demand-solution-resource matching.

2. The solution generation method based on integrated demand-resource matching as described in claim 1, characterized in that, The step of extracting features from the multimodal data to obtain the demand element vector includes: A demand element identification template is determined, which is based on morphology, function, parameters and standard four-dimensional elements; The multimodal data is structured and parsed using the demand element identification template to obtain a standardized element vector; The standardized element vector is subjected to implicit demand analysis to obtain the demand element vector.

3. The solution generation method based on integrated demand-resource matching as described in claim 2, characterized in that, The implicit demand analysis processing of the standardized element vector to obtain the demand element vector includes: The standardized element vector is matched with multiple historical cases in a preset multimodal knowledge base to calculate the demand similarity. When the demand similarity is greater than a preset threshold, the implicit demands in the historical cases corresponding to the demand similarity are extracted. The implicit demands include at least one of potential performance upgrade demands and cost control demands. By combining the implicit requirements with the customer's business scenario, the standardized element vector is optimized to obtain the requirement element vector.

4. The solution generation method based on integrated demand-resource matching as described in claim 1, characterized in that, The step of inputting the demand element vector into a preset solution generation model to obtain the target solution includes: The required element vector is input into a preset solution generation model. Based on the preset solution generation model, the required element vector is processed to output a preliminary solution that meets the requirements. The preliminary scheme is adapted to design specifications to obtain a revised scheme; The modified solution is compared with the competitor's solution information, and the modified solution is updated based on the comparison results to obtain the target solution.

5. The solution generation method based on integrated demand-resource matching as described in claim 4, characterized in that, The step of comparing the revised solution with competitor solutions and updating the revised solution based on the comparison results to obtain the target solution includes: Extract competitor solution information associated with the corrected solution from a preset multimodal knowledge base; The comparison data between the competitor's solution and the target solution is analyzed from multiple dimensions to allow the customer to adjust the solution based on the comparison data and determine the adjustment parameters. Receive the customer's adjustment parameters, update the correction scheme based on the adjustment parameters, and generate the target scheme after customer confirmation.

6. The solution generation method based on integrated demand-resource matching as described in claim 1, characterized in that, The step of calculating the matching score between the R&D task flow and candidate R&D resources based on the skill tag vector and the demand vector includes: Calculate the cosine similarity between the skill tag vector and the demand vector; The cosine similarity is used as the matching score between the R&D task flow and the candidate R&D resources.

7. The solution generation method based on integrated demand-resource matching as described in claim 1, characterized in that, The step of determining the target R&D resource among the candidate R&D resources based on the matching score includes: Determine the maximum value of the matching scores; The candidate R&D resource corresponding to the maximum matching score is selected as the target R&D resource.

8. The solution generation method based on integrated demand-resource matching as described in claim 1, characterized in that, After determining the target R&D resource among the candidate R&D resources based on the matching score, the process further includes: If it is determined that there is a gap in the R&D capabilities of the target R&D resources, R&D resource data is extracted from a preset multimodal knowledge base; The R&D resource data corresponding to the R&D capability gap will be merged into the target R&D resources.

9. The solution generation method based on integrated demand-resource matching as described in claim 1, characterized in that, Before performing feature extraction on the multimodal data to obtain the demand element vector, the method further includes: The multimodal data is semantically segmented to separate statements containing requirement parameters from redundant statements.

10. A solution generation system based on integrated demand-resource matching, characterized in that, The system includes a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of the method as described in any one of claims 1 to 9.

Images