Automatic price comparison supply management system based on supplier price and address

Abstract

The present invention discloses an automatic price inquiry and comparison supply management system based on supplier prices and addresses, and specifically relates to the field of price inquiry and comparison supply management, and is used to solve the problems of lagging geographic information and insufficient matching between suppliers and purchasing units in existing traditional supply chain management; by parsing supplier addresses into standardized geographic information, and clustering and regionalizing suppliers based on geographic distribution, by retrieving the area where the purchasing unit is located, associating its location with that of related suppliers, and defining the purchasing mode, a multi-dimensional logistics network topology map is generated, and the transportation path is optimized and evaluated in real time, and the supplier price trend is dynamically predicted based on historical price data and market fluctuations, and the prediction model is adjusted in combination with real-time data, and the optimal supplier list is generated according to the preset purchasing scenario and multi-dimensional constraints, thereby achieving accurate and efficient supply chain decision-making.

Description

Technical Field

[0001] The present invention relates to the technical field of automatic price inquiry and comparison supply management, and more specifically, to an automatic price inquiry and comparison supply management system based on supplier prices and addresses. Background Art

[0002] In modern supply chain management, supplier selection and price comparison processes increasingly rely on data-driven decision-making systems, especially in environments with wide geographical distribution, frequent market fluctuations, and rapid changes in demand. Existing supplier management systems often focus on static comparisons of prices and supplier information, lacking in-depth exploration and real-time response to geographical factors and market dynamics forecasts.

[0003] Therefore, it is necessary to establish an automatic price inquiry and comparison supply management system to solve the problems of market information lag and insufficient dynamic matching capabilities between suppliers and purchasing units in traditional supply chain management, ensure the price inquiry and comparison flexibility and response speed of the supply chain, and balance the supply and demand stability in the region.

[0004] In order to solve the above problems, a technical solution is now provided. Summary of the invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a supply management system for automatically inquiring and comparing prices based on supplier prices and addresses to solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] The supplier address resolution module builds a hierarchical storage architecture based on the supplier's original address, classifies and stores geographic data according to the set addressing logic level and establishes index relationships;

[0008] The address clustering and regional division module generates spatial groupings of suppliers’ geographical distribution based on clustering algorithms, converts clustering outputs into dynamic partitions of regionalized polygonal boundaries, and establishes an associated architecture for real-time dynamic update of geographic information and real-time adjustment of regional boundaries;

[0009] The location association module retrieves the dynamic partition of the purchasing unit, associates the purchasing unit with the start and end coordinates of all suppliers in the dynamic partition to which it belongs, and defines the purchasing mode of each purchasing unit;

[0010] The transport link selection module generates a multi-dimensional logistics network topology map based on the association between the supplier location and the purchasing unit location, and performs comprehensive quantitative calculations on the recommended scores of logistics paths in real time to form a continuously optimized transport link evaluation framework;

[0011] The supply price forecasting module establishes a dynamic price forecasting model based on the historical price data of alternative suppliers, dynamically introduces the latest market data to adjust the model parameters, and forms a price forecasting feedback mechanism based on rolling updates;

[0012] The price inquiry and comparison decision module integrates the address clustering and regional division module, the transportation link selection module and the supply price prediction module into a framework to establish a comprehensive decision engine for price inquiry and comparison supply management. It defines a multi-dimensional constraint set for pre-purchase scenarios based on the purchasing mode of each purchasing unit and generates a pre-purchase supplier list through the decision engine.

[0013] In a preferred embodiment, a hierarchical storage architecture is constructed based on the supplier's original address, and geographic data is classified and stored according to a set addressing logic level and an index relationship is established, which specifically includes:

[0014] The original address information provided by the supplier is split into standardized address units, including country, region, city and specific location fields, through a hierarchical approach based on logical rules at the geographic administrative level;

[0015] Apply nonlinear matching algorithms, combined with preset dynamic mapping rules, to convert standardized addresses into multi-dimensional geographic coordinate points, and correct coordinate errors based on regional characteristics;

[0016] With the help of multi-model comparison mechanism, duplicate, incomplete or conflicting data in supplier addresses can be automatically marked and corrected;

[0017] Cross-correlate geographic coordinates with regional logistics characteristics, transportation network structure and administrative boundaries to generate geographic information objects with multi-dimensional dynamic attributes, build a hierarchical storage architecture, store the geographic data of geographic information objects in layers according to the set addressing logic level and establish index relationships.

[0018] In a preferred embodiment, generating spatial groupings of supplier geographic distribution based on clustering algorithms, converting clustering outputs into dynamic partitions of regionalized polygonal boundaries, and establishing an associated architecture for real-time dynamic update of geographic information and real-time adjustment of regional boundaries specifically include:

[0019] Based on the calculation results of the supplier's geographical distribution density and quantity characteristics, density clustering and center point clustering are dynamically called to generate spatial grouping;

[0020] The spatial parsing rules are used to convert the clustering output into regionalized polygonal boundaries, and the boundary rules between regions are defined by recursive fitting to establish smooth partition boundaries.

[0021] Based on the logistic cost and supply density variables, the initial partition is iteratively updated using a multi-objective optimization model to form a dynamic regional partition with multi-factor associations, which is marked as dynamic partitioning;

[0022] The clustering areas are associated with the supplier information and stored, and an associated architecture for real-time dynamic update of geographic information and real-time adjustment of regional boundaries is established through a distributed database.

[0023] In a preferred embodiment, the dynamic partition of the purchasing unit is retrieved, the purchasing unit is associated with the start and end points of the location coordinates of all suppliers in the dynamic partition to which it belongs, and the purchasing mode of each purchasing unit is defined, specifically including:

[0024] Dynamically obtain newly registered purchasing units, extract the standardized address units of the purchasing units and convert them into geographic coordinate points;

[0025] Based on the index relationship in the hierarchical storage architecture of the supplier's geographic location, a progressive index search is performed by comparing the distance between coordinate points to select the supplier location with the closest coordinate point;

[0026] Based on the dynamic partition where the supplier closest to the purchasing unit is located, establish the start and end association relationship between the purchasing unit and the location coordinates of all suppliers in the dynamic partition;

[0027] Based on the historical goods purchasing time and goods purchasing price of all purchasing units, time domain analysis is performed on the goods purchasing price data that constitute the time series, and the purchasing pattern of the purchasing units is defined. The purchasing pattern includes the purchasing periodicity, purchasing price range, purchasing price change trend, and purchasing price fluctuation degree of the time domain analysis results.

[0028] In a preferred embodiment, a multi-dimensional logistics network topology diagram is generated based on the association between the supplier location and the purchasing unit location, and a comprehensive quantitative calculation is performed on the recommended score of the logistics path in real time to form a continuously optimized transportation link evaluation framework, which specifically includes:

[0029] Analyze the start-end relationship between the location coordinates of all purchasing units and their associated suppliers, extract the location of logistics sites in the associated space, and generate a multi-dimensional logistics network topology map that includes site distance, transportation mode, and road traffic characteristics;

[0030] According to the traversal time complexity of the topological structure of the logistics network, the path optimization algorithm is called to search for links, and all optional transportation link sets are determined through multiple rounds of iterations;

[0031] According to the distance, mode of transportation and traffic characteristics of the transportation links in the optional transportation link set, the transportation time and transportation cost data of all transportation links are calculated. The comprehensive quantitative value of the transportation link recommendation score of transportation time, transportation cost and transportation reliability is calculated using a multi-objective weighted model. The transportation links in the transportation link set are prioritized and the quality of the links is marked.

[0032] Select alternative transportation plans under different target conditions from the sorted transportation link set, where the target conditions include transportation cost constraints, delivery time requirements, and transportation reliability requirements, and generate a library of transportation link alternative plans based on different target conditions;

[0033] The suppliers corresponding to the transport link alternatives with a higher quality level than the set alternative are marked as alternative suppliers, and the target condition labels to which they belong are marked. The target condition labels can be superimposed;

[0034] By monitoring the real-time logistics status, we can identify situations where transportation links are interrupted or efficiency is reduced, and update the logistics transportation links based on the re-evaluated dynamic optimization results to form a continuously optimized transportation link evaluation framework.

[0035] In a preferred embodiment, a price dynamic prediction model is established based on the historical price data of alternative suppliers, and the latest market data is dynamically introduced to adjust the model parameters to form a price prediction feedback mechanism based on rolling updates, which specifically includes:

[0036] Based on the target condition selection of the transport link, the historical price list of goods published by the candidate suppliers carrying the target condition is automatically collected and generated;

[0037] Obtain dynamic data on the historical prices of goods from each candidate supplier, use multi-level filtering rules to remove outliers, and organize the data into a time series format;

[0038] Based on segmented statistics and frequency analysis, the historical price data is analyzed for periodicity, trend and discrete points, and a multi-dimensional feature set containing fluctuation rules is extracted;

[0039] A price dynamic prediction model with multi-variable input is established based on a multi-dimensional feature set, and a recursive fitting algorithm is used to optimize the model parameters during the training process;

[0040] Collect suppliers' real-time market data and input it into the training model to generate future price trend curves. Calculate the fluctuation range through comprehensive algorithms to form data output for decision-making.

[0041] Dynamically introduce the latest market data to readjust the model parameters, use cross-validation method to compare and correct errors in the prediction results, and form a price prediction feedback mechanism based on rolling updates.

[0042] In a preferred embodiment, the address clustering and regional division module, the transportation link selection module and the supply price prediction module are integrated into a framework to establish a comprehensive decision engine for price comparison and supply management. The multidimensional constraint set of the pre-purchase scenario is defined according to the purchase mode of each purchasing unit. The pre-purchase supplier list generated by the decision engine specifically includes:

[0043] The address clustering and regional division module, transportation link selection module and supply price prediction module are integrated into a framework to establish a comprehensive decision-making engine for price comparison and supply management;

[0044] Based on the purchasing mode of all purchasing units in the location association module, define the multi-dimensional constraint set of the purchasing unit's pre-purchasing scenario, including expected purchasing time, supplier quotation range constraint, and price stability constraint;

[0045] The multidimensional constraint set is organized into multidimensional input parameters, and the input parameter range is adjusted using a normalization algorithm;

[0046] The adjusted input parameters are input into the price inquiry and comparison supply management comprehensive decision engine to generate a recommended list of pre-purchase suppliers based on different transportation target condition labels.

[0047] The technical effects and advantages of the automatic price inquiry and comparison supply management system based on supplier prices and addresses of the present invention are as follows:

[0048] By combining geographic clustering with market fluctuation analysis, the system solves the problems of lagging geographic information, inaccurate market fluctuation prediction, and insufficient matching between suppliers and purchasing units in traditional supply chain management. The system uses dynamic geographic data analysis, real-time market trend prediction, and multi-dimensional constraint decision-making engine to efficiently select suppliers and compare prices in complex environments, ensuring more accurate logistics routes, more reasonable price predictions, and more optimized procurement decisions. It not only optimizes the supplier selection process, but also improves the intelligence of price inquiries, and promotes the transformation of supply chain management from static pricing to dynamic response and optimized decision-making, thereby providing the supply chain with procurement solutions that meet the supply and demand relationship. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 This is a schematic diagram of a supply management system for automatically inquiring and comparing prices based on supplier prices and addresses according to the present invention. DETAILED DESCRIPTION

[0050] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0051] Example 1

[0052] Figure 1 The present invention provides an automatic price inquiry and comparison supply management system based on supplier prices and addresses, which includes the following steps:

[0053] The supplier address resolution module builds a hierarchical storage architecture based on the supplier's original address, classifies and stores geographic data according to the set addressing logic level and establishes index relationships;

[0054] The address clustering and regional division module generates spatial groupings of suppliers’ geographical distribution based on clustering algorithms, converts clustering outputs into dynamic partitions of regionalized polygonal boundaries, and establishes an associated architecture for real-time dynamic update of geographic information and real-time adjustment of regional boundaries;

[0055] The location association module retrieves the dynamic partition of the purchasing unit, associates the purchasing unit with the start and end coordinates of all suppliers in the dynamic partition to which it belongs, and defines the purchasing mode of each purchasing unit;

[0056] The transport link selection module generates a multi-dimensional logistics network topology map based on the association between the supplier location and the purchasing unit location, and performs comprehensive quantitative calculations on the recommended scores of logistics paths in real time to form a continuously optimized transport link evaluation framework;

[0057] The supply price forecasting module establishes a dynamic price forecasting model based on the historical price data of alternative suppliers, dynamically introduces the latest market data to adjust the model parameters, and forms a price forecasting feedback mechanism based on rolling updates;

[0058] The price inquiry and comparison decision module integrates the address clustering and regional division module, the transportation link selection module and the supply price prediction module into a framework to establish a comprehensive decision engine for price inquiry and comparison supply management. It defines a multi-dimensional constraint set for pre-purchase scenarios based on the purchasing mode of each purchasing unit and generates a pre-purchase supplier list through the decision engine.

[0059] A hierarchical storage architecture is built based on the supplier's original address, and geographic data is classified and stored according to the logical rule level and index relationships are established. Through logical rules based on the geographic administrative level, the original address information provided by the supplier (such as "No. 10 Jianguo Road, Chaoyang District, Beijing") is split into standardized address units, including fields such as country, region, city, district, and street. This splitting process uses regular expressions and geographic information libraries (such as Geonames) to automatically parse addresses at each level. After standardization, each field is stored as an independent address unit, such as "Country = China, City = Beijing, District = Chaoyang District, Street = No. 10 Jianguo Road", providing a basis for subsequent geographic information processing and matching.

[0060] Traditional geographic information matching methods often assume that the relationship between addresses and coordinates is linear, that is, one address corresponds to one specific coordinate. However, geographic information in some areas may not be accurately mapped by a simple linear model. For example, in mountainous areas, the gap between addresses and actual geographic coordinates may be large. Nonlinear matching algorithms optimize this matching process by using more complex mathematical models (such as curve fitting, support vector machines, neural networks, etc.) to improve accuracy. These algorithms flexibly adjust the mapping method according to geographical conditions (such as terrain, population density, historical data, etc.) to avoid simple linear derivation errors. Convert standardized address units (such as cities and districts) into multi-dimensional geographic coordinate points (such as latitude and longitude coordinates).

[0061] Combined with the geographic information database, the standard address is matched with its corresponding geographic coordinates. During the matching process, the coordinates are appropriately corrected taking into account regional characteristics (such as terrain, climate, transportation, etc.). For example, the coordinates in remote areas may have errors due to complex road networks or large changes in geographical features. The system can adjust the coordinate results by combining real-time data from multiple sources such as transportation networks, historical address data, GPS data, and meteorological conditions, thereby improving positioning accuracy.

[0062] The corrected geographic coordinates are cross-correlated with regional logistics characteristics (such as road networks, traffic nodes, etc.), traffic network structures and administrative boundaries to generate geographic information objects with multi-dimensional dynamic attributes. These information objects include not only geographic coordinates, but also key data such as logistics routes, traffic conditions, and resource distribution in the region. These data are linked to external APIs (such as Google Maps) through the integrated GIS (Geographic Information System) module to dynamically update the geographic information in the region to ensure that changes in the geographic and logistics networks are reflected in real time.

[0063] All processed geographic information objects are stored in layers according to geographic logic rule levels (such as administrative level, longitude and latitude levels, etc.), and index relationships are established. Through the hierarchical storage architecture, address data, geographic coordinates, regional characteristics, transportation networks and other information are stored in a distributed database in a modular form, which is convenient for query and real-time update. The system establishes an index for each layer of data, making geographic information query efficient and flexible. Specifically, data at the country, region, city, district and other levels will be stored according to the hierarchical relationship to ensure efficient geographic information positioning and retrieval.

[0064] Based on the clustering algorithm, spatial grouping of suppliers' geographical distribution is generated, and the clustering output is converted into dynamic partitioning of regionalized polygonal boundaries, and an associated architecture for real-time dynamic update of geographic information and real-time adjustment of regional boundaries is established.

[0065] By analyzing the supplier's geographic location data, the supplier's geographic distribution density and quantity characteristics are calculated. Density calculation is based on the supplier's distribution pattern in space, that is, the ratio of the number of suppliers in a certain area to the area of ​​the area. Quantity characteristics include the number distribution and clustering of suppliers in each cluster area.

[0066] According to the density and quantity characteristics of supplier distribution, select an appropriate clustering algorithm for spatial grouping. Based on the clustering results, use density clustering (such as DBSCAN) or center point clustering (such as K-Means) to generate geographic spatial grouping. If the suppliers are densely distributed and the clustering areas are irregular, select the DBSCAN algorithm to identify the core points by setting a minimum number of points and distance thresholds, and dynamically form clustering areas. If the supplier distribution is relatively uniform and shows a more obvious concentration trend, use the K-Means algorithm to cluster the suppliers by calculating the Euclidean distance.

[0067] Using spatial parsing rules, the boundaries of each cluster area are converted into polygonal structures. The recursive fitting method is used to smooth the boundaries to avoid irregular jagged shapes on the regional boundaries. This process uses the point set within the region to calculate the regional boundaries through the convex hull algorithm, and recursively adjusts the boundaries to reduce the angle difference of the convex hull to ensure that the final generated boundaries conform to geographical logic.

[0068] Based on the variables of logistics cost and supplier density, a multi-objective optimization model is used to iteratively update the preliminary divided regions to generate dynamic regional divisions. The model comprehensively considers factors such as the distribution density of suppliers in the region, logistics costs, and transportation networks, and dynamically adjusts regional boundaries to optimize overall efficiency.

[0069] The clustering area is associated with the supplier information and stored, and a distributed database is used to establish an associated architecture for real-time dynamic update of geographic information and real-time adjustment of regional boundaries. When the supplier address or transportation network changes, the regional division will be automatically adjusted according to the preset update rules to ensure that the system can provide the latest regional division and supplier matching information at any time.

[0070] Retrieve the dynamic partitions of the purchasing units, associate the purchasing units with all suppliers in the dynamic partitions to which they belong, and define the purchasing mode for each purchasing unit.

[0071] Dynamically obtain the address of the newly registered purchasing unit, decompose the address through natural language processing (NLP) technology or manual annotation, and use the geocoding service to convert the decomposed address field into the corresponding geographic coordinate point.

[0072] Based on the index relationship in the hierarchical storage architecture of the supplier's geographic location, a progressive index search is performed through coordinate point distance comparison. The geographic coordinate points of the purchasing unit are compared with the supplier coordinates in the storage architecture to calculate the geographic distance between the two. Using the progressive indexing method, starting from a larger area, the search range is gradually narrowed through the spatial clustering algorithm until the supplier location closest to the purchasing unit is found. Progressive indexing can be achieved through a preset spatial indexing method (such as R-tree or Quad tree).

[0073] Assume that the coordinates of the purchasing unit are , the coordinates of a supplier are , use the great circle distance formula on the earth's surface to calculate the distance D between the two. The specific expression is:

[0074]

[0075] In the formula, is the radius of the Earth, is the latitude difference between the coordinates of the purchasing unit and the supplier, is the longitude difference between the coordinates of the purchasing unit and the supplier, , are the latitude and longitude of the purchasing unit, , are the latitude and longitude of the supplier respectively.

[0076] Based on the dynamic partition where the closest supplier of the purchasing unit is located, the purchasing unit is associated with all suppliers in the dynamic partition. The purchasing unit's historical purchasing time data and purchasing volume data are collected and stored in the form of time series. The time series data is Fourier transformed to convert it from the time domain to the frequency domain in order to identify periodic characteristics. Based on the results of the frequency domain analysis, the main periodic and volatility characteristics are extracted to define the purchasing mode of the purchasing unit, such as periodic procurement (with high requirements for stable supply and reliable transportation), random or sudden procurement (with high requirements for transportation time). At the same time, it is necessary to define the purchasing price range, purchasing price change trend, and purchasing price fluctuation degree for the subsequent decision engine price comparison input.

[0077] A multi-dimensional logistics network topology map is generated based on the association between the supplier location and the purchasing unit location, and the logistics path is comprehensively and quantitatively calculated in real time to form a continuously optimized transportation link evaluation framework.

[0078] Analyze the location associations between all purchasing units and their associated suppliers, and extract the logistics sites between these locations to form a multi-dimensional logistics network topology map. The system obtains the geographic coordinates of all purchasing units and their associated suppliers and associates them based on geographical location relationships. Then, the main logistics sites, distribution centers, warehousing facilities, etc. in the area are extracted and added to the logistics network. The system forms a topology map based on the geographical distance between each site, traffic routes, transportation methods (such as roads, railways, aviation, etc.) and traffic conditions (such as capacity, congestion). This topology map contains distance information between sites, optional transportation methods (for example: roads, railways) and traffic characteristics (such as road quality, traffic peak hours, etc.) for multi-dimensional analysis.

[0079] The system calculates possible transport links based on the path optimization algorithm and optimizes the link selection process through multiple rounds of iterations. The system selects a suitable path optimization algorithm based on the structural complexity of the logistics network topology, such as the Dijkstra algorithm (when the structural complexity is high) or the A* algorithm (when the structural complexity is low). Then, the system optimizes the path selection by iterative calculation based on the connection between each logistics site, covering all possible transport links. In multiple rounds of iterations, the system continuously adjusts the path selection, taking into account factors such as actual traffic conditions and weather changes, thereby generating a set of optional transport links.

[0080] Comprehensively quantify the available transport links, calculate the weighted comprehensive score based on transport time, cost and reliability, and prioritize the links. The system calculates the transport time (based on traffic conditions, route length, etc.), transport cost (considering transport tools, fuel costs, tolls, etc.) and transport reliability (such as the impact of weather and traffic accidents) of each link based on factors such as site distance, transport mode, and road traffic characteristics of each transport link. The system performs parameter conversion and numerical normalization on the above conditions, and uses a multi-objective weighted model to comprehensively evaluate the parameters and calculate the comprehensive score.

[0081] According to different target conditions (such as maximum transportation cost, shortest delivery time, and optimal transportation reliability), the appropriate transportation plan is selected from the sorted transportation link set to generate an alternative plan library. The system selects from the priority-sorted transportation link set according to different target conditions of procurement needs. According to different target conditions, multiple alternative transportation plan libraries are generated, and each plan library contains the optimal transportation link that meets specific conditions. Each transportation link alternative plan will be associated with the corresponding target condition label (such as "low cost", "high reliability", "short delivery time"), and can be dynamically combined according to actual business needs.

[0082] Real-time monitoring of key indicators in the logistics network, such as transportation timeliness, delays, and cost fluctuations of transportation links. By analyzing real-time data, once the system detects a disruption or decline in efficiency of a transportation link, it will automatically re-evaluate the link based on new traffic information, weather data, etc., adjust the logistics path, and optimize the transportation plan. After dynamic optimization, the system updates the transportation link, and re-evaluates and sorts it, thus forming a continuously optimized transportation link evaluation framework to ensure efficient supply chain management in a changing environment.

[0083] A dynamic price forecasting model is established based on the historical price data of alternative suppliers, and the latest market data is dynamically introduced to adjust the model parameters to form a price forecasting feedback mechanism based on rolling updates.

[0084] Based on the target conditions of the transport link, such as transport cost, delivery time, reliability, etc., the system automatically screens out alternative suppliers that meet the target conditions. The system automatically collects and summarizes the historical price data of these suppliers from the supplier database to form a historical price list of goods. The historical price list records the price information of goods released by each supplier at different time points and under different conditions.

[0085] The historical price data of each candidate supplier is captured in real time and processed initially. Using multi-level filtering rules, the system analyzes the captured data and removes abnormal values ​​(such as sudden price increases, missing data, input errors, etc.). Specifically, the filtering rules include but are not limited to: data-based outlier detection, abnormal fluctuation identification, continuity analysis between data points, etc. The filtered data is converted into a time series format to ensure that it can be analyzed and predicted in chronological order.

[0086] Through segmented statistical methods, historical data is divided into time periods (such as daily, weekly, and monthly), and frequency analysis methods are used to identify periodic fluctuations and trend changes. The system also analyzes discrete points in the data to identify abnormal patterns of price fluctuations. Based on these analysis results, feature sets of multiple dimensions are extracted, including periodic features, trend features, fluctuation amplitude, etc. These feature sets will provide input basis for subsequent price dynamics predictions.

[0087] The model input includes features of various dimensions (such as periodicity, trend, discrete fluctuations, historical prices, etc.), and is trained through machine learning algorithms. During the training process, the system applies recursive fitting algorithms (such as recursive least squares or recursive neural networks) to continuously optimize model parameters to improve the accuracy of predictions. The recursive fitting algorithm dynamically adjusts the model weights so that the model can be continuously improved and accurately capture price trends.

[0088] Collect supplier market data (such as raw material prices, supply and demand, transportation costs, etc.) in real time. These real-time data are input into the trained model to generate a trend curve of future price changes. Through comprehensive analysis, the system not only calculates future price trends, but also calculates the price fluctuation range and related risk factors (optional other decision parameters, such as supplier delivery delays, market demand fluctuations, etc.). These data will be converted into the output required for decision-making, helping decision makers formulate procurement strategies, assess risks and optimize supplier selection.

[0089] Dynamically introduce the latest market data to readjust model parameters, and use cross-validation methods to compare and correct forecast results. The cross-validation method includes dividing historical data into multiple subsets, training and validating in turn, so as to improve the generalization ability of the model. Through this rolling update method, the system can continuously optimize the forecast model, maintain sensitivity to future market price changes, and form a price forecasting mechanism with continuous feedback and optimization.

[0090] Establish a comprehensive decision-making engine for price inquiry and comparison supply management, combine it with a multi-dimensional constraint set that defines the pre-purchase scenario, and generate a pre-purchase supplier list through the decision engine.

[0091] The framework integrates the association architecture of real-time dynamic update of geographic information and real-time adjustment of regional boundaries, the continuously optimized transportation link evaluation framework, and the price forecast feedback mechanism based on rolling updates. Through the integration of these three modules, an efficient and dynamically responsive price comparison and supply management comprehensive decision-making engine is built to provide real-time and comprehensive data support for subsequent procurement decisions.

[0092] Based on the purchasing unit information obtained by the location association module, the purchasing mode of each purchasing unit is analyzed, and the purchasing history, purchasing frequency, demand and other data are extracted to define its pre-purchasing scenario. The pre-purchasing scenario includes multiple key constraints: first, the expected purchasing time of the purchasing unit, which determines the time frame of the purchasing activity; second, the supplier quotation range constraint, which is used to limit the range of the quotation provided by the supplier; then, the price stability constraint, which ensures that the fluctuation range of the supplier's quotation within a certain period of time does not exceed the preset stability range; in addition, link transportation time constraints and logistics cost constraints can be freely added according to demand to evaluate the acceptable range of time and cost of the transportation path, and transportation reliability constraints, which are used to ensure that the transportation reliability of the logistics supplier meets the needs of the purchasing unit. All these constraints are constructed into a multidimensional constraint set through logical association, providing clear parameter boundaries for subsequent price comparison decisions.

[0093] After arranging and defining the multidimensional constraint set, the system converts each constraint into specific input parameters. These input parameters include data such as time, price, cost, and transportation reliability, which need to be normalized. The normalization algorithm is used to convert the value range of each constraint parameter to a unified scale to make it comparable. Through normalization, data of different units and dimensions will be converted into standardized values ​​to avoid the impact caused by different data dimensions and ensure the stability and accuracy of subsequent calculations and decision-making processes.

[0094] The system integrates the aforementioned geographic information, transportation link evaluation and price prediction mechanism, comprehensively considers all constraints, and makes dynamic optimization decisions. Through the optimization algorithm, the system will screen out suppliers that meet the needs of the purchasing unit under multi-dimensional constraints and generate a pre-purchase supplier list. This list not only meets the purchasing unit's requirements for time, cost, reliability, etc., but can also be dynamically adjusted according to real-time market fluctuations to ensure the rationality and flexibility of procurement decisions.

[0095] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters and thresholds in the formula are set by technicians in this field according to actual conditions.

[0096] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium, or may be transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from one website, computer, server or data center to another website, computer, server or data center by wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium may be a solid-state hard disk.

[0097] Those of ordinary skill in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0098] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and modules described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0099] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the modules is only a logical function division. There may be other division methods in actual implementation, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or modules, which can be electrical, mechanical or other forms.

[0100] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, and may be located in one place or distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0101] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0102] If the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage media include: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, and other media that can store program codes.

[0103] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0104] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. Automatic price inquiry and comparison supply management system based on supplier price and address, characterized by: include: The supplier address resolution module builds a hierarchical storage architecture based on the supplier's original address, classifies and stores geographic data according to the set addressing logic level and establishes index relationships; The address clustering and regional division module generates spatial groupings of suppliers’ geographical distribution based on clustering algorithms, converts clustering outputs into dynamic partitions of regionalized polygonal boundaries, and establishes an associated architecture for real-time dynamic update of geographic information and real-time adjustment of regional boundaries; The location association module retrieves the dynamic partition of the purchasing unit, associates the purchasing unit with the start and end coordinates of all suppliers in the dynamic partition to which it belongs, and defines the purchasing mode of each purchasing unit; The transport link selection module generates a multi-dimensional logistics network topology map based on the association between the supplier location and the purchasing unit location, and performs comprehensive quantitative calculations on the recommended scores of the transport links in real time to form a continuously optimized transport link evaluation framework; The supply price forecasting module establishes a dynamic price forecasting model based on the historical price data of alternative suppliers, dynamically introduces the latest market data to adjust the model parameters, and forms a price forecasting feedback mechanism based on rolling updates; The price inquiry and comparison decision module integrates the address clustering and regional division module, the transportation link selection module and the supply price prediction module into a framework to establish a comprehensive decision engine for price inquiry and comparison supply management. It defines a multi-dimensional constraint set for pre-purchase scenarios based on the purchase mode of each purchasing unit and generates a pre-purchase supplier list through the decision engine. Retrieve the dynamic partition of the purchasing unit, associate the purchasing unit with the start and end coordinates of all suppliers in the dynamic partition to which it belongs, and define the purchasing mode of each purchasing unit, including: Dynamically obtain newly registered purchasing units, extract the purchasing units' standardized address units and convert them into geographic coordinate points; Based on the index relationship in the hierarchical storage architecture of the supplier's geographic location, a progressive index search is performed by comparing the distance between coordinate points to select the supplier location with the closest coordinate point; Based on the dynamic partition where the supplier closest to the purchasing unit is located, establish the start and end association relationship between the purchasing unit and the location coordinates of all suppliers in the dynamic partition; Based on the historical goods purchasing time and goods purchasing price of all purchasing units, the time domain analysis is performed on the goods purchasing price data that constitute the time series, and the purchasing mode of the purchasing unit is defined. The purchasing mode includes the purchasing periodicity, purchasing price range, purchasing price change trend, and purchasing price fluctuation degree of the time domain analysis results; A multi-dimensional logistics network topology map is generated based on the association between supplier locations and purchasing unit locations, and the recommended scores of transportation links are comprehensively and quantitatively calculated in real time to form a continuously optimized transportation link evaluation framework. Specifically, it includes: Analyze the start-end relationship between the location coordinates of all purchasing units and their associated suppliers, extract the location of logistics sites in the associated space, and generate a multi-dimensional logistics network topology map that includes logistics site distances, transportation methods, and road traffic characteristics; According to the traversal time complexity of the topological structure of the logistics network, the path optimization algorithm is called to search for links, and all optional transportation link sets are determined through multiple rounds of iterations; According to the distance, mode of transportation and road traffic characteristics of the transportation links in the optional transportation link set, the transportation time and transportation cost data of all transportation links are calculated. The comprehensive quantitative value of the transportation link recommendation score of transportation time, transportation cost and transportation reliability is calculated using a multi-objective weighted model. The transportation links in the transportation link set are prioritized and the quality of the links is marked. Selecting transport link alternatives under different target conditions from the sorted transport link set, the target conditions include transport cost constraints, delivery time requirements and transport reliability requirements, and generating a transport link alternative library based on different target conditions; The suppliers corresponding to the transport link alternatives with a higher quality level than the set alternative are marked as alternative suppliers, and the target condition labels to which they belong are marked. The target condition labels can be superimposed; By monitoring the real-time logistics status, we can identify situations where transportation links are interrupted or efficiency is reduced, and update the logistics transportation links based on the re-evaluated dynamic optimization results to form a continuously optimized transportation link evaluation framework.

2. The automatic price inquiry and comparison supply management system based on supplier price and address according to claim 1 is characterized in that: A hierarchical storage architecture is built based on the supplier's original address, and geographic data is classified and stored according to the set addressing logic level and index relationships are established, including: The original address information provided by the supplier is split into standardized address units, including country, region, city and specific location fields, through a hierarchical approach based on logical rules at the geographic administrative level; Apply nonlinear matching algorithms, combined with preset dynamic mapping rules, to convert standardized addresses into multi-dimensional geographic coordinate points, and correct coordinate errors based on regional characteristics; With the help of multi-model comparison mechanism, duplicate, incomplete or conflicting data in supplier addresses can be automatically marked and corrected; The corrected multi-dimensional geographic coordinate points are cross-correlated with regional logistics characteristics, transportation network structure and administrative boundaries to generate geographic information objects with multi-dimensional dynamic attributes, build a hierarchical storage architecture, store the geographic data of the geographic information objects in layers according to the set addressing logic level and establish an index relationship.

3. The automatic price inquiry and comparison supply management system based on supplier price and address according to claim 2 is characterized in that: Generate spatial grouping of supplier geographic distribution based on clustering algorithm, convert clustering output into dynamic partition of regional polygon boundary, and establish the associated architecture of real-time dynamic update of geographic information and real-time adjustment of regional boundaries. Specifically include: Based on the calculation results of the supplier's geographical distribution density and quantity characteristics, density clustering and center point clustering are dynamically called to generate spatial grouping; The spatial parsing rules are used to convert the clustering output into regionalized polygonal boundaries, and the boundary rules between regions are defined by recursive fitting to establish smooth partition boundaries. Based on the logistic cost and supply density variables, the initial partition is iteratively updated using a multi-objective optimization model to form a dynamic regional partition with multi-factor associations, which is marked as dynamic partitioning; The clustering areas are associated with the supplier information and stored, and an associated architecture for real-time dynamic update of geographic information and real-time adjustment of regional boundaries is established through a distributed database.

4. The automatic price inquiry and comparison supply management system based on supplier prices and addresses according to claim 3 is characterized in that: A dynamic price forecasting model is established based on the historical price data of alternative suppliers, and the latest market data is dynamically introduced to adjust the model parameters to form a price forecasting feedback mechanism based on rolling updates. Specifically, it includes: Based on the target condition selection of the transport link, the historical price list of goods published by the candidate suppliers carrying the target condition tag is automatically collected and generated; Obtain dynamic data on the historical prices of goods from each candidate supplier, use multi-level filtering rules to remove outliers, and organize the data into a time series format; Based on segmented statistics and frequency analysis, the historical price data is analyzed for periodicity, trend and discrete points, and a multi-dimensional feature set containing fluctuation rules is extracted; A price dynamic prediction model with multi-variable input is established based on a multi-dimensional feature set, and a recursive fitting algorithm is used to optimize the model parameters during the training process; Collect suppliers' real-time market data and input it into the training model to generate future price trend curves. Calculate the fluctuation range through comprehensive algorithms to form data output for decision-making. Dynamically introduce the latest market data to readjust the model parameters, use cross-validation method to compare and correct errors in the prediction results, and form a price prediction feedback mechanism based on rolling updates.

5. The automatic price inquiry and comparison supply management system based on supplier prices and addresses according to claim 4 is characterized in that: The address clustering and regional division module, transportation link selection module and supply price prediction module are integrated into a framework to establish a comprehensive decision engine for price comparison supply management. The multi-dimensional constraint set of the pre-purchase scenario is defined according to the purchasing mode of each purchasing unit. The pre-purchase supplier list generated by the decision engine includes: The address clustering and regional division module, transportation link selection module and supply price prediction module are integrated into a framework to establish a comprehensive decision-making engine for price comparison and supply management; Based on the purchasing mode of all purchasing units in the location association module, define the multi-dimensional constraint set of the purchasing unit's pre-purchasing scenario, including expected purchasing time, supplier quotation range constraint, and price stability constraint; The multidimensional constraint set is organized into multidimensional input parameters, and the input parameter range is adjusted using a normalization algorithm; The adjusted input parameters are input into the price inquiry and comparison supply management comprehensive decision engine to generate a recommended list of pre-purchase suppliers based on different target condition tags.

Images