Intelligent bid invitation purchasing method and device

Abstract

The invention discloses an intelligent bid-invitation purchasing method and device. The device comprises an infrastructure module, a data resource module and a platform supporting module. The infrastructure module is connected with the data resource module and is used for providing high-performance computing power, training power, reasoning power and a network environment; the data resource module is connected with the platform support module and is used for realizing data circulation and data storage service of the bidding data set; and the platform support module is used for realizing bid invitation evaluation on the acquired bidding document information and the bidding document information to obtain an evaluation result value of the bidding document information.

Description

Technical Field

[0001] The present invention relates to the fields of data processing and bidding and procurement evaluation, and in particular to an intelligent bidding and procurement method and device. Background Art

[0002] In traditional bidding and procurement processes, bidding evaluation primarily relies on manual review of bid and tender documents, which presents numerous challenges. On the one hand, manual review is inefficient. Faced with a large number of technical indicators and complex document content, it is difficult to accurately complete the evaluation work in a short period of time, resulting in a prolonged bidding cycle. On the other hand, manual evaluation is highly subjective, and different evaluators may have different understandings and evaluation criteria for the same technical indicators. This can easily lead to inconsistent evaluation results, impacting the fairness and impartiality of the bidding process. Furthermore, with the continuous development of bidding and procurement, the amount of data has exploded. Traditional evaluation methods struggle to effectively process and utilize this massive amount of data, failing to fully leverage the data's value to support more accurate bidding decisions. Summary of the Invention

[0003] The present invention mainly solves the problem of how to achieve accurate and rapid decision-making in the bidding and procurement process based on historical data of bidding and procurement business. The present invention discloses an intelligent bidding and procurement method and device.

[0004] In a first aspect of an embodiment of the present invention, an intelligent bidding and procurement device is disclosed, comprising: an infrastructure module, a data resource module, and a platform support module;

[0005] The infrastructure module is connected to the data resource module and is used to provide high-performance computing power, training computing power, inference computing power and network environment;

[0006] The data resource module is connected to the platform support module and is used to realize the data flow and data storage service of bidding data resources;

[0007] The platform support module is used to implement bidding evaluation on the acquired bidding document information and tender document information, and obtain the evaluation result value of the bidding document information.

[0008] The infrastructure module includes CPU servers, GPU servers, network servers, and cloud platforms; the GPU servers are deployed with training clouds and inference clouds;

[0009] The platform support module includes an AI large model base, an AI development layer, a general AI component middle platform, a service layer, and an access layer;

[0010] The AI big model base is used to provide a multi-source big model engine capability base for intelligent bidding;

[0011] The AI development layer includes a prompt word submodule, an AI operator library submodule, an AI framework submodule, and a large model reasoning submodule;

[0012] The general AI component middle platform is constructed by integrating and arranging services for multi-source large model engines, and has intelligent question-answering function components, intelligent authentication function components, intelligent interaction function components, intelligent search function components, text generation components, and text parsing components.

[0013] The access layer is used to provide unified external gateway API interface services and SDK services.

[0014] The service layer is used to receive bid document information and tender document information, perform text parsing on the bid document information and tender document information using a text parsing component, and obtain a technical indicator capability information set and a technical indicator requirement information set respectively; perform evaluation processing on the technical indicator capability information set and the technical indicator requirement information set to obtain an evaluation result value of the bid document information; the technical indicator capability information set includes a capability value sequence of each technical indicator; the technical indicator requirement information set includes a requirement value sequence of each technical indicator.

[0015] The evaluating and processing the technical indicator capability information set and the technical indicator requirement information set to obtain the evaluation result value of the bidding document information includes:

[0016] The technical indicator capability information set and the technical indicator requirement information set are represented as a capability information matrix and an indicator requirement matrix respectively; the row vector of the capability information matrix is a capability value sequence of a technical indicator; the row vector of the indicator requirement matrix is a requirement value sequence of a technical indicator;

[0017] Performing difference calculation on the capability information matrix and the indicator requirement matrix to obtain a difference value matrix;

[0018] The difference value matrix is quantitatively evaluated to obtain an evaluation result value of the bidding document information.

[0019] The data resource module consists of a data collection and aggregation system, a big data platform, and bidding data resources;

[0020] The data collection and aggregation system is connected to the big data platform and is the external interface of the data resource module. It is used to implement and configure the data collection engine, collect and aggregate historical bidding data, historical bid winning data, historical bid document data, and historical bidding process data, and connect to relevant data from other modules to build basic supporting data and complete the collection and compilation of bidding data resources.

[0021] The bidding data resources are stored in the data collection and aggregation system; the bidding data resources include bidding history data, bid winning history data, bid document history data, and bidding process history data;

[0022] The big data platform is connected to the cloud platform of the infrastructure module and is used to implement database storage management and data analysis and mining services. According to user data demand information, corresponding data is extracted from the data collection and aggregation system and sent to the platform support module.

[0023] A second aspect of the embodiments of the present invention discloses an intelligent bidding and procurement method, which is implemented using the intelligent bidding and procurement device, and includes:

[0024] S1, using the infrastructure module, provides high-performance computing power, training computing power, inference computing power, and network environment for the implementation of intelligent bidding and procurement methods;

[0025] S2, using the data resource module to realize the data flow and data storage service of bidding data resources;

[0026] S3, using the platform support module to implement bidding evaluation of the acquired bidding document information and tender document information, and obtain the evaluation result value of the bidding document information.

[0027] The platform support module is used to implement bidding evaluation of the acquired bidding document information and tender document information, and obtain the evaluation result value of the bidding document information, which is implemented by using the service layer of the platform support module, including:

[0028] Receiving bid document information and tender document information, and performing text parsing processing on the bid document information and tender document information using a text parsing component to obtain a technical indicator capability information set and a technical indicator requirement information set respectively;

[0029] The technical indicator capability information set and the technical indicator requirement information set are evaluated and processed to obtain the evaluation result value of the bidding document information; the technical indicator capability information set includes a capability value sequence of each technical indicator; the technical indicator requirement information set includes a requirement value sequence of each technical indicator.

[0030] The evaluating and processing the technical indicator capability information set and the technical indicator requirement information set to obtain the evaluation result value of the bidding document information includes:

[0031] The technical indicator capability information set and the technical indicator requirement information set are represented as a capability information matrix and an indicator requirement matrix respectively; the row vector of the capability information matrix is a capability value sequence of a technical indicator; the row vector of the indicator requirement matrix is a requirement value sequence of a technical indicator;

[0032] Performing difference calculation processing on the capability information matrix and the indicator requirement matrix to obtain a difference value matrix;

[0033] The difference value matrix is quantitatively evaluated to obtain an evaluation result value of the bidding document information.

[0034] The expression for the difference calculation is:

[0035]

[0036] Among them, z ij is the difference value matrix, b ij and n ij are the elements in the i-th row and j-th column of the capability information matrix and the indicator requirement matrix respectively.

[0037] The calculation expression of the quantitative bid evaluation process is:

[0038]

[0039] Among them, α j and β j are the jth first weighting factor and the jth second weighting factor, pb is the evaluation result value of the bidding document information, P and N are the row dimension and column dimension of the capability information matrix, respectively. and are the minimum and maximum values of the jth column of the difference value matrix, respectively.

[0040] The beneficial effects of the present invention are:

[0041] The intelligent bidding and procurement device of the present invention provides powerful computing power support through the infrastructure module, including high-performance computing power, training computing power, and reasoning computing power, as well as a stable network environment, laying the foundation for the efficient operation of the entire bidding and evaluation process. The hybrid cloud architecture it adopts realizes the organic combination of public cloud for training and private cloud for reasoning, and controls costs to the greatest extent while ensuring data security and privacy, and improves resource utilization efficiency. The data resource module can realize efficient data flow and reliable data storage services for bidding data sets, ensuring data integrity and availability, and providing a high-quality data foundation for subsequent bidding and evaluation.

[0042] The platform support module fully leverages the advantages of AI technology. The AI big model base integrates multi-source big model engines, including commercial and open-source big models, providing a powerful multi-source big model capability foundation for intelligent bidding. It also features service orchestration and scheduling capabilities, enabling flexible scheduling and combination of models based on bidding requirements, improving evaluation flexibility and adaptability. The rich submodules of the AI development layer, such as the prompt word submodule, AI operator library submodule, AI framework submodule, and big model inference submodule, as well as various functional components of the general AI component middleware, such as intelligent question-and-answer components, intelligent verification components, intelligent interaction components, intelligent search components, text generation components, and text parsing components, provide comprehensive intelligent support for bidding evaluation. They automatically parse and process bid document information, rapidly extract technical indicator capability information sets and technical indicator requirement information sets, and evaluate these information sets using scientific evaluation algorithms to obtain accurate bid document information evaluation results. This significantly improves the efficiency and accuracy of bidding evaluation, reduces manual intervention, reduces evaluation costs, and enhances the fairness and impartiality of bidding.

[0043] The service layer calls the large model reasoning submodule during the evaluation process, making full use of the powerful reasoning ability of the large model and further improving the reliability of the evaluation results. The unified external gateway API interface service and SDK service provided by the access layer facilitate the unified authentication, access, current limiting and fuse mechanism of the application, enhance the scalability and compatibility of the device, and can better adapt to different bidding and procurement scenarios and changes in business needs. In short, the intelligent bidding and procurement device of the present invention effectively solves the problems of low efficiency, strong subjectivity, and difficulty in processing massive data in the traditional bidding evaluation process, and provides an efficient, accurate and intelligent evaluation solution for the bidding and procurement field. It has significant innovation and practicality, and can provide strong support for the digital transformation and intelligent upgrading of bidding and procurement business. It has broad application prospects and important practical significance. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 1. It is a diagram of the device composition of the method of the present invention;

[0045] Figure 2 4 is an implementation flow chart of the method of the present invention. DETAILED DESCRIPTION

[0046] In order to better understand the content of the present invention, an embodiment is given here.

[0047] Figure 1 This is a diagram of the device composition of the method of the present invention; Figure 2 4 is an implementation flow chart of the method of the present invention.

[0048] In a first aspect of an embodiment of the present invention, an intelligent bidding and procurement device is disclosed, comprising: an infrastructure module, a data resource module, and a platform support module;

[0049] The infrastructure module is connected to the data resource module and is used to provide high-performance computing power, training computing power, inference computing power and network environment;

[0050] The data resource module is connected to the platform support module and is used to realize the data flow and data storage service of bidding data resources;

[0051] The platform support module is used to implement bidding evaluation on the acquired bidding document information and tender document information, and obtain the evaluation result value of the bidding document information.

[0052] The infrastructure module includes CPU servers, GPU servers, network servers, and a cloud platform. The architecture of the infrastructure module adopts a hybrid cloud architecture, namely a public cloud for training and a private cloud for inference, which can maximize cost control and efficient utilization. The training cloud and inference cloud are deployed on the GPU server.

[0053] The platform support module includes an AI large model base, an AI development layer, a general AI component middle platform, a service layer, and an access layer;

[0054] The AI large model base of the platform support module is connected to the data acquisition and aggregation system of the data resource module.

[0055] The AI big model base is connected to the AI development layer, the AI development layer is connected to the general AI component middle platform, the service layer is connected to the general AI component middle platform, and the service layer is connected to the access layer.

[0056] The AI big model base is used to provide a multi-source big model engine capability base for intelligent bidding, realizing the management and service orchestration and scheduling capabilities of multi-source big models; the multi-source big model engine includes commercial big models and open source big models;

[0057] The AI development layer includes a prompt word submodule, an AI operator library submodule, an AI framework submodule, and a large model reasoning submodule;

[0058] The general AI component middle platform is constructed by integrating and arranging services for multi-source large model engines, and has intelligent question-answering function components, intelligent authentication function components, intelligent interaction function components, intelligent search function components, text generation components, and text parsing components.

[0059] The service layer is configured to receive bid document information and tender document information, perform text parsing on the bid document information and tender document information using a text parsing component to obtain a technical indicator capability information set and a technical indicator requirement information set, respectively; and perform evaluation processing on the technical indicator capability information set and the technical indicator requirement information set to obtain an evaluation result value of the bid document information; the technical indicator capability information set includes a capability value sequence of each technical indicator; and the technical indicator requirement information set includes a requirement value sequence of each technical indicator.

[0060] The service layer evaluates and processes the technical indicator capability information set and the technical indicator requirement information set to obtain the evaluation result value of the bidding document information, which can be achieved by calling the large model reasoning submodule.

[0061] The access layer is used to provide unified external gateway API interface services and SDK services to facilitate unified application authentication, access, current limiting and circuit breaker mechanisms.

[0062] The AI development layer also includes AI plug-ins;

[0063] The evaluating and processing the technical indicator capability information set and the technical indicator requirement information set to obtain the evaluation result value of the bidding document information includes:

[0064] The technical indicator capability information set and the technical indicator requirement information set are represented as a capability information matrix and an indicator requirement matrix respectively; the row vector of the capability information matrix is a capability value sequence of a technical indicator; the row vector of the indicator requirement matrix is a requirement value sequence of a technical indicator;

[0065] Performing difference calculation on the capability information matrix and the indicator requirement matrix to obtain a difference value matrix;

[0066] The difference value matrix is quantitatively evaluated to obtain an evaluation result value of the bidding document information.

[0067] The expression for the difference calculation is:

[0068]

[0069] Among them, z ij is the difference value matrix, b ij and n ij are the elements of the i-th row and j-th column of the capability information matrix and the indicator requirement matrix respectively;

[0070] This expression uses the form of exponential function to perform nonlinear transformation on the difference between the corresponding elements of the capability information matrix and the indicator requirement matrix. ij -nij When the difference is small, the exponential part approaches 0, z ij Approaching 1, it means the difference between the two is small; when b ij -n ij When the difference is large, the exponential part is a large negative number, z ij This approach can highlight the key differences between the technical indicator capabilities of the bid document and the technical indicator requirements of the tender document, making it easier to quickly identify indicators that do not meet the requirements during subsequent bid evaluation and avoid overlooking key issues.

[0071] Compared to simple difference calculations, this expression is more sensitive to data discrepancies. In tender evaluations, subtle differences in technical indicators can have a significant impact on project implementation. The introduction of an exponential function allows even small discrepancies to be reflected in the calculation results, helping to more accurately assess the compatibility of bid and tender documents, without missing any details that could affect the tender outcome.

[0072] The calculation expression of the quantitative bid evaluation process is:

[0073]

[0074] Among them, α j and β j are the jth first weighting factor and the jth second weighting factor, pb is the evaluation result value of the bidding document information, P and N are the row dimension and column dimension of the capability information matrix, respectively. and are the minimum and maximum values of the jth column of the difference value matrix, respectively.

[0075] The calculation expression of the quantitative evaluation process is expressed by the product term This approach considers the combined impact of differences across different dimensions for the same technical indicator (row dimension P represents different dimensions). Taking the exponent and raising it to the power of P smoothly reflects the overall level of variation. The summation term combines the ratio of indicator capability to demand with the sine function to further refine the specific manifestations of variation across different dimensions. This multi-dimensional approach provides a more comprehensive assessment of bid document technical indicators, avoiding biased evaluation results due to a single-dimensional approach. The introduction of the sine function adds a nonlinear adjustment mechanism to the calculation process. In tender evaluations, the importance and impact of differences in different technical indicators may not be linearly related. The sine function flexibly adjusts based on the size and direction of the differences. When the differences are small, the changes are relatively gradual; when the differences are large, the changes are more dramatic. This more accurately reflects the actual impact of varying degrees of difference on the evaluation results, ensuring that the evaluation results are more aligned with actual business needs.

[0076] β j By directly selecting the maximum difference across different dimensions for the same technical indicator as the second weighting factor, we can quickly identify the most prominent issues with that technical indicator. During the bidding process, the maximum difference in a technical indicator can often be a key factor affecting project implementation. This approach allows us to focus on indicators most likely to present risks or non-compliance, improving the relevance and effectiveness of the assessment.

[0077] The calculation of pb comprehensively considers the maximum values of the first and second weighting factors, as well as the variance matrix. The numerator combines the weighting factors and variances using tangent and sine functions, further amplifying the impact of significant variances. The denominator balances the weighting of different technical indicators by calculating the relationship between the weighting factors and the maximum variance values. This comprehensive approach allows for a comprehensive and objective assessment of the bid document's performance across various technical indicators, resulting in a more reasonable evaluation result.

[0078] The cloud platform is connected to the CPU server, GPU server, and network server respectively, and is used to realize cloud connection and cloud login functions of the intelligent bidding and procurement device for users;

[0079] The network server is used to realize network connection of all CPU servers and GPU servers;

[0080] The CPU server and GPU server are used to provide computing resources for the intelligent bidding and procurement device.

[0081] Specifically, the network server is connected to the data collection and aggregation system of the data resource module.

[0082] The data resource module consists of a data collection and aggregation system, a big data platform and bidding data resources.

[0083] The data collection and aggregation system is connected to the big data platform and is the external interface of the data resource module. It is used to implement and configure the data collection engine, collect and aggregate historical bidding data, historical winning bid data, historical bid document data, and historical bidding process data, and connect to relevant data from other modules to construct basic supporting data and complete the collection and compilation of bidding data resources.

[0084] The bidding data resources are stored in a data collection and aggregation system; the bidding data resources include bidding history data, bid winning history data, bid document history data, and bidding process history data.

[0085] The big data platform is connected to the cloud platform of the infrastructure module and is used to implement database storage management and data analysis and mining services. According to user data demand information, the corresponding data is extracted from the data collection and aggregation system and sent to the platform support module;

[0086] The prompt word submodule is used to manage and generate prompt words that interact with the multi-source large model engine.

[0087] The function implementation process of the prompt word submodule includes:

[0088] Prompt word template management: Build a prompt word template database to store prompt word templates for various scenarios, such as templates corresponding to bidding project types and review stages; provide template addition, deletion, modification and query functions to facilitate administrators to manage according to actual needs;

[0089] Prompt word generation: Based on the bidding project information input by the user, a suitable template is selected from the template database; using natural language processing technology, the user input information is filled into the template to generate specific prompt words;

[0090] Prompt word optimization: Collect the output results of the large model and user feedback to optimize the prompt words. Use machine learning algorithms, such as reinforcement learning, to continuously adjust the parameters of the prompt words to improve the output quality of the large model.

[0091] The AI operator library submodule provides several AI algorithms and operators for building and training large models.

[0092] The implementation process of the AI operator library submodule includes: encapsulating AI algorithms and operators into reusable operators and storing them in the operator library; selecting corresponding operators from the operator library based on user needs and the characteristics of the bidding project; providing a visual interface to facilitate user combination and configuration of operators and build customized large models; optimizing the performance of operators to improve computing efficiency and accuracy. The operator library is regularly updated to introduce new algorithms and operators to adapt to the continuous development of AI technology;

[0093] The AI framework submodule provides a unified development environment to support the development and training of large models.

[0094] The implementation process of the AI framework sub-module includes: selecting popular AI frameworks, such as TensorFlow, PyTorch, etc., as the development basis; customizing and expanding the selected framework to meet the specific needs of intelligent bidding and procurement; providing model development tools, such as visual editors, code editors, etc., to facilitate users to design and implement models; utilizing bidding data resources for distributed training and parallel computing to improve the training efficiency of the model; providing model evaluation indicators and tools, such as accuracy, recall rate, F1 value, etc., for evaluating the performance of the model; optimizing and adjusting the model based on the evaluation results to improve the accuracy and stability of the model.

[0095] The large model reasoning submodule is used to apply the trained large model to the actual bidding review;

[0096] The implementation process of the large model inference submodule includes: loading the trained large model into the inference environment, such as a GPU server or cloud platform; providing model deployment tools, such as containerization technology, to facilitate rapid deployment and management of the model; receiving the user's inference request, including prompt words and bid data; preprocessing the inference request, such as data cleaning and feature extraction; inputting the preprocessed request into the loaded large model for inference calculation; post-processing the inference results, such as result parsing and formatting, to generate a final evaluation report.

[0097] The general AI component middle platform is constructed by integrating and arranging services for multi-source large model engines, and has intelligent question-answering function components, intelligent authentication function components, intelligent interaction function components, intelligent search function components, text generation components, and text parsing components.

[0098] The intelligent question-answering functional component is used to perform natural language processing on questions input by users, such as word segmentation, part-of-speech tagging, named entity recognition, etc.; use semantic understanding technology to extract the key information and intention of the question; use information retrieval algorithms to retrieve relevant answers from the knowledge base based on the key information of the question; filter and sort the retrieved answers to select the most suitable answer; use natural language generation technology to convert the answers into corresponding text data, and display the text data to the user.

[0099] The intelligent verification functional component is used to collect bidding data from data sources such as bidding documents and databases; clean and preprocess the collected data to remove noise and erroneous information; extract features from the preprocessed data, such as text features, numerical features, image features, etc.; use feature engineering technology to select and extract the most representative features; use historical bidding data and annotation information to train verification models, such as machine learning models and deep learning models; evaluate and optimize the trained models to improve the accuracy and reliability of the models; input the bidding data to be verified into the trained model for verification calculations; based on the verification results, judge the authenticity and legality of the bidding data and provide corresponding prompts and suggestions.

[0100] The intelligent interactive functional component is used to implement information interaction between the user and the intelligent bidding and procurement system; the intelligent interactive functional component is used to convert the user's voice input into text using speech recognition technology. The speech recognition results are corrected and optimized to improve recognition accuracy. Natural language processing, such as word segmentation, part-of-speech tagging, and named entity recognition, is performed on the text input by the user. Semantic understanding technology is used to extract the key information and intent of the text. Corresponding interaction strategies are formulated based on the user's intent and historical interaction records. Dialogue management technology is used to achieve smoothness and coherence in the dialogue. Speech synthesis technology is used to convert the system's responses into speech output. The speech synthesis results are optimized to improve the naturalness and intelligibility of the speech.

[0101] The intelligent search function component is used to assist users in obtaining information corresponding to their search requirements;

[0102] The intelligent search function component is used to pre-process retrieval requirements, such as word segmentation, part-of-speech tagging, named entity recognition, etc.; use indexing technologies, such as inverted index and B-tree index, to build an index library of bidding information; receive user search requests and pre-process the requests, such as keyword extraction and query expansion; retrieve relevant bidding information from the index library according to the user's request; sort the retrieved search results and perform weighted sorting based on factors such as relevance and timeliness; use machine learning algorithms, such as ranking learning, to continuously optimize the sorting algorithm of search results; display the sorted search results to users, and provide detailed bidding information and links; support paging, filtering, sorting and other operations of search results to facilitate users to find the required information.

[0103] The text generation component is used to build a text template database to store various types of bidding text templates, such as bidding document templates, evaluation report templates, etc.; provide template addition, deletion, modification and query functions to facilitate administrators to manage according to actual needs; select a suitable template from the template database based on the bidding project information entered by the user; use natural language processing technology to fill the user input information into the template to generate specific text content; perform grammar checking and semantic analysis on the generated text content to discover and correct errors and problems in the text; use text generation algorithms, such as generative adversarial networks (GANs) and variational autoencoders (VAEs), to optimize and polish the text content to improve the quality and readability of the text.

[0104] The text parsing component cleans and preprocesses the tender text to remove noise and format information, such as HTML tags, spaces, line breaks, etc.; uses natural language processing technologies, such as word segmentation, part-of-speech tagging, named entity recognition, etc., to analyze and annotate the text; defines information extraction rules and templates based on the structure and content of the tender text; uses rule matching, machine learning and other methods to extract key information and data from the text, such as the tender project name, tender scope, bid deadline, technical indicator requirement information and technical indicator capability information; integrates and associates the extracted information and data to construct a structured representation of the tender information; uses data mining technologies, such as cluster analysis and association rule mining, to discover potential relationships and patterns between information.

[0105] The information retrieval algorithms used include TF-IDF, BM25, etc., to improve the accuracy of answer retrieval;

[0106] The data collection engine can be implemented using Flume.

[0107] A second aspect of the embodiments of the present invention discloses an intelligent bidding and procurement method, which is implemented using the intelligent bidding and procurement device, and includes:

[0108] S1, using the infrastructure module, provides high-performance computing power, training computing power, inference computing power, and network environment for the implementation of intelligent bidding and procurement methods;

[0109] S2, using the data resource module to realize the data flow and data storage service of bidding data resources;

[0110] S3, using the platform support module to implement bidding evaluation of the acquired bidding document information and tender document information, and obtain an evaluation result value of the bidding document information;

[0111] The platform support module is used to implement bidding evaluation of the acquired bidding document information and tender document information, and obtain the evaluation result value of the bidding document information, which is implemented by using the service layer of the platform support module, including:

[0112] Receiving bid document information and tender document information, and performing text parsing processing on the bid document information and tender document information using a text parsing component to obtain a technical indicator capability information set and a technical indicator requirement information set respectively;

[0113] Evaluating the technical indicator capability information set and the technical indicator requirement information set to obtain an evaluation result value of the bidding document information; the technical indicator capability information set includes a capability value sequence of each technical indicator; the technical indicator requirement information set includes a requirement value sequence of each technical indicator;

[0114] The evaluating and processing the technical indicator capability information set and the technical indicator requirement information set to obtain the evaluation result value of the bidding document information includes:

[0115] The technical indicator capability information set and the technical indicator requirement information set are represented as a capability information matrix and an indicator requirement matrix respectively; the row vector of the capability information matrix is a capability value sequence of a technical indicator; the row vector of the indicator requirement matrix is a requirement value sequence of a technical indicator;

[0116] Performing difference calculation on the capability information matrix and the indicator requirement matrix to obtain a difference value matrix;

[0117] The difference value matrix is quantitatively evaluated to obtain an evaluation result value of the bidding document information.

[0118] The expression for the difference calculation is:

[0119]

[0120] Among them, z ij is the difference value matrix, b ij and n ij are the elements of the i-th row and j-th column of the capability information matrix and the indicator requirement matrix respectively;

[0121] The calculation expression of the quantitative bid evaluation process is:

[0122]

[0123] Among them, α j and β jare the jth first weighting factor and the jth second weighting factor, pb is the evaluation result value of the bidding document information, P and N are the row dimension and column dimension of the capability information matrix, respectively. and are the minimum and maximum values of the jth column of the difference value matrix, respectively.

[0124] The performing of quantitative bid evaluation processing on the difference value matrix to obtain an evaluation result value of the bid document information includes:

[0125] Performing singular value calculation on the difference value matrix to obtain a singular value set and a singular value vector set;

[0126] Performing cross-covariance calculation on the difference value matrix to obtain a cross-covariance matrix; the elements in the i-th row and j-th column of the cross-covariance matrix are cross-covariance values of the i-th row vector and the j-th row vector of the difference value matrix;

[0127] Performing eigenvalue calculation processing on the cross-covariance matrix to obtain an eigenvalue set and an eigenvalue vector set;

[0128] Performing evaluation and calculation processing on the singular value set, singular value vector set, eigenvalue set, and eigenvalue vector set to obtain an evaluation result value of the bidding document information;

[0129] The expression for the evaluation calculation process is:

[0130]

[0131] Among them, h i is the i-th weighting factor, α i is the i-th element of the singular value set, β ij is the jth element of the i-th singular value vector in the singular value vector set, γ i is the i-th element of the eigenvalue set, δ i is the mean of the i-th eigenvalue vector in the eigenvalue vector set, β i is the mean of the i-th singular value vector in the singular value vector set, T i () represents the i-th order polynomial of the first kind Chebyshev polynomial, and pb represents the evaluation result value of the bidding document information.

[0132] According to a third aspect of an embodiment of the present invention, an intelligent bidding and procurement method and apparatus is disclosed, the apparatus comprising:

[0133] a memory storing executable program code;

[0134] a processor coupled to the memory;

[0135] The processor calls the executable program code stored in the memory to execute the intelligent bidding and procurement method.

[0136] According to a fourth aspect of an embodiment of the present invention, a computer-storable medium is disclosed, wherein the computer-storable medium stores computer instructions. When the computer instructions are called by a computer, they are used to execute the intelligent bidding and procurement method.

[0137] According to a fifth aspect of the embodiments of the present invention, an information data processing terminal is disclosed, which is used to implement the intelligent bidding and procurement method.

[0138] The foregoing is merely an embodiment of the present invention and is not intended to limit the present invention. It will be apparent to those skilled in the art that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are intended to be included within the scope of the claims of the present invention.

Claims

1. An intelligent bidding and procurement device, characterized in that: include: Infrastructure module, data resource module, platform support module; The infrastructure module is connected to the data resource module and is used to provide high-performance computing power, training computing power, inference computing power and network environment; The data resource module is connected to the platform support module and is used to realize the data flow and data storage service of bidding data resources; The platform support module is used to implement bidding evaluation on the acquired bidding document information and tender document information, and obtain the evaluation result value of the bidding document information.

2. The intelligent bidding and procurement device according to claim 1, characterized in that: The infrastructure module includes CPU servers, GPU servers, network servers, and cloud platforms; the GPU servers are deployed with training clouds and inference clouds; The platform support module includes an AI large model base, an AI development layer, a general AI component middle platform, a service layer, and an access layer; The AI big model base is used to provide a multi-source big model engine capability base for intelligent bidding; The AI development layer includes a prompt word submodule, an AI operator library submodule, an AI framework submodule, and a large model reasoning submodule; The general AI component middle platform is constructed by integrating and arranging services for multi-source large model engines, and has intelligent question-answering function components, intelligent authentication function components, intelligent interaction function components, intelligent search function components, text generation components, and text parsing components. The access layer is used to provide unified external gateway API interface services and SDK services.

3. The intelligent bidding and procurement device according to claim 2, characterized in that: The service layer is used to receive bid document information and tender document information, and perform text parsing on the bid document information and tender document information using a text parsing component to obtain a set of technical indicator capability information and a set of technical indicator requirement information respectively; Evaluating the technical indicator capability information set and the technical indicator requirement information set to obtain an evaluation result value of the bidding document information; the technical indicator capability information set includes a capability value sequence of each technical indicator; The technical indicator requirement information set includes a requirement value sequence for each technical indicator.

4. The intelligent bidding and procurement device according to claim 3, characterized in that: The evaluating and processing the technical indicator capability information set and the technical indicator requirement information set to obtain the evaluation result value of the bidding document information includes: The technical indicator capability information set and the technical indicator requirement information set are represented as a capability information matrix and an indicator requirement matrix respectively; the row vector of the capability information matrix is a capability value sequence of a technical indicator; the row vector of the indicator requirement matrix is a requirement value sequence of a technical indicator; Performing difference calculation on the capability information matrix and the indicator requirement matrix to obtain a difference value matrix; The difference value matrix is quantitatively evaluated to obtain an evaluation result value of the bidding document information.

5. The intelligent bidding and procurement device according to claim 1, characterized in that: The data resource module consists of a data collection and aggregation system, a big data platform, and bidding data resources; The data collection and aggregation system is connected to the big data platform and is the external interface of the data resource module. It is used to implement and configure the data collection engine, collect and aggregate historical bidding data, historical bid winning data, historical bid document data, and historical bidding process data, and connect to relevant data from other modules to build basic supporting data and complete the collection and compilation of bidding data resources. The bidding data resources are stored in the data collection and aggregation system; The bidding data resources include historical bidding data, historical winning bid data, historical bid document data, and historical bidding process data; The big data platform is connected to the cloud platform of the infrastructure module and is used to implement database storage management and data analysis and mining services. According to user data demand information, corresponding data is extracted from the data collection and aggregation system and sent to the platform support module.

6. An intelligent bidding and procurement method, implemented using the intelligent bidding and procurement device according to any one of claims 1 to 5, characterized in that: include: S1, using the infrastructure module, provides high-performance computing power, training computing power, inference computing power, and network environment for the implementation of intelligent bidding and procurement methods; S2, using the data resource module to realize the data flow and data storage service of bidding data resources; S3, using the platform support module to implement bidding evaluation of the acquired bidding document information and tender document information, and obtain the evaluation result value of the bidding document information.

7. The intelligent bidding and procurement method according to claim 6, characterized in that: The platform support module is used to implement bidding evaluation of the acquired bidding document information and tender document information, and obtain the evaluation result value of the bidding document information, which is implemented by using the service layer of the platform support module, including: Receiving bid document information and tender document information, and performing text parsing processing on the bid document information and tender document information using a text parsing component to obtain a technical indicator capability information set and a technical indicator requirement information set respectively; The technical indicator capability information set and the technical indicator requirement information set are evaluated and processed to obtain the evaluation result value of the bidding document information; the technical indicator capability information set includes a capability value sequence of each technical indicator; the technical indicator requirement information set includes a requirement value sequence of each technical indicator.

8. The intelligent bidding and procurement method according to claim 7, characterized in that: The evaluating and processing the technical indicator capability information set and the technical indicator requirement information set to obtain the evaluation result value of the bidding document information includes: The technical indicator capability information set and the technical indicator requirement information set are represented as a capability information matrix and an indicator requirement matrix respectively; the row vector of the capability information matrix is a capability value sequence of a technical indicator; the row vector of the indicator requirement matrix is a requirement value sequence of a technical indicator; Performing difference calculation processing on the capability information matrix and the indicator requirement matrix to obtain a difference value matrix; The difference value matrix is quantitatively evaluated to obtain an evaluation result value of the bidding document information.

9. The intelligent bidding and procurement method according to claim 7, characterized in that: The expression for the difference calculation is: Among them, z ij is the difference value matrix, b ij and n ij are the elements in the i-th row and j-th column of the capability information matrix and the indicator requirement matrix respectively.

10. The intelligent bidding and procurement method according to claim 7, characterized in that: The calculation expression of the quantitative bid evaluation process is: Among them, α j and β j are the jth first weighting factor and the jth second weighting factor, pb is the evaluation result value of the bidding document information, P and N are the row dimension and column dimension of the capability information matrix, respectively. and are the minimum and maximum values of the jth column of the difference value matrix, respectively.

Images