Decoration engineering material supply chain collaboration system and method

Abstract

The invention provides a decoration engineering material supply chain collaboration system and method, and the method comprises the steps: collecting the loss, transportation and use data of materials in a supply chain from a purchasing link to a construction link, and carrying out the grouping of the characteristics of the materials through a clustering algorithm, and obtaining a material value contribution classification; obtaining the predicted value sequence, and performing priority ranking on materials in the predicted value sequence by integrating project demand adjustment data to obtain a resource allocation priority list; inputting market change response data to a construction matching evaluation link through the supplementary purchasing instruction sequence to obtain a matching degree scoring matrix; and obtaining the optimized resource allocation plan, integrating material value tracking data and feeding back the data to the acquisition link to form a closed-loop resource optimization mechanism.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to a collaborative system and method for the supply chain of decorative engineering materials. Background Technology

[0002] Supply chain management for decorative engineering materials, as a crucial area concerning project cost, schedule, and quality, occupies an indispensable position in the construction industry. With the increasing scale and complexity of engineering projects, efficient supply chain collaboration directly impacts a company's competitiveness and project success rate; its criticality is self-evident. However, current management methods often struggle to cope with the volatile market environment and complex project demands, exposing some deep-seated problems.

[0003] Existing methods for managing the entire materials supply chain often lack in-depth analysis of the dynamic relationships between different stages, particularly in material valuation and resource allocation, making precise matching and real-time adjustments difficult. This not only leads to resource waste but can also impact project progress and cost control due to information asymmetry. Especially when multiple material types are involved, management methods often fail to adapt to the differences in material characteristics, resulting in delayed and inefficient decision-making.

[0004] In this field, the core technical challenges lie primarily in achieving end-to-end value tracking of materials from procurement to construction, and in dynamically optimizing resource allocation based on market and project changes. The difficulty in material value tracking stems from the varying losses, transportation, and usage patterns of different materials throughout the supply chain, lacking a unified standard to comprehensively assess their actual contribution to the project. For example, some materials may experience high losses during transportation, while others may suffer increased waste due to improper construction matching. This makes it difficult to accurately assess the true value of each batch of materials based solely on experience. Furthermore, this problem evolves into the challenge of dynamically adjusting resource allocation. Because the fluctuation trend of material value cannot be accurately grasped, it is difficult to optimize inventory and procurement strategies in a timely manner when market prices change or project schedules are adjusted, often leading to material stockpiles or shortages.

[0005] Therefore, how to build a collaborative mechanism in the supply chain of decorative engineering materials that can comprehensively assess the value contribution of different materials and dynamically adjust resource allocation according to market and project needs has become a key issue that urgently needs to be addressed. Summary of the Invention

[0006] This invention provides a collaborative system and method for the supply chain of decorative engineering materials, mainly including: By collecting data on material losses, transportation, and usage from procurement to construction in the supply chain, clustering algorithms are used to group the material characteristics to obtain a classification of material value contribution. Based on the material value contribution classification, the trend of change is extracted from market price fluctuation data and project progress data. If the trend of change exceeds a preset threshold, a machine learning prediction model is applied to the materials in the material value contribution classification to predict the value fluctuation and obtain the predicted value sequence. Obtain the predicted value sequence, and prioritize the materials in the predicted value sequence by integrating project demand adjustment data to obtain a resource allocation priority list; Inventory adjustment parameters are extracted from the resource allocation priority list. The procurement strategy optimization data is combined with the resource allocation priority list using an information fusion method. If the inventory level is lower than the threshold corresponding to the predicted value sequence, a supplementary procurement instruction sequence is generated. By inputting market change response data into the construction matching assessment stage through the supplementary procurement instruction sequence, a matching degree score matrix is ​​obtained. Based on the matching score matrix, a unified measurement index is obtained from the entire supply chain process data, and the matching score matrix is ​​weighted and calculated. If the calculated score is lower than a preset threshold, the resource allocation scheme is adjusted to obtain an optimized resource allocation plan. The optimized resource allocation plan is obtained, and the material value tracking data is integrated and fed back to the collection stage to form a closed-loop resource optimization mechanism.

[0007] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects: This invention discloses a closed-loop optimization method for building material resources based on data driven throughout the entire supply chain. By collecting data on material losses, transportation, and usage from procurement to construction, a clustering algorithm is used to classify the value contribution of material characteristics. Market price fluctuations and project progress data are combined to extract trends. When a trend exceeds a threshold, a machine learning model is applied to predict material value fluctuations, generating a predicted value sequence. Subsequently, project demand adjustments are integrated to prioritize materials, forming a resource allocation priority list. Inventory adjustment parameters are extracted based on this list, and procurement strategies are optimized through information fusion. When inventory falls below the predicted threshold, a supplementary procurement instruction sequence is generated. This instruction sequence is further input into the construction matching evaluation stage, generating a matching score matrix. Weighted calculations are performed based on unified supply chain metrics. If the score is below the threshold, resource allocation is adjusted, ultimately forming an optimized resource allocation plan. Simultaneously, material value tracking data is fed back to the initial data collection stage, realizing a closed-loop resource optimization mechanism throughout the entire process. This invention addresses the problems of inaccurate material resource allocation, delayed inventory response, and low construction matching efficiency under supply chain fluctuations. Through data-driven forecasting, prioritization, dynamic procurement, and closed-loop feedback, it significantly improves resource utilization efficiency, reduces costs, and enhances the project's ability to respond quickly to market changes. Attached Figure Description

[0008] Figure 1 This is a flowchart of a collaborative system and method for the supply chain of decorative engineering materials according to the present invention.

[0009] Figure 2 This is a schematic diagram of a collaborative system and method for the supply chain of decorative engineering materials according to the present invention.

[0010] Figure 3 This is another schematic diagram of a collaborative system and method for the supply chain of decorative engineering materials according to the present invention. Detailed Implementation

[0011] The technical solutions of the present invention will be clearly and thoroughly described below with reference to the accompanying drawings. The described embodiments are merely some embodiments of the present invention.

[0012] like Figures 1-3 This embodiment of a collaborative system and method for the supply chain of decorative engineering materials may specifically include: Step S101: By collecting data on material loss, transportation, and usage from the procurement to the construction stages in the supply chain, a clustering algorithm is used to group the material characteristics to obtain a material value contribution classification.

[0013] By deploying sensors and recording systems at various stages of the supply chain, data on material loss, transportation, and usage during procurement, transportation, and construction are collected to obtain a raw multi-dimensional dataset. Missing values ​​are imputed and outliers are removed from the raw multi-dimensional dataset to obtain a cleaned dataset. Standardization is then applied to unify the dimensions of the loss, transportation, and usage data in the cleaned dataset to obtain a standardized dataset. A clustering algorithm is used to group the material loss characteristics in the standardized dataset, resulting in multiple material characteristic clusters. The material value loss rate within each cluster is calculated based on the average values ​​of the loss, transportation, and usage data, yielding a value loss index for each cluster. If the value loss index of a cluster exceeds a preset threshold, the materials in that cluster are classified as belonging to the high-value-contribution loss category; otherwise, they are classified as belonging to the low-value-contribution loss category, resulting in a material value contribution classification result. Based on the material value contribution classification result, corresponding supply chain stages are identified to determine the procurement, transportation, or construction stages where high-value-contribution loss materials are concentrated, resulting in a list of key loss stages.

[0014] For example, in supply chain management, data collected by sensors and recording systems during the procurement, transportation, and construction stages of materials can provide a comprehensive understanding of material loss, transportation, and usage. Suppose that in a building materials supply chain, sensors record the quantity of steel procured, losses during transportation, and usage efficiency at the construction site, forming a raw dataset containing multiple dimensions such as weight, time, and location.

[0015] In one possible implementation, missing value imputation and outlier removal for the original dataset can be achieved using mean imputation and box plot methods.

[0016] For example, if some weight records are missing from the transportation data of a batch of steel, they can be filled in by averaging the values ​​of other records in the batch. Outliers, such as losses exceeding 100 times the normal range during transportation, are identified and removed using box plots to ensure the accuracy of the cleaned dataset. This approach effectively improves the reliability of subsequent analyses.

[0017] Specifically, the cleaned dataset needs to be standardized to unify the units of measurement.

[0018] For example, converting loss data from tons to percentages and transportation time from hours to standardized fractions ensures data comparability across different dimensions. After standardization, differences between data reflect only actual characteristics rather than unit variations, which helps improve the accuracy of clustering algorithms.

[0019] In one possible implementation, the K-means clustering algorithm is used to group the material loss characteristics in the standardized dataset.

[0020] For example, steel data is divided into three clusters: the first cluster represents high loss and low utilization, the second cluster represents low loss and high utilization, and the third cluster represents medium loss and medium utilization. This grouping clearly reveals the differences in material performance throughout the supply chain, laying the foundation for subsequent analysis.

[0021] For example, when calculating the material value loss rate within a cluster, the value loss can be estimated based on the average loss data, transportation data, and usage data of each cluster, combined with the unit price of the material.

[0022] For example, the first cluster has an average loss rate of 20%, a high transportation delay rate, and a utilization rate of only 60%. Its value loss rate is far higher than the preset threshold of 10%, and it is classified as a high-value contribution loss category. The other two clusters are below the threshold and are classified as low-value contribution loss categories. This helps to accurately identify the categories of problematic materials.

[0023] In one possible approach, by linking the classification results to supply chain stages, it was found that the high-value-contributing loss-making steel in the first cluster is mainly concentrated in the transportation stage. For example, multiple records show that improper packaging during transportation led to increased losses. Therefore, the transportation stage was identified as a critical loss-making stage and added to the list. Such analysis directly addresses the root cause of the problem, providing a clear direction for optimizing the supply chain.

[0024] For example, for key loss-making processes, further improvement measures can be developed, such as strengthening transportation packaging standards or optimizing route planning, thereby reducing loss rates and improving overall supply chain efficiency. This complete process, from data collection to process optimization, not only improves resource utilization but also significantly reduces economic losses, bringing long-term benefits.

[0025] Step S102: Based on the material value contribution classification, extract the trend of change from market price fluctuation data and project progress data. If the trend of change exceeds a preset threshold, apply a machine learning prediction model to the materials in the material value contribution classification to predict the value fluctuation and obtain the predicted value sequence.

[0026] Step 1: Obtain the fluctuation data and progress data related to the material value from the market price database and project progress records. Through data cleaning and time series alignment, obtain standardized fluctuation and progress datasets. Step 2: For the standardized fluctuation and progress datasets, use time series analysis methods to extract the changing trends and generate corresponding trend feature sequences. Step 3: If the changing trend in the trend feature sequence exceeds a preset threshold, anomaly markers are applied to the fluctuation data related to the material value, resulting in an anomaly marker dataset. Step 4: Based on the anomaly marker dataset, apply a pre-established support vector machine model to predict value fluctuations, generating a preliminary predicted value sequence. Step 5: Smooth the preliminary predicted value sequence to eliminate short-term noise interference, obtaining an optimized predicted value sequence. Step 6: For the optimized predicted value sequence, group and map it using contribution classification data to generate value fluctuation prediction results under each category. Step 7: Based on the value fluctuation prediction results under each category, construct a sequence generation report and determine the final predicted value sequence output.

[0027] For example, in the field of building materials supply chain management, obtaining fluctuation data and progress data related to material value from market price databases and project schedule records first requires data cleaning to remove invalid records and time series alignment to ensure that the fluctuation data and progress data correspond on the same time axis, thus obtaining standardized fluctuation datasets and progress datasets. This alignment process helps improve the accuracy of subsequent analysis.

[0028] Specifically, for the standardized fluctuation dataset and progress dataset, time series analysis methods are used to extract the changing trends, and corresponding trend feature sequences can be generated through methods such as moving average or exponential smoothing.

[0029] In one embodiment, it is assumed that the monthly price fluctuation data of a certain steel product is a sequence from January to December. The trend feature sequence shows a pattern of stability in the first half of the year and an increase in the second half, which reflects the impact of changes in market supply and demand.

[0030] For example, if the trend in the trend feature sequence exceeds a preset threshold, such as a monthly increase exceeding 15%, then the fluctuation data related to the material value is marked as an anomaly, resulting in an anomaly-marked dataset. This marking can promptly identify price anomalies caused by sudden events, facilitating accurate subsequent predictions.

[0031] In one possible implementation, a pre-built support vector machine model is applied to predict value fluctuations based on an anomaly-labeled dataset. This model, trained on historical labeled data, can capture non-linear relationships and generate a preliminary predicted value sequence.

[0032] Preferably, for the marked steel data, the model may predict a 10% price increase in the next quarter, providing an initial reference.

[0033] Specifically, by smoothing the initial predicted value sequence, such as by using low-pass filtering to eliminate short-term noise interference, an optimized predicted value sequence is obtained. This step removes the influence of short-term market speculation, making the prediction closer to the true trend, thereby improving the reliability of decision-making.

[0034] For example, for the optimized predicted value sequence, group mapping is performed by combining contribution classification data, that is, the predicted sequence is associated with high-value contribution loss class or low-value contribution loss class materials obtained from historical dialogue, and value fluctuation prediction results under the classification are generated.

[0035] In one embodiment, the forecast for cement materials with high-value contribution loss shows greater volatility, which helps to prioritize their risk.

[0036] In one possible implementation, a sequence generation report is constructed based on the value fluctuation prediction results under each category, determining the final predicted value sequence output. This report integrates trend, anomaly, and category information to form a complete output, supporting supply chain managers in adjusting procurement strategies in advance and reducing the risk of losses due to value fluctuations. This prediction mechanism effectively improves the predictability of material value management, optimizes resource allocation, and reduces unnecessary inventory backlogs or emergency procurement costs.

[0037] Step S103: Obtain the predicted value sequence, and prioritize the materials in the predicted value sequence by integrating project demand adjustment data to obtain a resource allocation priority list.

[0038] By acquiring predicted value sequence data from the system, a batch processing method is used to initially classify each element in the sequence, resulting in a classified value dataset. Based on the classified value dataset, project data matching is performed for each category. If the correlation between the matched project data and the predicted value is higher than a preset threshold, it is marked as high priority, resulting in a marked priority dataset. The marked priority dataset is then used to further evaluate the high-priority data using a logistic regression model, determining the potential value ranking of each material. Based on the value ranking results, the system's built-in ranking logic is used to adjust the material priorities, resulting in an adjusted priority sequence. Based on the adjusted priority sequence and resource planning constraints, if a material's priority sequence position is higher than other materials, resources are allocated to it first, resulting in a preliminary resource allocation plan. Using the preliminary resource allocation plan, data is integrated according to the rules for generating the allocation list to determine the final resource allocation priority list.

[0039] For example, when processing predicted value sequence data, relevant data can be extracted from the system first. Suppose a material's value sequence includes data from the past three months, with values ​​of 100, 120, and 110 units respectively. Through batch processing, this data can be initially classified into high, medium, and low value categories, with 100 and 110 classified as medium value and 120 as high value, forming a categorized value dataset. This classification method facilitates subsequent differentiated processing strategies for different categories.

[0040] For example, when matching project data within a categorized value dataset, it's assumed that data in a high-value category must have a correlation of 0.8 or higher with project schedule data to be marked as high priority. If a material in the high-value category has a correlation of 0.85 with project A, exceeding the threshold, it's marked as high priority, while another material with a correlation of only 0.6 is not marked. This matching process helps identify materials that contribute significantly to the project, ensuring the rationality of resource allocation.

[0041] For example, when using a logistic regression model to assess the value of high-priority data, the potential value ranking can be determined by analyzing multiple attributes of the materials, such as market scarcity and project demand. Assuming material X has high market scarcity and a demand of 80%, while material Y has low scarcity and a demand of 60%, then X's potential value ranking is higher than Y's. This assessment method can more accurately reflect the actual importance of the materials.

[0042] For example, when adjusting priorities based on value ranking results, the system's built-in ranking logic might consider material inventory levels and procurement cycles. Suppose material X has low inventory and a long procurement cycle, its priority is raised to first place, while material Y has sufficient inventory, its priority is lowered. This adjustment logic ensures the rationality and timeliness of resource allocation.

[0043] For example, in the resource allocation phase, considering resource planning constraints, such as a budget limit of 500,000 units, material X, which has the highest priority, is allocated 300,000 units, while material Y, which has the second highest priority, is allocated only 100,000 units. This priority allocation method ensures that critical materials receive sufficient support, improving project efficiency.

[0044] For example, when generating the final resource allocation priority list, the allocation scheme data is integrated to ensure that the list clearly reflects the resource proportion and priority order of each material. Assuming the list shows that material X accounts for 60%, material Y accounts for 20%, and other materials together account for 20%, this integration method allows project managers to quickly understand the resource distribution, thereby optimizing the decision-making process.

[0045] For example, in the logical progression from the core solution to the extended solution, the core solution is based on the classification and prioritization of predicted value sequences, while the extended solution can introduce more dimensions of data, such as supplier stability, as a supplementary basis for resource allocation. This approach ensures the integrity of the solution while improving the rationality of allocation through diversified considerations.

[0046] Step S104: Extract inventory adjustment parameters from the resource allocation priority list, and combine the procurement strategy optimization data with the resource allocation priority list using an information fusion method. If the inventory level is lower than the threshold corresponding to the predicted value sequence, generate a supplementary procurement instruction sequence.

[0047] Step 1: Obtain inventory adjustment parameters from the resource allocation priority list, and use data filtering methods to separate parameter fields directly related to inventory levels, obtaining preliminary inventory adjustment basis. Step 2: Based on the preliminary inventory adjustment basis, integrate the procurement strategy optimization data with the resource allocation priority list using information fusion methods to determine the integrated procurement strategy adjustment plan. Step 3: According to the integrated procurement strategy adjustment plan, obtain the current inventory level data and compare it with the predicted value sequence. If the inventory level is lower than the threshold corresponding to the predicted value sequence, a shortage inventory judgment result is generated. Step 4: Based on the shortage inventory judgment result, use preset supplementary procurement rules to generate a corresponding supplementary procurement instruction sequence and determine the priority order of the procurement instructions. Step 5: Based on the priority order of the supplementary procurement instruction sequence, obtain the available resource data of relevant supply chain nodes. If the resource data of the supply chain nodes meets the requirements of the procurement instruction sequence, a resource allocation plan is generated. Step 6: Through the resource allocation plan, obtain the updated inventory data records and determine whether the inventory level has reached the threshold requirement of the predicted value sequence. If not, a secondary adjustment of supplementary procurement instructions is generated. Step 7: Based on the supplementary procurement instructions from the second adjustment, obtain the final inventory adjustment execution plan, and automatically distribute it to the supply chain nodes through the system to complete the dynamic balancing of inventory levels.

[0048] The core of retrieving inventory adjustment parameters from the resource allocation priority list lies in identifying the key fields that directly affect inventory fluctuations. For example...

[0049] In one possible implementation, the system filters parameters such as safety stock coefficient and demand volatility, which are derived from a priority list previously generated based on the predicted value sequence, thereby ensuring that the adjustment basis is highly correlated with the material value.

[0050] Specifically, after separating parameters through data filtering methods, preliminary basis for inventory adjustment can be obtained.

[0051] In one embodiment, assuming the predicted value sequence of a high-priority material A shows a monthly demand of 500 units, while the current inventory is only 300 units, the separated parameters after filtering include a leadtime of 7 days and a safety stock rate of 20%. These fields are directly related to the inventory level, providing a reliable basis for adjustment. This separation helps to accurately pinpoint the problem and avoid interference from irrelevant parameters. Based on the preliminary evidence, the procurement strategy optimization data and priority list are integrated using information fusion methods to determine the integrated procurement strategy adjustment plan.

[0052] For example, information about material A being ranked first in the priority list can be integrated with bulk discount data in the procurement strategy to create an adjustment plan that favors large-volume purchases of material A. This integration strengthens value-driven procurement decisions and effectively reduces overall holding costs. Based on the integration plan, current inventory level data is compared with the predicted value sequence; if it falls below a threshold, an inventory shortage judgment is generated.

[0053] Specifically, if the predicted value sequence sets the threshold for material A at 400 units, while the actual inventory is 300 units, a shortage determination is triggered. This comparison mechanism promptly captures deviations, ensuring agile supply chain response. For shortage determinations, a sequence of instructions is generated using pre-defined replenishment procurement rules, and a priority ranking is determined.

[0054] For example, rules can specify that high-priority materials should be replenished first, generating instructions such as purchasing 200 units of A first, then 100 units of B in sequence. This prioritization helps maintain a continuous supply of high-value materials and reduces the risk of production disruptions. Based on the priority of the instruction sequence, available resource data for each node in the supply chain is obtained, and if the requirements are met, a resource allocation plan is generated.

[0055] In one possible implementation, a node displays 300 units of A available. If this exceeds the 200-unit demand from the instruction, allocation is directly initiated, forming a plan and executing it. This matching improves resource utilization efficiency and avoids blind procurement. Inventory update data is obtained through the resource allocation plan to determine if the threshold has been reached. If not, a secondary adjustment and replenishment is generated.

[0056] For example, if the inventory rises to 450 units after the initial adjustment, but is still below the target threshold of 500 units, then an additional 50 units will be purchased. This iterative adjustment achieves dynamic balance, significantly improving inventory accuracy. Based on the second adjustment and replenishment, the final inventory adjustment execution plan is obtained and automatically distributed to nodes for processing.

[0057] Specifically, the final solution integrates all instructions and pushes them to supplier nodes for execution through the system, achieving closed-loop management of inventory from insufficient to balanced. This automatic distribution helps accelerate response and optimizes the stability and efficiency of the entire resource allocation chain.

[0058] Step S105: Input market change response data into the construction matching assessment stage through the supplementary procurement instruction sequence to obtain the matching degree score matrix.

[0059] Market change signals are collected by supplementing the procurement instruction sequence to obtain change response data. The market change type is analyzed based on the change response data to determine the response data feature vector. Vector similarity is calculated between the response data feature vector and a pre-stored construction matching template to obtain an initial matching score. If the initial matching score is lower than a preset threshold, the procurement instruction sequence is adjusted to obtain an adjusted instruction sequence. Change response data is re-extracted based on the adjusted instruction sequence to determine the updated response data feature vector. Vector similarity is calculated between the updated response data feature vector and the construction matching template to obtain an optimized matching score matrix. All construction matching options are sorted according to the optimized matching score matrix to determine the final matching score matrix.

[0060] For example, in the business areas of inventory management and procurement optimization, collecting market change signals by supplementing procurement order sequences is a key operation. Suppose a company is responsible for the supply chain management of building materials. Fluctuations in raw material prices or a sudden increase in demand can lead to inventory pressure. The company obtains market change signals through procurement order sequences, such as a 5% increase in steel prices or a 10% increase in cement demand, thus generating change response data. This data reflects the dynamic changes in market supply and demand, providing a basis for subsequent decision-making.

[0061] Specifically, when analyzing market change types, response data can be categorized into price fluctuation type and demand surge type. For example, if steel price increases fall under the price fluctuation type, the system will extract the magnitude and frequency of price changes as part of the feature vector. Conversely, an increase in cement demand would be classified as a demand surge type, and its feature vector might include the demand growth rate and changes in order volume within a given time window. Such feature vectors can quantify the specific impact of market changes, providing a basis for subsequent matching.

[0062] For example, when calculating vector similarity with construction matching templates, the system pre-stores multiple construction scenario templates, such as high-rise building construction templates and road construction templates. Each template includes material demand ratios and priority rules. Suppose the current response data feature vector shows a high proportion of steel demand, resulting in a similarity score of 80 with the high-rise building construction template, but only 50 with the road construction template. If the preset threshold is 75, the initial matching score meets the condition, and there's no need to adjust the procurement instruction sequence; the system directly proceeds to the next matching and sorting step.

[0063] Specifically, if the initial matching score is below a threshold, such as only 60 points, the system will trigger an adjustment to the procurement instruction sequence. Assuming the adjustment increases the proportion of steel procurement, after re-extracting the change response data, the updated feature vector shows that the steel demand ratio is more consistent with high-rise building construction templates, increasing the similarity score to 82 points. This adjustment ensures a higher degree of alignment between the procurement strategy and actual construction needs.

[0064] For example, in optimizing the ranking of the matching score matrix, the system will score multiple construction matching options comprehensively. Suppose the final matrix shows that high-rise building construction templates score 85 points, road construction templates score 55 points, and bridge construction templates score 70 points, then the high-rise building construction template will be prioritized as the final matching result. This ranking method ensures accurate alignment between resource allocation and construction needs.

[0065] Specifically, multiple comparisons and optimizations of the updated response data feature vector with the template help companies respond quickly to changes in a dynamic market environment. For example, by adjusting the procurement order sequence, companies avoid inventory shortages caused by market price fluctuations while ensuring the continuity of material supply for construction projects. This method has significant advantages in ensuring supply chain stability and resource utilization efficiency, especially in the building materials sector, where it effectively addresses the challenges posed by market uncertainty.

[0066] For example, determining the final matching score matrix can also provide a reference for subsequent inventory adjustments. Suppose that high-rise building construction formwork requires a steel inventory ratio of 60%, while the current inventory is only 40%. The system will generate a supplementary procurement plan based on the matching results to ensure that the inventory level is consistent with construction needs. This closed-loop mechanism plays a crucial role in the dynamic balancing of inventory.

[0067] Step S106: Based on the matching score matrix, obtain a unified measurement index from the entire supply chain process data, perform a weighted calculation on the matching score matrix, and if the calculated score is lower than a preset threshold, adjust the resource allocation scheme to obtain an optimized resource allocation plan.

[0068] By leveraging supply chain data and end-to-end information, operational status data for each stage is acquired. Data cleaning and standardization processes are then employed to obtain a structured supply chain dataset. Based on this dataset, a matching score matrix is ​​constructed. For each node's resource allocation, a matching score is calculated to determine the distribution of the score matrix values. A unified metric is introduced as an evaluation benchmark, and a weighted calculation method is used to obtain a comprehensive score, determining whether the score meets the expected standard. If the comprehensive score falls below a threshold, a resource allocation adjustment mechanism is triggered, acquiring deviation data in the current resource allocation to determine the direction of the adjustment strategy. Based on the adjustment strategy, an optimized allocation method is used to reallocate resource nodes, generating a new resource plan and obtaining an optimized configuration. The optimized configuration results are used to update the operational status of the supply chain data, obtaining real-time feedback on end-to-end information to determine whether the resource plan meets business needs. For the real-time feedback on end-to-end information, changes in the matching score are continuously monitored, and a logistic regression model is used to analyze potential deviations and determine the stability of the resource allocation plan.

[0069] For example, in the field of supply chain resource allocation optimization, data on the operational status of each link can be obtained by collecting information throughout the entire process, such as key indicators like supplier delivery cycle, inventory turnover rate, and production node capacity utilization rate.

[0070] Specifically, the raw data is first cleaned to remove outliers and missing items, and then standardized to transform data of different dimensions into a unified scale, forming a structured supply chain dataset.

[0071] In one embodiment, a matching score matrix is ​​constructed based on the dataset. For example, the supply chain is divided into four nodes: procurement, logistics, production, and distribution. The matching degree of resource allocation is evaluated for each node. A matching score is calculated for each node. For example, the current supplier resource allocation score for the procurement node is 85 points, and for the logistics node it is 72 points. The score distribution of all nodes is presented in matrix form, which facilitates intuitive identification of low-scoring links.

[0072] For example, a unified metric can be introduced as an evaluation benchmark, such as setting a standard score of 100 points. This is combined with weighted indicators, such as capacity utilization (0.4 weight), on-time delivery rate (0.3 weight), and cost control (0.3 weight), to obtain a comprehensive score through weighted calculation. If the calculated comprehensive score is 78 points, which is lower than the preset threshold of 80 points, it indicates a deviation in resource allocation, requiring further optimization.

[0073] Specifically, when the overall score is below the threshold, an adjustment mechanism is triggered. The current deviation data is analyzed. If it is found that the low score of a logistics node is due to insufficient transportation resources, the adjustment direction is determined to be to increase the investment of transportation capacity.

[0074] In one embodiment, an optimized allocation method is used to reallocate resources, such as allocating idle spare vehicles to logistics nodes, generating a new resource plan, and improving the logistics node score to 88 points after optimization, resulting in a more balanced overall configuration.

[0075] Understandably, optimizing supply chain operations and obtaining real-time feedback, such as reducing inventory turnover from 15 days to 12 days after implementing a new plan, helps determine whether business needs are met. This real-time feedback helps verify the effectiveness of adjustments and avoids wasting resources.

[0076] For example, by continuously monitoring changes in matching scores based on feedback information, a logistic regression model can be used to analyze potential deviations, such as identifying future risks that may arise from seasonal demand fluctuations, and determining the long-term stability of resource allocation. This monitoring mechanism can promptly identify potential problems, ensure supply chain resilience, improve overall operational efficiency and responsiveness, and reduce delay costs caused by resource mismatches.

[0077] Step S107: Obtain the optimized resource allocation plan, integrate material value tracking data and feed it back to the collection stage to form a closed-loop resource optimization mechanism.

[0078] By extracting optimization scheme data from the resource allocation system and analyzing its application effects at different stages, preliminary allocation and matching results are obtained. Based on these preliminary results and material value tracking data, the value contribution ratio of each stage is calculated, and the value distribution of key stages is determined. If the value distribution of key stages is lower than a preset threshold, a data integration process is triggered, comparing the tracking data with the data collected to determine if any deviation exists. The comparison results reveal the specific range and affected stages of the deviation, and a pre-established correction model is used to obtain adjusted resource optimization parameters. Based on these adjusted parameters, the feedback path in the closed-loop mechanism is updated, and the priority order of value feedback is determined. The priority order of value feedback is then used to adjust the resource allocation execution logic for higher-priority stages, resulting in the final optimized execution scheme. Finally, the data flow path in the mechanism loop is updated using the final optimized execution scheme, and data consistency during the loop is assessed, completing the closed-loop processing of resource optimization.

[0079] For example, in the business area of ​​supply chain resource allocation, the effectiveness of an application can be evaluated by analyzing operational data from different stages of the resource allocation system, specifically regarding the extraction of optimization solutions from the system. Suppose a supply chain network involves three stages: procurement, warehousing, and distribution. Data extracted by the system shows a response time of 3 days for procurement, an inventory turnover of 5 days for warehousing, and a delivery rate of 90% for distribution. By comparing historical data with industry benchmarks, it can be preliminarily determined that the efficiency of the distribution stage is relatively low and requires close attention.

[0080] For example, regarding the topic of calculating the value contribution ratio by combining material value tracking data.

[0081] Understandably, material value tracking typically involves allocating costs across the entire process of raw material procurement to final delivery. Assume the total cost of raw materials is 1 million yuan, with procurement accounting for 40%, warehousing for 30%, and delivery for 30%. Analysis shows that if delivery delays lead to a 10% increase in additional costs, its value contribution decreases, necessitating further optimization of resource allocation.

[0082] For example, to determine if the value distribution of key processes falls below a threshold and trigger the data integration process, a preset threshold, such as a value contribution ratio of no less than 25%, can be used to screen problematic processes. If the proportion of the delivery process drops to 20%, data integration is initiated, comparing the tracking data with the original data from the collection process to determine if there are any recording discrepancies or operational errors, thereby identifying the root cause of the problem.

[0083] For example, in a scenario where the deviation range is determined and a correction model is used to adjust resource optimization parameters, it is assumed that the deviation in the delivery process is mainly due to unreasonable transportation route planning, resulting in a 15% increase in time costs. By using a correction model to replan the route, adjust the number and frequency of delivery vehicles, and optimize the parameters, it is expected that the time cost can be reduced to less than 5%.

[0084] For example, regarding the feedback path and priority order of value feedback in the closed-loop update mechanism, feedback data from the delivery stage can be processed first to ensure that the effects of resource allocation adjustments are reflected in the system in real time. Assuming the priority order is delivery, warehousing, and procurement, the resource allocation logic will prioritize ensuring the allocation of vehicles and personnel for the delivery stage.

[0085] For example, under the theme of adjusting the execution logic of resource allocation and forming the final optimized execution plan, for the high priority of the delivery link, the number of night delivery shifts can be increased from the original 2 shifts to 3 shifts to ensure that the delivery rate is increased to more than 95%, thus forming the final plan.

[0086] For example, regarding the data flow path and data consistency assessment in the update mechanism loop, assuming the adjusted data flow path is from delivery feedback to warehousing and then to procurement, by comparing the data update time and content at each stage, it is ensured that no information is missed, completing the closed-loop process. This approach helps improve the responsiveness and accuracy of resource allocation, ensuring the smooth operation of the entire supply chain.

[0087] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A collaborative system and method for the supply chain of decorative engineering materials, characterized in that, The method includes: By collecting data on material losses, transportation, and usage from procurement to construction in the supply chain, clustering algorithms are used to group the material characteristics to obtain a classification of material value contribution. Based on the material value contribution classification, the trend of change is extracted from market price fluctuation data and project progress data. If the trend of change exceeds a preset threshold, a machine learning prediction model is applied to the materials in the material value contribution classification to predict the value fluctuation and obtain the predicted value sequence. Obtain the predicted value sequence, and prioritize the materials in the predicted value sequence by integrating project demand adjustment data to obtain a resource allocation priority list; Inventory adjustment parameters are extracted from the resource allocation priority list. The procurement strategy optimization data is combined with the resource allocation priority list using an information fusion method. If the inventory level is lower than the threshold corresponding to the predicted value sequence, a supplementary procurement instruction sequence is generated. By inputting market change response data into the construction matching assessment stage through the supplementary procurement instruction sequence, a matching degree score matrix is ​​obtained. Based on the matching score matrix, a unified measurement index is obtained from the entire supply chain process data, and the matching score matrix is ​​weighted and calculated. If the calculated score is lower than a preset threshold, the resource allocation scheme is adjusted to obtain an optimized resource allocation plan. The optimized resource allocation plan is obtained, and the material value tracking data is integrated and fed back to the collection stage to form a closed-loop resource optimization mechanism.

2. The collaborative system and method for the supply chain of decorative engineering materials according to claim 1, characterized in that, The process involves collecting data on material losses, transportation, and usage from procurement to construction within the supply chain. A clustering algorithm is then used to group the material characteristics to obtain a classification of material value contribution, including: By deploying sensors and recording systems at various stages of the supply chain, data on material loss, transportation, and usage during procurement, transportation, and construction are collected to obtain a raw, multi-dimensional dataset. The original multidimensional dataset is imputed for missing values ​​and outliers are removed to obtain a cleaned dataset. Standardization processing was used to unify the dimensions of loss data, transportation data, and usage data in the cleaned dataset, resulting in a standardized dataset. The material loss characteristics in the standardized dataset are grouped using a clustering algorithm to obtain multiple material characteristic clusters. The value loss rate of materials within each cluster is calculated based on the average values ​​of loss data, transportation data, and usage data within each material characteristic cluster, thus obtaining the value loss index for each cluster. If the value loss index of a certain cluster is higher than the preset threshold, the material in that cluster is determined to belong to the high value contribution loss category; otherwise, it is determined to belong to the low value contribution loss category, and the material value contribution classification result is obtained. Based on the classification results of material value contribution, we can identify the corresponding supply chain links, determine the procurement, transportation or construction links where high-value contribution loss materials are concentrated, and obtain a list of key loss links.

3. The collaborative system and method for the supply chain of decorative engineering materials according to claim 1, characterized in that, The process involves classifying materials based on their value contribution, extracting trends from market price fluctuation data and project progress data, and if these trends exceed a preset threshold, applying a machine learning prediction model to predict value fluctuations for the materials in the value contribution classification to obtain a predicted value sequence, including: Step 1: Obtain the fluctuation data and progress data related to the value of the material from the market price database and project progress records. Through data cleaning and time series alignment, obtain the standardized fluctuation dataset and progress dataset. Step 2: For the standardized fluctuation dataset and progress dataset, time series analysis methods are used to extract the changing trends and generate corresponding trend feature sequences; Step 3: If the trend in the trend feature sequence exceeds a preset threshold, then anomaly markers are applied to the fluctuation data related to the material value to obtain an anomaly marker dataset; Step 4: Based on the anomaly-labeled dataset, apply a pre-established support vector machine model to predict value fluctuations and generate a preliminary predicted value sequence; Step 5: By smoothing the preliminary predicted value sequence to eliminate short-term noise interference, the optimized predicted value sequence is obtained; Step Six: For the optimized predicted value sequence, group and map it with contribution classification data to generate value fluctuation prediction results under each category; Step 7: Based on the value fluctuation prediction results under the classification, construct a sequence generation report and determine the final predicted value sequence output.

4. The collaborative system and method for the supply chain of decorative engineering materials according to claim 1, characterized in that, The process of obtaining the predicted value sequence involves prioritizing the materials in the predicted value sequence by integrating project demand adjustment data to obtain a resource allocation priority list, including: By obtaining the predicted value sequence data from the system, each element in the sequence is initially classified using a batch processing method to obtain the classified value dataset; Based on the classified value dataset, the project data is matched for each category. If the correlation between the matched project data and the predicted value is higher than the preset threshold, it is marked as high priority, and the marked priority dataset is obtained. After obtaining the labeled priority dataset, a logistic regression model is used to further evaluate the value of high-priority data and determine the potential value ranking of each material. Based on the value ranking results, the priority of materials is adjusted through the system's built-in ranking logic to obtain an adjusted priority sequence; Based on the adjusted priority sequence and the constraints of resource planning, if a certain material has a higher priority sequence than other materials, resources will be allocated to it first, resulting in a preliminary resource allocation plan. Based on the preliminary resource allocation plan, data is integrated according to the rules for generating the allocation list to determine the final resource allocation priority list.

5. The collaborative system and method for the supply chain of decorative engineering materials according to claim 1, characterized in that, The process involves extracting inventory adjustment parameters from the resource allocation priority list, combining procurement strategy optimization data with the resource allocation priority list using an information fusion method, and generating a supplementary procurement instruction sequence if the inventory level is lower than the threshold corresponding to the predicted value sequence. This sequence includes: Step 1: Obtain inventory adjustment parameters from the resource allocation priority list, and use data filtering methods to separate the parameter fields directly related to the inventory level to obtain the preliminary basis for inventory adjustment; Step 2: Based on the initial inventory adjustment basis, integrate the procurement strategy optimization data with the resource allocation priority list using information fusion methods to determine the integrated procurement strategy adjustment plan; Step 3: Based on the integrated procurement strategy adjustment plan, obtain the current inventory level data and compare it with the predicted value series. If the inventory level is lower than the threshold corresponding to the predicted value series, generate a judgment result of insufficient inventory. Step 4: Based on the determination of insufficient inventory, use the preset replenishment purchase rules to generate a corresponding replenishment purchase instruction sequence and determine the priority order of the purchase instructions; Step 5: Based on the priority of the supplementary procurement order sequence, obtain the available resource data of the relevant supply chain nodes. If the resource data of the supply chain nodes meets the requirements of the procurement order sequence, generate a resource allocation plan. Step Six: Obtain updated inventory data records through the resource allocation plan, determine whether the inventory level has reached the threshold requirement of the predicted value sequence, and if not, generate a supplementary purchase order for secondary adjustment. Step 7: Based on the supplementary procurement instructions from the second adjustment, obtain the final inventory adjustment execution plan, and automatically distribute it to the supply chain nodes through the system to complete the dynamic balancing of inventory levels.

6. The collaborative system and method for the supply chain of decorative engineering materials according to claim 1, characterized in that, The process involves inputting market change response data into the construction matching assessment stage through the supplementary procurement instruction sequence to obtain a matching degree scoring matrix, including: Market change signals are collected by supplementing the procurement order sequence to obtain change response data; Based on the analysis of market change types using change response data, the feature vector of the response data is determined; An initial matching score is obtained by calculating the vector similarity between the response data feature vector and the pre-stored construction matching template. If the initial matching score is lower than the preset threshold, the procurement instruction sequence will be adjusted to obtain the adjusted instruction sequence. Based on the adjusted instruction sequence, the changed response data is re-extracted, and the updated response data feature vector is determined. The optimized matching score matrix is ​​obtained by calculating the vector similarity between the updated response data feature vector and the construction matching template. All construction matching options are sorted according to the optimized matching score matrix to determine the final matching score matrix.

7. The collaborative system and method for the supply chain of decorative engineering materials according to claim 1, characterized in that, The process involves obtaining a unified measurement index from the entire supply chain data based on the matching score matrix, performing a weighted calculation on the matching score matrix, and adjusting the resource allocation scheme if the calculated score is lower than a preset threshold to obtain an optimized resource allocation plan, including: By acquiring operational status data for each stage through supply chain data and end-to-end information, and then using data cleaning and standardization processes, a structured supply chain dataset is obtained. Based on the structured supply chain dataset, a matching score matrix is ​​constructed. For the resource allocation of each node, the matching score is calculated, and the distribution of the score matrix values ​​is determined. For the scoring matrix values, a unified measurement value is introduced as the evaluation benchmark. Combined with the weight ratio of the indicators, a weighted calculation method is used to obtain the comprehensive scoring result and determine whether the score meets the expected standard. If the overall score is lower than the score threshold, the resource allocation adjustment mechanism will be triggered to obtain the deviation data of the current resource allocation and determine the direction of the adjustment plan. Based on the direction of the adjustment plan, the optimization allocation method is adopted to reallocate resource nodes, generate a new resource plan table, and obtain the optimized configuration result. Based on the optimized configuration results, update the operational status of the supply chain data, obtain real-time feedback on the entire process information, and determine whether the resource plan meets business needs. Based on the real-time feedback of the entire process information, we continuously monitor changes in the matching score, use a logistic regression model to analyze potential deviations, and determine the stability of the resource allocation plan.

8. The collaborative system and method for the supply chain of decorative engineering materials according to claim 1, characterized in that, The process of obtaining the optimized resource allocation plan and integrating material value tracking data back to the data collection stage to form a closed-loop resource optimization mechanism includes: By extracting data on optimization schemes from the resource allocation system and analyzing their application effects at different stages, preliminary allocation and matching results are obtained. Based on the preliminary allocation and matching results, combined with material value tracking data, the value contribution ratio of each link is calculated, and the value distribution of key links is determined. If the value distribution of key links is lower than the preset threshold, the data integration process is triggered to compare the tracking data with the data collected in the collection process to determine whether there is a deviation. By comparing the results, the specific range and influencing factors of the deviation are obtained, and the adjusted resource optimization parameters are obtained by using a pre-established correction model. Based on the adjusted resource optimization parameters, update the feedback paths in the closed-loop mechanism and determine the priority order of value feedback; Prioritize the acquisition of value feedback, adjust the execution logic of resource allocation for higher priority stages, and obtain the final optimized execution plan; By optimizing the execution plan, updating the data flow path in the mechanism loop, judging the data consistency in the loop process, and completing the closed-loop processing of resource optimization.

Images