A material supply chain intelligent recommendation method and system for the construction industry

Abstract

This invention relates to the field of building materials supply chain technology, specifically to an intelligent recommendation method and system for building materials supply chains, comprising the following steps: suppliers upload reports to a platform, the platform generates a data package of information and synchronizes it to a procurement cloud platform; the procurement cloud platform receives the data, generates industry tags, calculates indicator scores, and builds tables to form a classification index system; buyers submit requirements, the platform parses the requirements and executes a screening mechanism to screen suppliers and generate a list of solutions based on comprehensive scoring; buyers evaluate and select suppliers to generate orders, and the platform updates inventory and logistics information. This invention comprehensively records supplier information and constructs a classification index through multi-dimensional data processing technology, achieving accurate screening; it uses dynamic scoring and hierarchical screening technology to flexibly adjust screening criteria; it breaks down the global problem through grouping and parallel computing technology to improve computational efficiency; and it objectively evaluates suppliers through comprehensive scoring technology, ultimately improving procurement efficiency and optimizing supply chain matching.

Description

Technical Field

[0001] This invention relates to the field of building materials supply chain technology, and more specifically, to an intelligent recommendation method and system for building materials supply chains. Background Technology

[0002] In today's construction industry, materials supply chain management is a crucial element in ensuring the smooth progress of projects. However, existing methods and systems for managing construction materials supply chains have many problems that urgently need to be addressed, leading to low procurement efficiency, poor supply chain matching, and difficulties in guaranteeing project quality and schedule.

[0003] Firstly, regarding supplier information management, the data uploaded by suppliers in traditional methods is often insufficiently detailed and accurate. Many systems only record basic supplier information, with vague descriptions of their business scope, failing to clearly cover building material categories and specifications. This makes it difficult for procurement cloud platforms to accurately locate suitable suppliers. For example, when searching for suppliers of specific types of rebar, due to inaccurate information, buyers need to spend a significant amount of time and effort sifting through a massive number of suppliers, resulting in extremely low efficiency. Simultaneously, inventory and logistics information recording is not detailed enough, failing to record current inventory levels, average monthly shipments, and inventory status at the SKU level. There is also a lack of comprehensive understanding of logistics coverage areas, average delivery times, and carrier lists. This makes it difficult for buyers to accurately predict replenishment points, easily leading to work stoppages due to material shortages or inventory backlogs, increasing costs, and making it difficult to assess transportation reliability, potentially causing project delays due to logistical issues.

[0004] Secondly, the existing system lacks an effective classification index system for supplier classification and screening. The industry labeling is simplistic and crude, failing to meet the diverse and varied characteristics of building materials. For example, classification by broad building material categories lacks detail down to specific subcategories and models, making it difficult for buyers to conduct precise searches and reducing the matching efficiency of the material supply chain. Furthermore, the screening mechanism typically uses fixed rules and standards, unable to flexibly adjust to changes in procurement needs. It cannot target different projects with varying requirements for building material inventory, logistics, quality, and price, resulting in selected suppliers that do not match actual needs, significantly reducing the effectiveness and practicality of recommended solutions.

[0005] Furthermore, with the increasing number of suppliers, traditional processing methods suffer from severe bottlenecks in computational efficiency. Processing all suppliers directly involves enormous computational demands, resulting in lengthy screening processes that fail to meet real-time requirements. Simultaneously, the existing system lacks clear distinctions and corresponding processing mechanisms for different types of procurement needs, such as rigid, flexible, and combined demands. This fails to adequately consider various scenarios that may arise in actual procurement, limiting the buyer's choices and reducing the success rate of procurement.

[0006] Furthermore, existing supplier evaluation systems are neither comprehensive nor objective. They often focus only on one or a few aspects, such as price or quality, while neglecting other important factors, such as inventory matching and logistics timeliness. This makes it difficult for buyers to conduct a comprehensive and accurate evaluation of suppliers, hindering the selection of the most suitable ones. Moreover, the evaluation results cannot be effectively fed back to suppliers, failing to prompt them to improve their management and service strategies.

[0007] In view of this, a method and system for intelligent recommendation of material supply chains in the construction industry is proposed. Summary of the Invention

[0008] The purpose of this invention is to provide an intelligent recommendation method and system for the material supply chain in the construction industry, so as to improve procurement efficiency, optimize supply chain matching, ensure project quality and schedule, and promote suppliers to improve service quality.

[0009] To solve the above-mentioned technical problems, this invention provides an intelligent recommendation method for the material supply chain in the construction industry, comprising the following steps:

[0010] S1. Suppliers upload reports containing business scope, inventory and logistics information to the supplier management platform. The supplier management platform generates a supplier profile data package with the business license number as the primary key and synchronizes it to the procurement cloud platform.

[0011] S2. After receiving the supplier profile data package, the procurement cloud platform generates industry tags based on keywords in the business scope, calculates the index scores of inventory matching degree, logistics timeliness and material quality according to preset rules, and forms inventory matching degree table, logistics timeliness table and material quality table respectively in descending order of index scores. At the same time, it forms a price table based on the selling price and establishes a classification index system by material type.

[0012] S3. After the buyer submits its procurement requirements, the following screening process is performed: the supplier list is split using a hash algorithm and processed in parallel by distributed nodes; the screening rules and weight coefficients are adjusted in real time according to the requirements parameters to select suppliers that meet rigid, flexible, and combined requirements; the suppliers are comprehensively scored and a list of solutions is formed in descending order of the comprehensive scores.

[0013] S4. The buyer evaluates and selects from the list of options, and after confirmation, generates a final purchase order. The procurement cloud platform updates inventory and logistics information simultaneously.

[0014] As a further improvement to this technical solution, in S1, the scope of business includes building material categories and specifications; inventory is recorded at the SKU level, including current inventory quantity, average monthly shipment quantity, and inventory status; logistics information includes coverage area, average transportation time, and carrier list; the supplier management platform automatically verifies the data integrity of the reports, and triggers a manual review process when the data missing rate in the reports exceeds the preset value A.

[0015] As a further improvement to this technical solution, in step S1, a hash value is generated using the business license number as the primary key; the data in the supplier profile data package includes business scope, logistics information, inventory, upload timestamp, and data version number. When the inventory change is greater than the preset value B or the logistics information is updated, the supplier profile data package is automatically updated.

[0016] As a further improvement to this technical solution, in step S2, the industry classification label is generated based on the building material category entities in the business scope. At the same time, the label is constructed in three layers according to the building material category, building material subcategory, and building material model.

[0017] As a further improvement to this technical solution, in S2, the index score of inventory matching degree = current inventory quantity / average daily shipment volume of similar inventory on the procurement cloud platform × 100%, the index score of logistics timeliness = ∑ (number of cities covered in the region × timeliness coefficient) / total number of cities covered, and the index score of material quality is based on historical quality inspection data.

[0018] The classification index system is based on building material models, and separate tables are established for inventory matching, logistics timeliness, material quality, and price.

[0019] As a further improvement to this technical solution, in S3, the screening is specifically as follows:

[0020] Based on the procurement requirements parameters, data screening is carried out on the inventory matching table, logistics timeliness table, material quality table and price table with the help of dynamic scoring nodes. The parts of each table that meet the procurement requirements parameters are extracted to obtain multiple candidate supplier lists and complete the preliminary screening.

[0021] For each candidate supplier list, the hash value modulo N is taken as the group identifier. The candidate suppliers are assigned to the corresponding group identifier according to the corresponding hash value and stored independently to obtain the group directory of each candidate supplier list. In the modulo N grouping algorithm, when the number of suppliers exceeds the preset threshold, the hierarchical hashing strategy is automatically triggered.

[0022] In multiple candidate supplier lists, computing nodes with corresponding computing power are allocated according to the number of suppliers in the same group directory. After filtering irrelevant groups, the intersection calculation is performed on the group directories of each candidate supplier list of the same type to obtain suppliers that exist in multiple groups of the same type, which are then used as supplier groups.

[0023] As a further improvement to this technical solution, the supplier's exclusive data in each of the inventory matching table, logistics timeliness table, material quality table and price table is extracted independently, and corresponding information nodes are configured for the supplier's exclusive data in each table.

[0024] The dynamic scoring node comprises a demand receiving end, a parsing and mapping module, and a data interface end. The demand receiving end receives procurement demand parameters. The parsing and mapping module has filtering rule bases corresponding to inventory matching tables, logistics timeliness tables, material quality tables, and price tables, respectively. The parsing and mapping module uses natural language processing technology to parse the procurement demand parameters, transforming them into specific data filtering conditions, and then stores these conditions in the corresponding table's filtering rule base. The data interface end interacts with all information nodes in a single table, filtering the specific data bound to each information node based on the filtering conditions in the corresponding table's filtering rule base within the parsing and mapping module, thus obtaining information nodes that meet the filtering conditions.

[0025] As a further improvement to this technical solution, in S3, the conditions for rigid demand are: inventory ≥ demand, logistics ≤ time limit, quality ≥ standard, and price ≤ budget; the conditions for flexible demand are: inventory < demand but can be allocated, logistics ≤ time limit, quality ≥ standard, and price ≤ budget; the conditions for combined demand are: inventory of a single supplier ≥ demand × preset value C, logistics ≤ time limit, quality ≥ standard, and price ≤ budget × preset value D. By changing the procurement demand parameters and adjusting the insertion position of the dynamic scoring node in each table, different lists of candidate suppliers can be obtained; the comprehensive score = inventory matching index score × E + logistics timeliness index score × F + material quality index score × G + price index score × H, where E + F + G + H = 1.

[0026] A smart recommendation system for the material supply chain in the construction industry, wherein the smart recommendation system for the material supply chain in the construction industry is used to implement the above-mentioned smart recommendation method for the material supply chain in the construction industry, comprising:

[0027] The supplier management platform receives supplier reports and extracts information, generates supplier profile data packages with hash values ​​using business license numbers, and verifies data integrity.

[0028] The procurement cloud platform, connected to the supplier management platform, includes a preprocessing module, a screening mechanism, and a sorting module. The preprocessing module receives supplier profile data packages, generates industry classification tags based on the information within the data packages, calculates scores for inventory matching, logistics timeliness, and material quality according to preset rules, and generates inventory matching, logistics timeliness, and material quality tables in descending order of score. Simultaneously, it generates a price table based on selling prices and establishes a classification index system by material type. The screening mechanism matches corresponding industry classification tags to material types and sequentially filters suppliers that meet rigid, flexible, and combined needs based on procurement requirements. The sorting module performs a comprehensive evaluation of suppliers and generates a list of solutions for rigid, flexible, and combined needs in descending order of the comprehensive score.

[0029] The determination module connects to the procurement cloud platform and is used to display a list of solutions for rigid demand, flexible demand, and combined demand to the buyer. The buyer compares and selects the solutions and generates a final order. The procurement cloud platform updates the inventory and logistics status in real time.

[0030] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0031] 1. In this intelligent recommendation method and system for the material supply chain in the construction industry, dynamic scoring nodes are inserted into four core business tables: inventory matching degree, logistics timeliness, material quality, and price. By adjusting the position of the dynamic scoring nodes, multi-dimensional combination filtering is supported, covering all scenarios from rigid to flexible needs, significantly improving the flexibility of filtering. At the same time, through a hierarchical filtering strategy, it covers all scenarios from strict matching to flexible combination. Combined with the weight adjustment of dynamic scoring nodes, it provides buyers with diversified recommendation solutions.

[0032] 2. In this intelligent recommendation method and system for the material supply chain in the construction industry, hash grouping is used to decompose the global problem into independent group calculations. Combined with distributed parallel processing, the computational efficiency is reduced, breaking through the efficiency bottleneck of traditional screening.

[0033] 3. The intelligent recommendation method and system for the material supply chain in the construction industry achieves precise matching from material type to specific model through fine-grained classification index. Combined with dynamic scoring nodes, it ensures that the screening results are highly aligned with the details of the procurement needs, such as model and specifications. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of the method flow of the present invention;

[0035] Figure 2 This is a schematic diagram of the system structure of the present invention. Detailed Implementation

[0036] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all secondary embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0037] Currently, existing methods and systems for managing the construction materials supply chain have significant shortcomings in areas such as supplier information management, classification and screening mechanisms, computational efficiency, and supplier evaluation, failing to meet the growing demands for refined management and efficient operation in the construction industry. Therefore, there is an urgent need for a new, more intelligent, and efficient intelligent recommendation method and system for the construction materials supply chain to improve procurement efficiency, optimize supply chain matching, ensure project quality and schedule, and encourage suppliers to improve service quality.

[0038] In view of this, please refer to Figure 1 As shown, one of the objectives of this invention is to provide an intelligent recommendation method for the material supply chain in the construction industry, which includes the following steps:

[0039] S1. Suppliers upload reports containing business scope, inventory and logistics information to the supplier management platform. The supplier management platform generates a supplier profile data package with the business license number as the primary key and synchronizes it to the procurement cloud platform.

[0040] S2. After receiving the supplier profile data package, the procurement cloud platform generates industry tags based on keywords in the business scope, calculates the index scores of inventory matching degree, logistics timeliness and material quality according to preset rules, and forms inventory matching degree table, logistics timeliness table and material quality table respectively in descending order of index scores. At the same time, it forms a price table based on the selling price and establishes a classification index system by material type.

[0041] S3. After the buyer submits its procurement requirements, the following screening process is performed: the supplier list is split using a hash algorithm and processed in parallel by distributed nodes; the screening rules and weight coefficients are adjusted in real time according to the requirements parameters to select suppliers that meet rigid, flexible, and combined requirements; the suppliers are comprehensively scored and a list of solutions is formed in descending order of the comprehensive scores.

[0042] S4. The buyer evaluates and selects from the list of options, and after confirmation, generates a final purchase order. The procurement cloud platform updates inventory and logistics information simultaneously.

[0043] This intelligent recommendation method for the construction industry's material supply chain inserts dynamic scoring nodes into four core business tables: inventory matching degree, logistics timeliness, material quality, and price. By adjusting the position of these dynamic scoring nodes, it supports multi-dimensional combination filtering, covering all scenarios from rigid to flexible needs, significantly improving filtering flexibility. At the same time, through a hierarchical filtering strategy, it covers all scenarios from strict matching to flexible combination. Combined with the weight adjustment of dynamic scoring nodes, it provides buyers with diversified recommendation solutions. Furthermore, by using hash grouping to decompose the global problem into independent group calculations, combined with distributed parallel processing, it compresses computational efficiency and breaks through the efficiency bottleneck of traditional filtering.

[0044] Given the extremely precise requirements of building materials in construction projects, different architectural design and construction stages require specific categories and corresponding specifications of building materials. For example, the load-bearing structure of high-rise buildings must use steel bars of a specific strength grade that meet national standards (such as HRB400E). Only by clearly defining the categories and specifications of building materials can the procurement cloud platform accurately locate suitable suppliers, achieving a "needle in a haystack" screening from a massive number of suppliers, improving procurement efficiency and the matching degree of project quality. Therefore, in step S1, the scope of business includes building material categories and specifications.

[0045] Due to the large scale of construction projects and the long and fluctuating material consumption cycles, taking the construction of large commercial complexes as an example, the average monthly shipment volume of main materials such as cement and steel varies significantly at different construction stages. Knowing the current inventory level and average monthly shipment volume in real time allows buyers to accurately predict replenishment points, avoid work stoppages due to material shortages or inventory backlogs, and reduce costs. Inventory status (such as normal, slow-moving, and stockout warnings) provides suppliers with intuitive data for managing inventory and buyers for adjusting procurement plans. Therefore, inventory is recorded at the SKU level, including current inventory level, average monthly shipment volume, and inventory status.

[0046] Meanwhile, because the transportation of building materials is limited by site and construction period, there are stringent requirements for logistics coverage and timeliness. For materials like concrete with limited initial setting time, it is necessary to ensure that the supplier's logistics can deliver to the construction site in a short time. The list of carriers is related to the stability of logistics and service quality. The buyer uses this information to assess the reliability of transportation, avoid the risk of project delays caused by transportation delays, and ensure project progress. Therefore, logistics information includes coverage area, average transportation time, and carrier list.

[0047] Considering that supplier-uploaded data serves as the "raw material" for intelligent recommendations across the entire supply chain, data errors or omissions can trigger a chain reaction, leading to screening biases and ineffective recommendations. In the construction industry, erroneous data can result in serious consequences such as purchasing incorrect building materials and delaying construction schedules. For example, entering incorrect steel inventory data could prevent buyers from receiving goods on time after placing orders, thus delaying key construction milestones. Therefore, the supplier management platform automatically verifies the data integrity of reports. When the data omission rate in a report exceeds a preset value A, a manual review process is triggered. This process leverages the experience of professionals to ensure high data accuracy, laying a solid foundation for subsequent precise recommendations.

[0048] Considering that the business license number is the unique legal identifier of the supplier, and has uniqueness and stability, in step S1, a hash value is generated using the business license number as the primary key. In this embodiment, a hash algorithm, such as MD5 or SHA-256, can be used. The hash algorithm can ensure that each supplier corresponds to a unique hash identifier, which makes it easier for the system to quickly and accurately identify and manage suppliers, avoid identification confusion caused by duplicate or changed supplier names, and improve the accuracy and efficiency of data management.

[0049] To comprehensively record supplier-related information and facilitate subsequent analysis and screening by the procurement cloud platform, the supplier profile data package includes business scope, logistics information, inventory, upload timestamp, and data version number. Considering that inventory and logistics information are dynamic and have a significant impact on procurement decisions, the supplier profile data package is automatically updated when inventory changes exceed a preset value B or logistics information is updated. This avoids procurement errors due to information lag. In this embodiment, database trigger technology is used. Triggers are set in the database for tables related to inventory and logistics information. When the change in inventory data exceeds the preset value B, or when there is an update operation in the logistics information table, the trigger is automatically activated, thereby executing the operation of updating the supplier profile data package.

[0050] Because the construction industry produces a wide variety of materials with varying characteristics, uses, and market demands, step S2 generates industry classification tags based on building material categories within the business scope. These tags are constructed in three layers: major building material category, subcategory, and model. This embodiment employs Natural Language Processing (NLP) technology to parse the supplier's business scope text, extracting the building material categories. Generating tags based on these categories and constructing them in three layers is crucial for more accurate supplier classification. This allows buyers to more easily find suitable suppliers based on material type, facilitates effective management and integration of supplier resources by the platform, improves the matching efficiency of the material supply chain, and, more importantly, meets different levels of search and classification needs. Building material categories can be used for macro-level screening, building material subcategories can further refine the scope, and building material models can achieve precise positioning, adapting to various scenarios for buyers from fuzzy queries to precise searches. In this embodiment, database technology is used to establish the association between tags and supplier files, and the three-layer tag system is stored in the form of a tree structure or hierarchical table to facilitate data query, update and maintenance.

[0051] Because the demand for building materials is diverse and large in quantity, the procurement cloud platform needs to accurately assess the supply capacity of suppliers. Therefore, in step S2, the inventory matching index is calculated as: Current inventory quantity / Average daily shipment volume of similar inventory on the procurement cloud platform × 100%. By comparing the current inventory quantity with the average daily shipment volume of similar inventory, the degree to which the supplier's inventory meets the procurement demand can be measured, and it can be determined whether the supplier has sufficient inventory to cope with the procurement, thus avoiding the situation of insufficient supply.

[0052] Given the tight construction schedule and the critical importance of timely material delivery, the logistics timeliness index is calculated as follows: = ∑(number of cities covered in the region × timeliness coefficient) / total number of cities covered. By combining the number of cities covered in the region with the timeliness coefficient, the logistics timeliness can be calculated. This allows for a comprehensive consideration of the supplier's logistics coverage and transportation speed, making it easier for the buyer to select a supplier that can deliver materials to the construction site within the specified time, thus ensuring that the project progress is not affected by logistics.

[0053] Considering that the quality of building materials is directly related to the quality and safety of the project and is a key factor in procurement, the material quality index is scored based on historical quality inspection data. By scoring based on historical quality inspection data, the supplier's material quality level can be objectively reflected by utilizing past quality inspection results, providing the buyer with a quality reference and ensuring that the purchased materials meet the quality requirements of the construction project.

[0054] Since building material models are important identifiers for building material products, different models often have different characteristics in terms of inventory, logistics, quality, and price. In terms of inventory, the market demand and inventory turnover speed of different models of building materials vary greatly; in terms of logistics, different models may have different transportation requirements and delivery times; in terms of quality, each model has its own specific quality standards and inspection conditions; and prices vary from model to model. Therefore, the classification index system is based on building material models and establishes separate tables for inventory matching, logistics timeliness, material quality, and price. By building an index system with building material models as the core, various types of data can be managed and queried more accurately, making it easier for business personnel to quickly locate relevant information for specific building material models and meeting the needs of refined operations in supply chain management.

[0055] Due to the complex and ever-changing procurement needs of the construction industry, different projects have significantly different requirements for building material inventory, logistics, quality, and price. For example, urgent projects may place greater emphasis on logistics timeliness and the immediate supply capacity of inventory, while large, long-term projects may be more concerned with price and quality stability. Therefore, in step S3, the screening mechanism is specifically as follows:

[0056] Based on the procurement requirements parameters, data screening is performed on the inventory matching table, logistics timeliness table, material quality table, and price table using dynamic scoring nodes. The portions of each table that meet the procurement requirements parameters are extracted to obtain multiple candidate supplier lists, completing the initial screening. By inserting dynamic scoring nodes into each table, the screening criteria can be flexibly adjusted according to specific procurement requirements parameters, making the selected candidate supplier list more in line with actual needs and improving the accuracy and effectiveness of the screening.

[0057] Considering that inventory matching, logistics timeliness, material quality, and price are core independent dimensions for procurement decisions in the construction industry, and that the data structures, evaluation standards, and business logic of each dimension differ significantly (e.g., inventory focuses on quantity and turnover efficiency, logistics focuses on regional coverage and transportation timeliness), the specific data of suppliers in each of the inventory matching, logistics timeliness, material quality, and price tables are extracted independently. Corresponding information nodes are configured for the specific data of suppliers in each table. In this embodiment, the information node can be a single cloud server or a single storage port. By independently extracting the specific data of corresponding suppliers in each table and configuring the corresponding information nodes, cross-dimensional data mixing can be avoided, ensuring that the screening and scoring of each dimension are more focused and accurate, which meets the needs of the construction industry for multi-dimensional and refined management of the material supply chain. The specific data in each table means that the inventory matching table only extracts the inventory matching data corresponding to the supplier, the logistics timeliness table only extracts the logistics timeliness data corresponding to the supplier, the material quality table only extracts the material quality data corresponding to the supplier, and the price table only extracts the price data corresponding to the supplier.

[0058] Because procurement requirements often contain vague natural language descriptions (such as "high-quality steel" and "moderate price"), they need to be transformed into quantifiable filtering conditions through dynamic scoring nodes. Therefore, the dynamic scoring node includes a demand receiving end, a parsing and mapping module, and a data interface end. The demand receiving end receives procurement demand parameters. The parsing and mapping module has filtering rule bases corresponding to inventory matching tables, logistics timeliness tables, material quality tables, and price tables. Each filtering rule base matches the corresponding table by table name and is switched within the parsing and mapping module. Using natural language processing technology, the parsing and mapping module parses the procurement demand parameters, transforming them into specific data filtering conditions, and then stores these conditions in the corresponding table's filtering rule base. For example, if the procurement demand parameter is "a supplier capable of supplying at least 800 cubic meters of concrete within next week," natural language processing technology extracts keywords from the parameter, such as "concrete" (material type), "at least 800 cubic meters" (quantity requirement), and "within next week (supply time must be within 7 days)." The time frame was clearly defined, which translated into specific data filtering conditions: "the quantity of concrete in stock provided by the supplier is ≥800 cubic meters," and the inventory status is "available for immediate supply" or "available for dispatch within the next week." The data interface is used to exchange information with all information nodes in a single table for dedicated data. Based on the data filtering conditions in the corresponding table's filtering rule base in the parsing and mapping module, the dedicated data bound to the information nodes is filtered to obtain the information nodes that meet the data filtering conditions. Then, the information nodes that meet the data filtering conditions are reverse-located to extract the suppliers in the corresponding table, and finally, multiple candidate supplier lists are obtained, completing the initial screening. By setting the filtering rule base for each table in the parsing and mapping module and supporting dynamic switching, dedicated filtering rules can be matched for the business logic of different tables (such as the quantity threshold of the inventory table and the inspection standards of the quality table), realizing the precise mapping of "demand-rules-data" and solving the problem that traditional fixed rule filtering cannot adapt to complex needs.

[0059] As the number of suppliers increases, processing all candidate suppliers directly leads to enormous computational overhead and low efficiency. Therefore, grouping by hash value modulo N can distribute a large number of candidate suppliers into multiple groups. By grouping candidate suppliers, distributed storage and parallel processing of data are achieved, greatly improving the ability to process large-scale supplier data. When the number of suppliers exceeds a preset threshold, a hierarchical hashing strategy is automatically triggered to cope with dynamic changes in data scale, further optimize the grouping effect, and ensure the balance and efficiency of the grouping. The range of N is limited to the set of prime numbers [8, 32] because prime numbers can better guarantee the uniform distribution of data in hash operations, reduce hash collisions, and improve the accuracy and stability of grouping.

[0060] The hierarchical hashing strategy is automatically triggered when the number of suppliers in a single group after single-level hashing exceeds a preset threshold, such as a preset threshold of 10,000. Its core objective is to recursively split large-scale data into smaller subgroups by increasing the levels of hashing, thereby solving the problem of group imbalance in single-level hashing when data grows explosively (such as some groups having excessively large data volumes, leading to uneven load on computing nodes). This ensures that the data size of each subgroup is maintained within an efficient processing range (such as the data volume of a single group ≤ preset threshold / 10), thereby maximizing parallel computing efficiency.

[0061] Since the number of suppliers within different groups can vary significantly, a uniform allocation of computing resources could lead to resource waste or low computational efficiency. Therefore, allocating computing nodes with corresponding computing power based on the number of suppliers within the same group allows for the rational use of computing resources and improves computational efficiency. Filtering irrelevant groups reduces unnecessary computation, further improving the efficiency and accuracy of the screening process. Parallel computing fully utilizes computing resources, accelerates processing speed, and enables the system to handle complex screening tasks while meeting real-time requirements. Ultimately, it accurately identifies the supplier group that meets the criteria, providing the purchaser with high-quality supplier recommendations; the suppliers identified through intersection calculations are extracted to form the supplier group that meets the criteria.

[0062] In the construction industry, different projects have different characteristics and requirements. Rigid demand usually applies to projects with strict requirements on construction period and quality, and where there is no room for significant changes, such as key government projects and emergency rescue projects. These projects must ensure the stability and timeliness of material supply, so inventory, logistics, quality, and price must all be strictly met. Flexible demand, on the other hand, takes into account the situation in actual procurement where some suppliers may have temporary insufficient inventory, but the demand can be met by allocating resources. This requirement setting provides buyers with a certain degree of flexibility, broadens the range of suppliers to choose from, and increases the probability of successful procurement. Combined requirements are suitable for large-scale projects, where the material demand is large and a single supplier may not be able to fully meet the demand, but multiple suppliers can work together to meet the demand. By setting preset values ​​C for the inventory of a single supplier and D for the budget, buyers can be guided to adopt a multi-supplier cooperation model to optimize the procurement plan. Different procurement requirements have different focuses on inventory, logistics, quality, and price. Therefore, in step S3, the conditions for rigid requirements are: inventory ≥ demand, logistics ≤ time limit, quality ≥ standard, and price ≤ budget; the conditions for flexible requirements are: inventory < demand but can be allocated, logistics ≤ time limit, quality ≥ standard, and price ≤ budget; the conditions for combined requirements are: inventory of a single supplier ≥ demand × preset value C, logistics ≤ time limit, quality ≥ standard, and price ≤ budget × preset value D. By changing the procurement requirement parameters and adjusting the insertion position of the dynamic scoring node in each table, suppliers that meet the conditions can be flexibly selected according to specific requirements.

[0063] To provide buyers with a basis for decision-making, the comprehensive score is calculated as follows: Comprehensive Score = Inventory Matching Score × E + Logistics Timeliness Score × F + Material Quality Score × G + Price Score × H, where E + F + G + H = 1. This comprehensive score formula considers four aspects: inventory matching, logistics timeliness, material quality, and price, and uses weighted allocation (E + F + G + H = 1) to reflect the importance of different indicators in procurement decisions. This allows for a comprehensive and objective evaluation of suppliers. Buyers can rank suppliers based on the comprehensive score and select the most suitable one. Furthermore, the comprehensive score results can be fed back to suppliers, prompting them to focus on their inventory management, logistics, material quality, and pricing strategies, continuously improving service quality to achieve higher scores in the competition.

[0064] Please see Figure 2 As shown, a second objective of this invention is to provide an intelligent recommendation system for the material supply chain in the construction industry. This intelligent recommendation system for the material supply chain in the construction industry is used to implement the aforementioned intelligent recommendation method for the material supply chain in the construction industry, including:

[0065] The supplier management platform receives supplier reports and extracts information, generates supplier profile data packages with hash values ​​using business license numbers, and verifies data integrity.

[0066] The procurement cloud platform, connected to the supplier management platform, includes a preprocessing module, a screening mechanism, and a ranking module. The preprocessing module receives supplier profile data packages, generates industry classification tags based on the information within the packages, calculates scores for inventory matching, logistics timeliness, and material quality according to preset rules, and creates inventory matching, logistics timeliness, and material quality tables in descending order of score. Simultaneously, it generates a price table based on selling prices and establishes a classification index system by material type. The screening mechanism matches corresponding industry classification tags to material types and sequentially filters suppliers that meet rigid, flexible, and combined needs based on procurement requirements. The ranking module performs a comprehensive evaluation of suppliers and generates lists of solutions for rigid, flexible, and combined needs in descending order of the comprehensive score.

[0067] The determination module connects to the procurement cloud platform and is used to display a list of solutions for rigid demand, flexible demand, and combined demand to the buyer. The buyer compares and selects the solutions and generates a final order. The procurement cloud platform updates the inventory and logistics status in real time.

[0068] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for intelligent recommendation of material supply chains in the construction industry, characterized in that, Includes the following steps: S1. Suppliers upload reports containing business scope, inventory and logistics information to the supplier management platform. The supplier management platform generates a supplier profile data package with the business license number as the primary key and synchronizes it to the procurement cloud platform. In step S1, a hash value is generated using the business license number as the primary key. The supplier profile data package includes business scope, logistics information, inventory, upload timestamp, and data version number. When inventory changes > preset value B or logistics information is updated, the supplier profile data package is automatically updated. S2. After receiving the supplier profile data package, the procurement cloud platform generates industry tags based on keywords in the business scope, calculates the index scores of inventory matching degree, logistics timeliness and material quality according to preset rules, and forms inventory matching degree table, logistics timeliness table and material quality table respectively in descending order of index scores. At the same time, it forms a price table based on the selling price and establishes a classification index system by material type. S3. After the buyer submits its procurement requirements, the following screening process is performed: the supplier list is split using a hash algorithm and processed in parallel by distributed nodes; The selection rules and weighting coefficients are adjusted in real time based on the demand parameters to select suppliers that meet rigid, flexible, and combined requirements; the suppliers are comprehensively scored, and a list of solutions is formed in descending order of the comprehensive scores. In S3, the filtering mechanism is specifically as follows: Based on the procurement requirements parameters, data screening is carried out on the inventory matching table, logistics timeliness table, material quality table and price table with the help of dynamic scoring nodes. The parts of each table that meet the procurement requirements parameters are extracted to obtain multiple candidate supplier lists and complete the preliminary screening. For each candidate supplier list, the hash value modulo N is taken as the group identifier. The candidate suppliers are assigned to the corresponding group identifier according to the corresponding hash value and stored independently to obtain the group directory of each candidate supplier list. In the modulo N grouping algorithm, when the number of suppliers exceeds the preset threshold, the hierarchical hashing strategy is automatically triggered. In multiple candidate supplier lists, computing nodes with corresponding computing power are allocated according to the number of suppliers in the same group directory. After filtering irrelevant groups, the intersection calculation is performed on the group directory of each candidate supplier list in the same category to obtain suppliers that exist in multiple groups in the same category, which are then used as supplier groups. Extract the supplier's exclusive data from the inventory matching table, logistics timeliness table, material quality table, and price table, and configure the corresponding information nodes for the supplier's exclusive data in each table. The dynamic scoring node comprises a demand receiving end, a parsing and mapping module, and a data docking end. The demand receiving end receives procurement demand parameters. The parsing and mapping module has filtering rule bases corresponding to inventory matching tables, logistics timeliness tables, material quality tables, and price tables, respectively. The parsing and mapping module uses natural language processing technology to parse the procurement demand parameters, transforming them into specific data filtering conditions, and then stores these conditions in the corresponding table's filtering rule base. The data docking end interacts with all information nodes in a single table, filtering the specific data bound to each information node based on the filtering conditions in the corresponding table's filtering rule base within the parsing and mapping module, thus obtaining information nodes that meet the filtering conditions. S4. The buyer evaluates and selects from the list of options, and after confirmation, generates a final purchase order. The procurement cloud platform updates inventory and logistics information simultaneously.

2. The intelligent recommendation method for the material supply chain in the construction industry according to claim 1, characterized in that: In S1, the scope of business includes building material categories and specifications. Inventory is recorded at the SKU level, including current inventory, average monthly shipments, and inventory status. Logistics information includes coverage area, average transportation time, and carrier list. The supplier management platform automatically verifies the data integrity of the reports. When the data missing rate in the reports exceeds the preset value A, a manual review process is triggered.

3. The intelligent recommendation method for the material supply chain in the construction industry according to claim 1, characterized in that: In S2, the industry classification labels are generated based on the building material categories in the business scope. At the same time, the labels are constructed in three layers according to the building material category, building material subcategory, and building material model.

4. The intelligent recommendation method for the material supply chain in the construction industry according to claim 1, characterized in that: In S2, the inventory matching score = current inventory quantity / average daily shipment volume of similar inventory on the procurement cloud platform × 100%; the logistics timeliness score = ∑ (number of cities covered in the region × timeliness coefficient) / total number of cities covered; and the material quality score is based on historical quality inspection data. The classification index system is based on building material models, and separate tables are established for inventory matching, logistics timeliness, material quality, and price.

5. The intelligent recommendation method for the material supply chain in the construction industry according to claim 1, characterized in that: In S3, the conditions for rigid demand are: inventory ≥ demand, logistics ≤ time limit, quality ≥ standard, and price ≤ budget; the conditions for flexible demand are: inventory < demand but can be allocated, logistics ≤ time limit, quality ≥ standard, and price ≤ budget. The conditions for combining requirements are: inventory of a single supplier ≥ demand × preset value C, logistics ≤ time limit, quality ≥ standard, and price ≤ budget × preset value D. By changing the procurement requirement parameters and adjusting the insertion position of the dynamic scoring node in each table, different lists of candidate suppliers can be obtained. The overall score is calculated as follows: Inventory matching score × E + Logistics timeliness score × F + Material quality score × G + Price score × H, where E + F + G + H = 1.

6. A smart recommendation system for the material supply chain in the construction industry, wherein the smart recommendation system for the material supply chain in the construction industry is used to implement the smart recommendation method for the material supply chain in the construction industry according to any one of claims 1-5, characterized in that, include: The supplier management platform receives supplier reports and extracts information, generates supplier profile data packages with hash values ​​using business license numbers, and verifies data integrity. The procurement cloud platform, connected to the supplier management platform, includes a preprocessing module, a screening mechanism, and a sorting module. The preprocessing module receives supplier profile data packages, generates industry classification tags based on the information within the data packages, calculates scores for inventory matching, logistics timeliness, and material quality according to preset rules, and generates inventory matching, logistics timeliness, and material quality tables in descending order of score. Simultaneously, it generates a price table based on selling prices and establishes a classification index system by material type. The screening mechanism matches corresponding industry classification tags to material types and sequentially filters suppliers that meet rigid, flexible, and combined needs based on procurement requirements. The sorting module performs a comprehensive evaluation of suppliers and generates a list of solutions for rigid, flexible, and combined needs in descending order of the comprehensive score. The determination module, connected to the procurement cloud platform, is used to display a list of solutions for rigid demand, flexible demand, and combined demand to the buyer. The buyer compares and selects the solutions and generates a final order. The procurement cloud platform updates the inventory and logistics status simultaneously.

Images