A delivery period optimization method and system for a material procurement plan

Abstract

The application belongs to the field of supply chain and logistics management, and relates to a delivery period optimization method and system for material procurement plan. The method selects relevant variables which have influence on the delivery period of the material plan from the total historical material procurement data, and constructs a multi-dimensional relevant variable data set. The multi-dimensional relevant variable data set is subjected to data processing operation to obtain a standardized variable data set which can be used for modeling analysis. A delivery period planning model is established based on the standardized variable data set, and a parameter solving algorithm is used to solve the delivery period planning model to generate the planned delivery date of various materials. According to the generated planned delivery date, the optimized material procurement plan delivery period arrangement is output in combination with the engineering milestone plan and the material supply guide period. The application can effectively improve the rationality, stability and consistency of the material procurement plan delivery period arrangement.

Description

Technical Field

[0001] This invention relates to the field of supply chain and logistics management, specifically to a method and system for optimizing delivery times in material procurement planning. Background Technology

[0002] In the process of preparing the power grid material procurement plan, the determination of the delivery date needs to take into account multiple factors such as bidding batches, material supply lead time, construction progress milestones, and the matching relationship between material categories. This directly affects the connection between project progress and supply chain response efficiency.

[0003] In existing technologies, delivery dates for materials are typically calculated manually based on bidding schedules at various levels, project milestone plans, and material supply guidelines. This can roughly define delivery nodes. Alternatively, the supply chain can be divided into multiple stages, and the average actual operating time for each stage can be calculated based on the factors most relevant to the guidelines and compared with the standard operating time. A difference threshold is used to determine whether guidelines need to be optimized. This solves the problem of controlling the operating time of each stage of the supply chain and dynamically adjusting guidelines. However, the following problems exist: delivery dates in the material procurement plan are delayed due to factors such as poor coordination of delivery sequences between multiple categories of materials, differences in the actual performance capabilities of suppliers, and the urgency of different project types. In view of this, it is necessary to provide a method and system for optimizing delivery dates in material procurement plans to solve the above-mentioned defects in existing technologies. Summary of the Invention

[0004] The purpose of this invention is to solve the technical problems in the prior art where the delivery time of material procurement is delayed due to factors such as poor coordination of delivery time among multiple types of materials, differences in the actual performance capabilities of suppliers, and the urgency of different project types. The invention provides a method and system for optimizing the delivery time of material procurement plans to solve the technical problems existing in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: Firstly, this application provides a method for optimizing the delivery time of a material procurement plan, comprising the following steps: Step S1: Select the associated variables that affect the planned delivery time of materials from the full historical material procurement data, and construct a multi-dimensional associated variable dataset; Step S2: Perform data processing operations on the multi-dimensional correlation variable dataset to obtain a standardized variable dataset that can be used for modeling and analysis; Step S3: Based on the standardized variable dataset, establish a delivery schedule planning model, use a parameter solving algorithm to solve the delivery schedule planning model, and generate the planned delivery dates for various materials; Step S4: Based on the generated planned delivery date, and in conjunction with the engineering milestone plan and material supply schedule, output the optimized material procurement plan delivery schedule.

[0006] Furthermore, the present invention also provides a delivery time optimization system for material procurement planning, comprising: The module for selecting related variables contains: From the full historical material procurement data, we screened out the related variables that affect the planned delivery time of materials and constructed a multi-dimensional related variable dataset. The data processing module contains: Data processing operations are performed on the multidimensional correlation variable dataset to obtain a standardized variable dataset that can be used for modeling and analysis; The delivery schedule model planning module contains: Based on a standardized variable dataset, a delivery schedule planning model is established, and a parameter solving algorithm is used to solve the delivery schedule planning model to calculate the material plan delivery date. The planned output module contains: Based on the generated planned delivery date, the constraints are checked in conjunction with the engineering milestone plan and material supply lead time, and the optimized material procurement plan delivery schedule is output.

[0007] The beneficial effects of this invention are as follows: This invention constructs a multi-dimensional dataset of related variables, enabling a unified expression of key factors affecting the delivery date of material planning. This improves the data integrity and variable coverage of material planning delivery date analysis, effectively addressing the problem in existing technologies where material planning delivery dates do not fully consider factors such as poor delivery sequence among multiple categories of materials, differences in suppliers' actual performance capabilities, and the urgency of various project types, resulting in delivery delays.

[0008] This invention obtains a standardized variable dataset through data processing operations, which is consistently expressed in a unified numerical space, improving data comparability and model adaptability, and reducing the impact of abnormal data on material planning and delivery date analysis.

[0009] This invention establishes a delivery schedule planning model and uses a parametric solution algorithm to solve the model to generate planned delivery dates for various materials. This allows the planned delivery dates to be calculated by multi-dimensional related variables, thus improving the objectivity and consistency of the material delivery schedule results.

[0010] This invention outputs an optimized material procurement plan delivery schedule based on the generated planned delivery date, combined with the engineering milestone plan and material supply lead time. This allows the planned material delivery date to be verified and corrected under the constraints of the engineering milestone plan and material supply lead time, thereby improving the matching degree between the material procurement plan delivery schedule and the engineering implementation process.

[0011] This invention proposes a method for optimizing the delivery time of a material procurement plan, which effectively improves the rationality, stability, and consistency of the delivery time arrangement of the material procurement plan, and enhances the matching degree between the material plan delivery time and the actual execution process of the project.

[0012] Therefore, it is evident that the present invention has outstanding substantive features and significant progress compared with the prior art, and the beneficial effects of its implementation are also obvious. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0014] Figure 1 This is a flowchart of a method for optimizing delivery times in a materials procurement plan; Figure 2 This is a block diagram illustrating the principle of a delivery time optimization system for material procurement planning.

[0015] The module consists of 1-related variable selection module, 2-data processing module, 3-delivery date model planning module, and 4-plan output module. Detailed Implementation

[0016] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. The following embodiments are explanations of the present invention, but the present invention is not limited to the following implementation methods.

[0017] Example 1: like Figure 1 As shown, this embodiment provides a method for optimizing the delivery time of a material procurement plan, which can realize the collection and processing of multi-dimensional data information throughout the entire material procurement process. The data information includes the bid award date, contract signing time, supplier production cycle, transportation cycle, actual delivery dates of various materials, and construction milestones at each stage of the project. The method for optimizing the delivery time of a material procurement plan includes the following steps: Step S1: Select the associated variables that affect the planned delivery time of materials from the full historical material procurement data, and construct a multi-dimensional associated variable dataset; Step S1 specifically includes: S11. Based on the business rules for the procurement of electrical equipment and materials for power grids, extract all historical material procurement data from the historical material procurement database and construct a full dataset. The business rules for the procurement of electrical equipment and materials for power grids mentioned in step S11 are that during the construction of power grid projects, the procurement, production, delivery and use of various electrical equipment and materials for power grids shall follow established business processes, time sequences and constraints. The comprehensive historical material procurement data includes material procurement plan data, bidding batch schedule data, contract signing data, supply lead time data, actual delivery date data, and project milestone plan data. The material procurement plan data includes material categories, specifications, and required quantities. The bidding batch schedule data includes the time of bidding announcement, bid opening time, and bid award date. The supply lead time data includes production cycle, transportation cycle, and delivery cycle. The project milestone plan data includes commencement time, key milestone time, and completion time.

[0018] By constructing a full dataset based on the business rules for the procurement of electrical equipment and materials for the power grid, the data collection scope covers the entire process of procurement, bidding, supply and project implementation, improving the integrity of data sources and business consistency, and providing a reliable data foundation for subsequent operations.

[0019] S12. Filter the associated variables that affect the delivery time of the material plan from the full dataset and construct a set of associated variables; Step S12 specifically includes: S121. Extract the actual delivery date from the full dataset as the target variable, and extract the award date, supply lead time, engineering milestone plan, material category, delivery date of supporting equipment and delivery date of supporting materials from the bidding batch schedule as candidate variables. S122. Calculate the correlation between each candidate variable and the actual delivery date to obtain the correlation coefficient for each candidate variable. The correlation coefficient is calculated using the Pearson correlation coefficient, and the calculation formula is as follows:

[0020] in, This represents the candidate variable value for the i-th sample; This represents the actual delivery date value for the i-th sample; This represents the mean of the candidate variables; This represents the mean of the actual delivery dates; n represents the sample size.

[0021] S123. Based on the correlation coefficient, the candidate variables are sorted to characterize the influence of each candidate variable on the material plan delivery date; according to the time transmission relationship of bidding batch schedule - supply lead time - project milestone plan in the power grid electrical equipment and material procurement process, the candidate variables are matched with business links to identify the action path of each candidate variable in the formation of the material plan delivery date and determine whether it is in a direct or indirect action link; the candidate variables are checked for completeness and candidate variables with missing proportions exceeding the preset threshold or data anomalies are removed to ensure the data availability of the variables participating in the screening; S124. Select candidate variables that meet the requirements of correlation coefficient higher than the preset threshold, are in the business time transmission link and meet the data integrity requirements as related variables, and construct a set of related variables; The associated variables mentioned in step S12 include the award date, supply lead time, project milestone plan, material category, delivery date of supporting equipment, and delivery date of supporting materials in the bidding batch schedule.

[0022] By combining correlation calculation, business link matching, and data integrity screening, variables strongly correlated with the material plan delivery date are extracted from the full dataset, improving the accuracy and effectiveness of correlation variable screening and reducing the interference of irrelevant variables.

[0023] S13. Based on the degree of influence of the associated variables on the delivery time of the material plan, the associated variables are divided into primary variables and secondary variables, and weights are assigned to each associated variable to form a multi-dimensional associated variable dataset; Step S13 specifically includes: S131. Based on the strength of the influence of the related variables on the actual delivery date, their position in the time chain, and the controllability of the variables, the related variables are classified into primary variables and secondary variables. By calculating the correlation coefficients between each associated variable and the actual delivery date, sorting them according to the absolute value of the correlation coefficients, and combining them with the time transmission relationship, we can identify the position of each associated variable in the process of forming the material plan delivery date and determine whether it is in the direct or indirect link of the time transmission link. The correlation variables that rank first by absolute value of correlation coefficient and are directly involved in the time transmission link are classified as main variables. The main variables include the award date, supply lead time and project milestone plan in the bidding batch schedule. Related variables that constrain the main variables through matching relationships or affect the material delivery time through indirect paths are classified as secondary variables. The secondary variables include the delivery time of supporting equipment and the delivery time of supporting materials. The primary variable mentioned in step S131 is used to determine the time interval of the planned delivery period of materials, and the secondary variable is used to provide auxiliary constraints and corrections to the planned delivery period of materials.

[0024] By combining correlation ranking with time transmission link analysis, the related variables are classified, which provides a statistical basis and business logic support for the division of primary and secondary variables, thereby improving the rationality and interpretability of variable classification.

[0025] S132. Perform weight allocation on the partitioned primary and secondary variables to generate weight values ​​for each related variable: Step S132 specifically includes: S1321. Calculate the initial weights based on the correlation coefficients between each associated variable and the actual delivery date, and normalize the correlation coefficients to obtain the initial weights:

[0026] in, Let m represent the correlation coefficient between the i-th associated variable and the actual delivery date, and m represent the total number of associated variables. S1322. Based on historical delivery deviation data, the initial weights of each related variable are corrected, the error contribution rate of each related variable is calculated, and the initial weights are adjusted according to the error contribution rate. The corrected weights are:

[0027] in, This represents the historical error contribution of the i-th associated variable. Indicates the adjustment coefficient; S1323. Different weight constraints are set for primary and secondary variables, so that the overall weight of primary variables is higher than that of secondary variables, and secondary variables with multi-level matching relationships are subject to progressive weight decay according to the matching level; wherein, the multi-level matching weight decay formula is as follows:

[0028] Where k represents the depth of the associated hierarchy. This represents the attenuation coefficient, and 0 < <1; S1324. Normalize the weights of all associated variables to obtain the final weight distribution, and combine it with the corresponding associated variables to generate a multi-dimensional associated variable dataset.

[0029] Step S132 uses a weight allocation method that combines correlation normalization, historical error correction, and corresponding hierarchical decay to make the weight distribution of each associated variable more consistent with the actual delivery date impact mechanism, thereby improving the accuracy and stability of the weight allocation.

[0030] Step S13 quantifies the impact of different variables on the delivery date of the material plan by classifying the associated variables into hierarchical categories and introducing a weight allocation mechanism, thereby improving the structured nature of variable expression and the adaptability of modeling.

[0031] It should be noted that step S1 improves the completeness and temporal consistency of data by constructing a multi-source historical material procurement data system covering the entire procurement process, which includes information on key aspects such as bidding, supply, delivery, and project progress. By screening related variables and constructing a multi-dimensional variable system, the key factors affecting the planned delivery date of materials are expressed in a structured manner, providing a high-quality input data foundation for the subsequent delivery date optimization model and improving the accuracy of material delivery date prediction and planning.

[0032] Step S2: Perform data processing operations on the multi-dimensional correlation variable dataset to obtain a standardized variable dataset that can be used for modeling and analysis; S21. Perform cluster analysis and outlier removal on the multi-dimensional correlation variable dataset to construct a cleaned correlation variable dataset; Step S21 specifically includes: S211. Extract features from each associated variable in the multi-dimensional associated variable dataset. Take the award date, supply lead time, engineering milestone plan, material category, delivery time of supporting equipment and delivery time of supporting materials in the bidding batch schedule as the analysis objects and construct variable feature vectors. S212. Calculate the similarity between samples based on variable feature vectors, and use Euclidean distance or time feature distance to calculate the distance between samples to obtain the distance matrix between samples; S213. Based on the distance matrix between samples, use K-means clustering or hierarchical clustering to cluster the samples, so that samples with similar time characteristics or delivery characteristics belong to the same cluster. S214. Calculate the distance between the sample and the cluster center within each cluster, and identify samples whose distance exceeds the average distance range within the same cluster as abnormal samples; S215. Remove abnormal samples, retain normal sample data, and generate a cleaned dataset of related variables.

[0033] By combining cluster analysis with outlier removal, data cleaning is performed on multi-dimensional correlated variable datasets, making the data samples more consistent in terms of time and delivery characteristics, reducing the interference of outlier data on subsequent model training, and improving data quality and stability.

[0034] S22. Standardize the cleaned dataset of related variables to construct a standardized dataset of variables; Step S22 specifically includes: S221. Perform type classification processing on the variables in the cleaned associated variable dataset, dividing the variables into time-related variables, categorical variables, and numerical variables; Among them, time-related variables include calibration date, time nodes corresponding to supply lead time, project milestone plan node time, and delivery time of supporting equipment; category-related variables include material category information; and numerical variables include supply lead time duration, production cycle duration, transportation cycle duration, and delivery deviation quantification value.

[0035] S222. Process time-type variables, categorical variables, and numerical variables separately: Time-related variables are subjected to unified time quantification processing, transforming them into continuous numerical time features based on a unified time benchmark to characterize the temporal sequence and time interval differences between different time nodes; calibration dates, project milestone schedule dates, and delivery dates of supporting equipment are uniformly converted into time differences relative to the benchmark time, forming continuous time feature variables.

[0036] The categorical variables are numerically encoded to convert them into computable numerical feature variables. Specifically, the material categories are converted using either unique coding or serial number coding to represent the differences between different material categories.

[0037] Numerical variables are processed using one or more of the following methods: Z-Score standardization, min-max standardization, robust standardization, or maximum absolute value standardization, to eliminate the influence of different units on model calculation and construct numerical features of a unified scale. Among them, Z-Score standardization achieves numerical normalization by calculating the mean and standard deviation of variables; min-maximum standardization achieves interval compression through extreme value mapping; robust standardization reduces the impact of outliers based on the median and interquartile range; and maximum absolute value standardization achieves scaling through the maximum absolute value. S223. Map the processed variables to a unified numerical space to generate a standardized variable dataset for subsequent material planning delivery date prediction modeling and analysis.

[0038] By classifying and uniformly quantifying time-related, categorical, and numerical variables, different types of data can be uniformly expressed within the same numerical space, improving the computability and adaptability of multi-dimensional correlated variable datasets.

[0039] It should be noted that step S2 optimizes the quality and unifies the structure of the original associated variable data by clustering and standardizing the multi-dimensional associated variable dataset. This enables the data to have both low noise characteristics and uniform dimensional features, improves the comparability and fusion ability between different types of variables, and reduces the impact of data heterogeneity on subsequent material planning and delivery date modeling and analysis.

[0040] Step S3: Based on the standardized variable dataset, establish a delivery schedule planning model, use a parameter solving algorithm to solve the delivery schedule planning model, and generate the planned delivery dates for various materials; Step S3 specifically includes: S31. Based on a standardized variable dataset, a delivery date planning model is constructed, which is used to characterize the comprehensive influence of each related variable on the material plan delivery date; Step S31 specifically includes: S311. Extract calibration date, supply lead time, engineering milestone plan, material category, supporting equipment delivery date, and supporting material delivery date variables from the standardized variable dataset as input variables. Each input variable is represented in the numerical form found in the standardized variable dataset. Construct the input feature vector: ,in, Let represent the i-th standardized correlation variable, and n represent the number of correlation variables; Use the planned delivery date of materials as the output variable; S312. Construct a functional mapping relationship for the material plan delivery date based on the input feature vector, establish a delivery period planning model in the form of linear combination, and express the material plan delivery date as the weighted sum of each input variable; The delivery schedule planning model expression is: Where y represents the planned delivery date of the materials. represents the weight coefficient corresponding to the i-th standardized association variable, and b represents the bias term.

[0041] S313. Set constraints on the weighting coefficients in the delivery schedule planning model to ensure that the results of the delivery schedule planning model conform to actual business rules. The constraints include:

[0042]

[0043] By constraining the sum of the weight coefficients to 1, a normalized expression of the influence of each related variable is achieved. By constraining the weight coefficients to be non-negative, the direction of the influence of each variable on the delivery date is ensured to be consistent.

[0044] S314. Based on the weights of each associated variable obtained in step S13, use the weights of each associated variable as the initial values ​​of the weight coefficients of the delivery date planning model, and initialize the bias term to obtain the initial parameters of the delivery date planning model. S315. Construct an error function for the delivery schedule planning model to measure the deviation between the model output and the actual delivery date; The error function expression is as follows:

[0045] in, This indicates the actual delivery date of the k-th sample. This represents the model's predicted delivery date, and N represents the sample size.

[0046] By constructing a delivery schedule planning model, the standardized variable dataset is uniformly mapped to the input feature vector, realizing a structured expression of the factors affecting the material plan delivery date, and improving the consistency of expression and modeling foundation between various related variables and the material plan delivery date.

[0047] S32. Construct a multivariate regression model, and use quadratic programming to train and optimize the parameters in the multivariate regression model to obtain a parameter solving algorithm suitable for the delivery schedule planning model; Step S32 specifically includes: S321. Establish a linear mapping relationship between each associated variable and the actual delivery date, construct a multiple regression model, and perform parameter modeling on the weight coefficients and bias terms in the delivery schedule planning model; S322. Construct a quadratic programming optimization model with the objective function of minimizing the deviation between the predicted delivery date and the actual delivery date. Introduce weight constraints to construct a quadratic programming optimization model for solving the parameters of the multiple regression model. The quadratic programming objective function is as follows:

[0048] Where λ represents the regularization coefficient; The constraints are the same as the weighting coefficient constraints in the delivery schedule planning model, which are: This is used to ensure that the weights of each associated variable satisfy the normalization constraint and the non-negativity constraint.

[0049] S323. Based on the standardized variable dataset, the weight coefficients and bias terms in the multiple regression model are trained. The objective function is iteratively optimized using the quadratic programming solution method to gradually reduce the error between the predicted delivery date and the actual delivery date. During the iterative optimization process, the weight coefficients are constrained and corrected so that the model parameters meet the constraints and gradually converge to the optimal parameter solution. S324. When the parameters of the multivariate regression model meet the convergence condition or reach the number of iterations, output the weight coefficients and bias terms after training is completed, and generate a parameter solving algorithm suitable for the delivery schedule planning model.

[0050] By constructing a multiple regression model and introducing a quadratic programming optimization method to train and solve the model parameters, the weight coefficients and bias terms achieve optimal fitting under the conditions of normalization constraints and non-negativity constraints. This improves the stability and accuracy of the delivery time planning model parameter solution and enhances the model's adaptability to delivery times of different material types.

[0051] S33. Substitute the input feature vector directly into the delivery date planning model, call the weight coefficients and bias terms output by the parameter solving algorithm, perform weighted summation calculation, and superimpose the bias terms to calculate and process the delivery date planning model to obtain the planned delivery dates of various materials.

[0052] It should be noted that by constructing a delivery schedule planning model and combining it with a parameter solving algorithm to solve the model, the process of generating material plan delivery dates is transformed from experience-based judgment to data-driven quantitative calculation, which improves the objectivity and consistency of delivery schedule planning, as well as the rationality and feasibility of the delivery schedule planning results.

[0053] Step S4: Based on the generated planned delivery date, and in conjunction with the engineering milestone plan and material supply schedule, output the optimized material procurement plan delivery schedule.

[0054] Step S4 specifically includes: S41. Based on the generated material plan delivery date, match and verify it with the project milestone plan and material supply schedule to establish a delivery date constraint relationship; By comparing the planned delivery date of materials with the planned milestone dates of the project, it can be determined whether the planned delivery date of materials meets the project milestone date requirements; By combining the material supply schedule, the planned delivery dates of materials are checked to see if they meet the constraints of production and transportation cycles, and materials with time conflicts or mismatches are identified.

[0055] The engineering milestone plan mentioned in step S41 is used to characterize the time requirements of key engineering nodes, and the material supply lead time is used to characterize the time cycle required for materials from production to delivery.

[0056] By matching and verifying the planned delivery dates of materials with the project milestone plan and the material supply lead time, a delivery date constraint relationship is established, enabling feasibility testing of the delivery date results. This effectively identifies abnormal delivery arrangements that do not meet the project milestone requirements or supply lead time constraints, thereby improving the consistency of delivery date constraints.

[0057] S42. The planned delivery date of materials that has passed the delivery date constraint verification shall be directly used as the final material procurement plan delivery date; For material delivery dates that do not meet the project milestone plan or material supply lead time constraints, time adjustments are made based on the constraints to ensure that the adjusted delivery dates simultaneously meet the project milestone plan time requirements and material supply lead time constraints. Generate an optimized material procurement plan and delivery schedule.

[0058] It should be noted that by introducing engineering milestone plans and material supply lead times for constraint verification and correction, the delivery schedule planning results are transformed from single model predictions into optimized results that meet both engineering schedule and supply capacity constraints. By correcting conflicting delivery dates and retaining feasible delivery dates, the final material procurement plan delivery schedule simultaneously meets engineering schedule requirements and material supply lead time constraints, improving the matching degree between the material procurement plan delivery schedule and the actual engineering implementation process, and enhancing the rationality and engineering adaptability of the overall delivery schedule planning.

[0059] Example 2: like Figure 2 As shown in the figure, this embodiment provides a delivery time optimization system for material procurement planning, including: Module 1 for selecting related variables, in which: From the full historical material procurement data, we screened out the related variables that affect the planned delivery time of materials and constructed a multi-dimensional related variable dataset. The full historical material procurement data in the correlation variable selection module 1 includes: Material procurement plan data, bidding batch schedule data, contract signing data, supply lead time data, actual delivery date data, and project milestone plan data; The correlation variable selection module 1 specifically includes: Based on the business rules for power grid electrical equipment and materials procurement, a full dataset is constructed by extracting all historical material procurement data from the historical material procurement database. From the full dataset, the calibration date, supply lead time, engineering milestone plan, material category, delivery date of supporting equipment, and delivery date of supporting materials are extracted as candidate variables. The correlation coefficients between the candidate variables and the actual delivery dates are calculated. Candidate variables were selected based on the correlation coefficient ranking results, business time transmission link relationships, and data integrity constraints; Variables that meet the conditions of correlation threshold, consistency of time transmission relationship and data integrity are identified as associated variables; The related variables are divided into primary and secondary variables according to their degree of influence on the delivery date of the material plan, and corresponding weights are generated. This ultimately results in a multi-dimensional dataset of related variables.

[0060] Data processing module 2, in which: Data processing operations are performed on the multidimensional correlation variable dataset to obtain a standardized variable dataset that can be used for modeling and analysis; The data processing operations of the data processing module 2 include: Feature extraction is performed on a multi-dimensional correlated variable dataset to construct variable feature vectors; The similarity between samples is calculated based on the feature vectors, and a sample distance matrix is ​​generated. The samples are grouped based on a clustering algorithm using the distance matrix; Identify and remove outliers from the clustering results to generate a cleaned dataset of association variables. The cleaned dataset of related variables was subjected to type classification, dividing the variables into time-related variables, categorical variables, and numerical variables; Perform quantification processing on time-related variables using a unified time base; Encode the category variables; Standardize numerical variables; The processed variables are mapped to a unified numerical space to generate a standardized variable dataset.

[0061] Delivery schedule model planning module 3, in which: Based on a standardized variable dataset, a delivery schedule planning model is established, and a parameter solving algorithm is used to solve the delivery schedule planning model to calculate the material plan delivery date. The delivery schedule model planning module 3 specifically includes: Extract calibration date variables, supply lead time variables, engineering milestone plan variables, material category variables, supporting equipment delivery time variables, and supporting material delivery time variables from the standardized variable dataset, and construct the input feature vector; A delivery schedule planning model is constructed based on the input feature vector, and a linear mapping relationship is established between the material plan delivery date and various related variables. Set weight constraints for the delivery schedule planning model, including weight normalization constraints and non-negativity constraints; Initialize model parameters based on the weights of historical correlation variables, and construct an error function; A multiple regression model is constructed, and the model parameters are trained and solved using a quadratic programming optimization method to obtain the parameter solving algorithm; The input feature vector is substituted into the delivery date planning model, and the model parameters output by the parameter solving algorithm are called to calculate the planned delivery dates for various materials.

[0062] The planned output module 4 contains: Based on the generated planned delivery date, the constraints are checked in conjunction with the engineering milestone plan and material supply lead time, and the optimized material procurement plan delivery schedule is output.

[0063] The plan output module 4 specifically includes: Match and verify the material delivery dates with the project milestone schedule dates; Perform periodic constraint verification between the planned delivery date of materials and the material supply lead time; Identify planned delivery dates that do not meet engineering milestone time requirements or supply lead time constraints; The planned delivery date that meets the constraints will be directly used as the final material procurement plan delivery date; Adjust the time for planned delivery dates that do not meet the constraints to ensure that they simultaneously meet the project milestone plan and material supply schedule constraints. Output the optimized material procurement plan and delivery schedule.

[0064] The above-disclosed embodiments are merely preferred embodiments of the present invention, but the present invention is not limited thereto. Any non-creative variations that can be conceived by those skilled in the art, as well as any improvements and modifications made without departing from the principles of the present invention, should fall within the protection scope of the present invention.

Claims

1. A method for optimizing delivery time in a material procurement plan, characterized in that, Includes the following steps: Step S1: Select the associated variables that affect the planned delivery time of materials from the full historical material procurement data, and construct a multi-dimensional associated variable dataset; Step S2: Perform data processing operations on the multi-dimensional correlation variable dataset to obtain a standardized variable dataset that can be used for modeling and analysis; Step S3: Based on the standardized variable dataset, establish a delivery schedule planning model, use a parameter solving algorithm to solve the delivery schedule planning model, and generate the planned delivery dates for various materials; Step S4: Based on the generated planned delivery date, and in conjunction with the engineering milestone plan and material supply schedule, output the optimized material procurement plan delivery schedule.

2. The delivery time optimization method for a material procurement plan according to claim 1, characterized in that, Step S1 specifically includes: S11. Based on the business rules for the procurement of electrical equipment and materials for power grids, extract all historical material procurement data from the historical material procurement database and construct a full dataset. S12. Filter the associated variables that affect the delivery time of the material plan from the full dataset and construct a set of associated variables; S13. Based on the degree of influence of the associated variables on the material plan delivery time, the associated variables are divided into primary variables and secondary variables, and weights are assigned to each associated variable to form a multi-dimensional associated variable dataset.

3. The delivery time optimization method for a material procurement plan according to claim 2, characterized in that, Step S12 specifically includes: S121. Extract the actual delivery date from the full dataset as the target variable, and extract the award date, supply lead time, engineering milestone plan, material category, delivery date of supporting equipment and delivery date of supporting materials from the bidding batch schedule as candidate variables. S122. Calculate the correlation between each candidate variable and the actual delivery date to obtain the correlation coefficient for each candidate variable. The correlation coefficient is calculated using the Pearson correlation coefficient, and the calculation formula is as follows: in, This represents the candidate variable value for the i-th sample; This represents the actual delivery date value for the i-th sample; This represents the mean of the candidate variables; This represents the mean of actual delivery dates; n represents the sample size. S123. Based on the correlation coefficient, the candidate variables are sorted to characterize the influence of each candidate variable on the material plan delivery date; according to the time transmission relationship of bidding batch schedule - supply lead time - project milestone plan in the power grid electrical equipment and material procurement process, the candidate variables are matched with business links to identify the action path of each candidate variable in the formation of the material plan delivery date and determine whether it is in a direct or indirect action link; the candidate variables are checked for completeness and candidate variables with missing proportions exceeding the preset threshold or data anomalies are removed to ensure the data availability of the variables participating in the screening; S124. Select candidate variables that meet the requirements of correlation coefficient higher than the preset threshold, are in the business time transmission link and meet the data integrity requirements as related variables, and construct a set of related variables.

4. The delivery time optimization method for a material procurement plan according to claim 3, characterized in that, Step S13 specifically includes: S131. Based on the strength of the influence of the related variables on the actual delivery date, their position in the time chain, and the controllability of the variables, the related variables are classified into primary variables and secondary variables. By calculating the correlation coefficients between each associated variable and the actual delivery date, sorting them according to the absolute value of the correlation coefficients, and combining them with the time transmission relationship, we can identify the position of each associated variable in the process of forming the material plan delivery date and determine whether it is in the direct or indirect link of the time transmission link. The correlation variables that rank first by absolute value of correlation coefficient and are directly involved in the time transmission link are classified as main variables. The main variables include the award date, supply lead time and project milestone plan in the bidding batch schedule. Related variables that constrain the main variables through matching relationships or affect the material delivery time through indirect paths are classified as secondary variables. The secondary variables include the delivery time of supporting equipment and the delivery time of supporting materials. S132. Perform weight allocation processing on the primary and secondary variables after partitioning to generate the weight values ​​corresponding to each related variable.

5. The delivery time optimization method for a material procurement plan according to claim 4, characterized in that, Step S132 specifically includes: S1321. Calculate the initial weights based on the correlation coefficients between each associated variable and the actual delivery date, and normalize the correlation coefficients to obtain the initial weights: in, Let m represent the correlation coefficient between the i-th associated variable and the actual delivery date, and m represent the total number of associated variables. S1322. Based on historical delivery deviation data, the initial weights of each related variable are corrected, the error contribution rate of each related variable is calculated, and the initial weights are adjusted according to the error contribution rate. The corrected weights are: in, This represents the historical error contribution of the i-th associated variable. Indicates the adjustment coefficient; S1323. Different weight constraints are set for primary and secondary variables, so that the overall weight of primary variables is higher than that of secondary variables, and secondary variables with multi-level matching relationships are subject to progressive weight decay according to the matching level; wherein, the multi-level matching weight decay formula is as follows: Where k represents the depth of the associated hierarchy. This represents the attenuation coefficient, and 0 < <1; S1324. Normalize the weights of all associated variables to obtain the final weight distribution, and combine it with the corresponding associated variables to generate a multi-dimensional associated variable dataset.

6. The delivery time optimization method for a material procurement plan according to claim 5, characterized in that, Step S3 specifically includes: S31. Construct a delivery schedule planning model based on a standardized variable dataset; S32. Construct a multivariate regression model, and use quadratic programming to train and optimize the parameters in the multivariate regression model to obtain a parameter solving algorithm suitable for the delivery schedule planning model; S33. Substitute the input feature vector directly into the delivery date planning model, call the weight coefficients and bias terms output by the parameter solving algorithm, perform weighted summation calculation, and superimpose the bias terms to calculate and process the delivery date planning model to obtain the planned delivery dates of various materials.

7. The delivery time optimization method for a material procurement plan according to claim 6, characterized in that, Step S31 specifically includes: S311. Extract calibration date, supply lead time, engineering milestone plan, material category, supporting equipment delivery date, and supporting material delivery date variables from the standardized variable dataset as input variables. Each input variable is represented using the numerical form found in the standardized variable dataset. Construct the input feature vector: ,in, Let represent the i-th standardized correlation variable, and n represent the number of correlation variables; Use the planned delivery date of materials as the output variable; S312. Construct a functional mapping relationship for the material plan delivery date based on the input feature vector, establish a delivery period planning model in the form of linear combination, and express the material plan delivery date as the weighted sum of each input variable; The delivery schedule planning model expression is: Where y represents the planned delivery date of the materials. represents the weight coefficient corresponding to the i-th standardized association variable, and b represents the bias term; S313. Set constraints on the weighting coefficients in the delivery schedule planning model to ensure that the results of the delivery schedule planning model conform to actual business rules. The constraints include: By constraining the sum of the weight coefficients to 1, a normalized expression of the influence of each related variable is achieved. By constraining the weight coefficients to be non-negative, the direction of the influence of each variable on the delivery date is ensured to be consistent. S314. Based on the weights of each associated variable obtained in step S13, use the weights of each associated variable as the initial values ​​of the weight coefficients of the delivery date planning model, and initialize the bias term to obtain the initial parameters of the delivery date planning model. S315. Construct an error function for the delivery schedule planning model to measure the deviation between the model output and the actual delivery date; The error function expression is as follows: in, This indicates the actual delivery date of the k-th sample. This represents the model's predicted delivery date, and N represents the sample size.

8. The delivery time optimization method for a material procurement plan according to claim 7, characterized in that, Step S32 specifically includes: S321. Establish a linear mapping relationship between each associated variable and the actual delivery date, construct a multiple regression model, and perform parameter modeling on the weight coefficients and bias terms in the delivery schedule planning model; S322. Construct a quadratic programming optimization model with the objective function of minimizing the deviation between the predicted delivery date and the actual delivery date. Introduce weight constraints to construct a quadratic programming optimization model for solving the parameters of the multiple regression model. The quadratic programming objective function is as follows: Where λ represents the regularization coefficient; The constraints are the same as the weighting coefficient constraints in the delivery schedule planning model, which are: ； S323. Based on the standardized variable dataset, the weight coefficients and bias terms in the multiple regression model are trained. The objective function is iteratively optimized using the quadratic programming solution method to gradually reduce the error between the predicted delivery date and the actual delivery date. During the iterative optimization process, the weight coefficients are constrained and corrected so that the model parameters meet the constraints and gradually converge to the optimal parameter solution. S324. When the parameters of the multivariate regression model meet the convergence condition or reach the number of iterations, output the weight coefficients and bias terms after training is completed, and generate a parameter solving algorithm suitable for the delivery schedule planning model.

9. A delivery time optimization system for material procurement planning, characterized in that, include: The module for selecting related variables is (1), the module for data processing is (2), the module for planning the delivery date model is (3), and the module for plan output is (4). The associated variable selection module (1) includes: From the full historical material procurement data, we screened out the related variables that affect the planned delivery time of materials and constructed a multi-dimensional related variable dataset. The data processing module (2) includes: Data processing operations are performed on the multidimensional correlated variable dataset to obtain a standardized variable dataset that can be used for modeling and analysis; The delivery schedule model planning module (3) includes: Based on a standardized variable dataset, a delivery schedule planning model is established, and a parameter solving algorithm is used to solve the delivery schedule planning model to calculate the material plan delivery date. The plan output module (4) contains: Based on the generated planned delivery date, the constraints are checked in conjunction with the engineering milestone plan and material supply lead time, and the optimized material procurement plan delivery schedule is output.

10. The delivery time optimization system for material procurement planning according to claim 9, characterized in that, The data processing operations of the data processing module (2) include: Feature extraction is performed on a multi-dimensional correlated variable dataset to construct variable feature vectors; The similarity between samples is calculated based on the feature vectors, and a sample distance matrix is ​​generated. The samples are grouped based on a clustering algorithm using the distance matrix; Identify and remove outliers from the clustering results to generate a cleaned dataset of related variables. The cleaned dataset of related variables was subjected to type classification, dividing the variables into time-related variables, categorical variables, and numerical variables; Perform quantification processing on time-related variables using a unified time base; Encode the category variables; Standardize numerical variables; The processed variables are mapped to a unified numerical space to generate a standardized variable dataset. The plan output module (4) specifically includes: Match and verify the material delivery dates with the project milestone schedule dates; Perform periodic constraint verification between the planned delivery date of materials and the material supply lead time; Identify planned delivery dates that do not meet engineering milestone time requirements or supply lead time constraints; The planned delivery date that meets the constraints shall be taken as the final material procurement plan delivery date; Adjust the time for planned delivery dates that do not meet the constraints to ensure that they simultaneously meet the project milestone plan and material supply schedule constraints. Output the optimized material procurement plan and delivery schedule.

Images