Double-objective optimization method for combined replenishment of hydropower enterprises

Abstract

The invention relates to the technical field of supply chain management, and provides a dual-objective optimization method for joint replenishment of water and power enterprises. The method comprises the steps of querying a historical replenishment record, performing similar clustering on hydropower enterprises according to the record, and determining an enterprise group; material demand prediction is executed, and a demand material set is determined; historical supply records of suppliers are collected, and supply credibility is analyzed and determined; screening suppliers according to the supply credibility by taking meeting the demand material set as a constraint, and determining supply mapping; carrying out transportation path optimization based on the demand node set and the supply mapping, and determining an optimal replenishment path; and executing joint replenishment according to the demand material set and the optimal replenishment path. The technical problem of low replenishment efficiency caused by high demand complexity and high supplier screening difficulty in the combined replenishment process of hydropower enterprises is solved, efficient replenishment resource allocation and path optimization are achieved through historical data mining and supply and demand matching optimization, and the replenishment efficiency is improved. And the overall efficiency and reliability of combined replenishment are improved.

Description

Technical Field

[0001] The present application relates to the technical field of supply chain management, and in particular to a dual-objective optimization method for joint replenishment of hydropower enterprises. Background Art

[0002] With the development of modern supply chain management technology, the efficiency and cost control of hydropower enterprises in material replenishment have become important factors in enterprise operations. Due to the wide geographical distribution of hydropower enterprises and the wide variety of materials they need, the model of independent material procurement and replenishment by a single enterprise often leads to problems such as unreasonable inventory management, increased procurement costs, and low transportation efficiency. Especially when demand fluctuates greatly, it is difficult to accurately predict the replenishment cycle, which can easily cause shortages or surpluses of materials, thereby affecting production operations. At present, some hydropower enterprises have adopted supply chain optimization methods to improve replenishment efficiency through information technology. However, most of these methods are optimized for a single enterprise and fail to fully consider the integration and sharing of resources among multiple enterprises. In addition, supplier selection and transportation route planning are often carried out separately, lacking comprehensive consideration of supplier performance capabilities and supply reliability, which leads to increased uncertainty in supply chain operations. Summary of the invention

[0003] This application provides a dual-objective optimization method for joint replenishment of hydropower enterprises, aiming to solve the technical problem of low replenishment efficiency of hydropower enterprises due to high demand complexity and difficulty in supplier screening during the joint replenishment process.

[0004] In view of the above problems, this application provides a dual-objective optimization method for joint replenishment of hydropower enterprises.

[0005] The present application provides a dual-objective optimization method for joint replenishment of hydropower enterprises, the method comprising: accessing an information sharing platform to query historical replenishment records, performing similarity clustering on multiple hydropower enterprises according to the historical replenishment records, and determining multiple enterprise groups; executing material demand forecasting for the multiple enterprise groups, and determining multiple demand material sets, wherein each demand material includes a demand node; collecting multiple historical supply records of multiple suppliers, and determining multiple supply credibility based on analysis of the multiple historical supply records; taking satisfaction of the multiple demand material sets as a constraint, screening suppliers according to the multiple supply credibility, and determining multiple supply mappings; performing transportation path optimization analysis based on the multiple demand node sets and the multiple supply mappings, and determining multiple optimal replenishment paths, and executing joint replenishment of hydropower enterprises according to the multiple demand material sets and the multiple optimal replenishment paths.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0007] The above-mentioned dual-objective optimization method for joint replenishment of hydropower enterprises first accesses the information sharing platform to query the historical replenishment records of multiple hydropower enterprises, performs data analysis and similarity clustering on these enterprises, and thus divides enterprises with similar replenishment characteristics into multiple enterprise groups. This can effectively improve the accuracy of replenishment coordination and avoid redundancy or insufficiency problems when a single enterprise replenishes independently. Subsequently, material demand forecasts are conducted for each enterprise group to generate multiple demand material sets, each of which corresponds to multiple demand nodes, to ensure that the actual material demand of each enterprise can be accurately grasped. This process significantly reduces the prediction error, thereby improving the accuracy of the replenishment plan. Afterwards, the historical supply records of multiple suppliers are collected, and the supply credibility of the suppliers is analyzed based on these records. Through credibility analysis, enterprises can select suppliers with strong performance capabilities and reliable supply, and avoid supply delays or material shortages caused by the instability of suppliers. On the premise of meeting the needs of the enterprise, the supplier's supply credibility is screened and multiple supply mappings are determined, so that the enterprise can match the most suitable supplier for replenishment, further improving the stability and response speed of the supply chain. Finally, based on the demand node set and supply mapping, the transportation path is optimized and analyzed to determine the optimal replenishment path. This link can not only effectively shorten the transportation distance and reduce transportation costs, but also improve the efficiency of material distribution through path optimization to ensure that materials arrive in time. This method achieves efficient, accurate and stable material replenishment for hydropower enterprises through comprehensive joint optimization from demand forecasting, supplier screening to transportation path optimization.

[0008] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0010] Figure 1 A schematic flow chart of a dual-objective optimization method for joint replenishment of hydropower enterprises in one embodiment;

[0011] Figure 2 The figure is a flow chart of determining multiple enterprise groups in a dual-objective optimization method for joint replenishment of hydropower enterprises in one embodiment. DETAILED DESCRIPTION

[0012] The embodiment of the present application solves the technical problem of low replenishment efficiency of hydropower enterprises due to high demand complexity and difficulty in supplier screening during the joint replenishment process of hydropower enterprises by providing a dual-objective optimization method for joint replenishment of hydropower enterprises.

[0013] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0014] It should be noted that the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules that are not explicitly listed or inherent to these processes, methods, products or devices.

[0015] Examples, such as Figure 1 As shown, the present application provides a dual-objective optimization method for joint replenishment of hydropower enterprises, the method comprising:

[0016] The information sharing platform is accessed to query historical replenishment records, and multiple hydropower enterprises are clustered according to similarity based on the historical replenishment records to determine multiple enterprise groups.

[0017] In an embodiment of the present application, the system terminal accesses an information sharing platform, which contains historical replenishment records of multiple hydropower enterprises. These records usually include data such as the material demand, replenishment time, replenishment quantity, and replenishment frequency of each enterprise in the past. By analyzing these historical replenishment records, the characteristics of each enterprise in replenishment behavior can be extracted. Subsequently, using these analysis results, multiple hydropower enterprises are clustered by similarity. The purpose of clustering is to group enterprises with similar replenishment characteristics into the same group, such as close replenishment cycles, similar material demand types, and close geographical locations. This clustering method helps optimize the replenishment process because the demand and supply patterns of similar enterprises are more consistent, which facilitates unified planning and coordination of replenishment. In this way, multiple enterprises can be reasonably divided into different enterprise groups, and the enterprises in each group have a high degree of consistency in replenishment demand. In this way, not only can the efficiency of joint replenishment of enterprises be improved, but also the complexity caused by differences in demand of individual enterprises can be reduced.

[0018] Further, if Figure 2 As shown, the present application provides similar clustering of multiple hydropower enterprises based on historical replenishment records to determine multiple enterprise groups, including:

[0019] Acquire multiple historical replenishment record data sets of multiple hydropower enterprises, perform high-frequency analysis on the multiple historical replenishment record data sets, and determine multiple high-frequency replenishment features; collect multiple enterprise location coordinates, perform similarity clustering on the multiple hydropower enterprises according to the multiple enterprise location coordinates and the multiple high-frequency replenishment features, and determine multiple enterprise groups.

[0020] Preferably, the system terminal first obtains a historical replenishment record data set of multiple hydropower enterprises. Each enterprise will store past replenishment records, and these data sets contain information such as the time of each replenishment, the type of replenished materials, quantity, and replenishment frequency. Through these data, we can fully understand the changes in material demand of each enterprise in different time periods. Subsequently, these historical replenishment record data sets are subjected to high-frequency analysis, and the characteristics of the replenishment behavior of each enterprise that meet the predetermined coefficient threshold are found through statistics and analysis, and the high-frequency replenishment characteristics of each enterprise are determined. Common high-frequency characteristics include the demand for specific types of materials, the regular cycle of replenishment, etc. These high-frequency characteristics can reflect the key demand patterns of enterprise replenishment. After that, the geographical location information of each enterprise is collected, and the spatial information of the enterprise is supplemented by obtaining the longitude and latitude coordinates of the specific location of the enterprise, and then the high-frequency replenishment characteristics are combined with the geographical location information of each enterprise for similarity clustering. This process is to classify enterprises with similar replenishment characteristics and close geographical locations into the same group, thereby forming multiple enterprise groups. Specifically, the system terminal first standardizes these different types of data. Because the numerical ranges of geographical location and replenishment characteristics are different, direct clustering will cause data of a certain dimension to have unnecessary influence on the results. Through normalization or Z-score standardization, it can be ensured that the geographic coordinates and replenishment characteristics are compared on the same scale to eliminate numerical deviations. After standardization, the system terminal calculates the similarity between each pair of enterprises. The calculation of similarity is divided into two parts. The first part is the similarity of geographical locations. Euclidean distance is usually used to calculate the spatial distance between enterprises. This can measure the geographical proximity of two enterprises, which is recorded as geographical similarity. The second part is the similarity of replenishment characteristics, which can usually be measured using cosine similarity or Euclidean distance. In this step, the system terminal will compare the replenishment frequency, material type, replenishment cycle and other data of each enterprise to find out enterprises with similar replenishment demand patterns. Taking cosine similarity as an example, assuming that the replenishment characteristics of an enterprise are composed of three dimensions: replenishment frequency f, replenishment quantity q and replenishment cycle p, then the replenishment feature vectors of enterprise i and enterprise j are V respectively. i =[f i ,q i ,p i ] and V j =[f j ,q j ,p j], and then use the cosine similarity calculation formula to calculate the replenishment feature vector V of enterprise i i and the replenishment feature vector V of enterprise j j The cosine similarity between them is recorded as the replenishment feature similarity. In order to obtain the comprehensive similarity, the system terminal will assign different weights to the geographical location similarity and the replenishment feature similarity. By weighted calculation of the geographical location similarity and the replenishment feature similarity, the comprehensive similarity is obtained, where the sum of the weights of the geographical location similarity and the replenishment feature similarity is 1. The specific value is dynamically adjusted according to the different needs of the business scenario. For example, when the geographical location has a greater impact, the weight of the geographical location similarity can be increased. After completing the calculation of the comprehensive similarity between all enterprises, a comprehensive similarity matrix is ​​constructed. Once the comprehensive similarity matrix is ​​constructed, the system terminal will use a hierarchical clustering algorithm to group the enterprises. Taking the agglomerative hierarchical clustering algorithm as an example, the system terminal regards each enterprise as a separate class and gradually merges them according to the comprehensive similarity between the enterprises. In each iteration, the two enterprises or enterprise groups with the highest similarity will be found and merged into a new group. The comprehensive similarity of the merged new group can be calculated by the average distance method, that is, the comprehensive similarity between the enterprises in the new group is the average of the comprehensive similarities between all enterprises in the two original groups. This method can avoid the overall clustering results being affected by the extreme similarity of individual points. The clustering process is repeated until all enterprises are merged into a large cluster, or the number of enterprise groups preset according to business needs is reached. At this time, a dendrogram will be generated. This dendrogram shows the hierarchical structure of enterprise clustering. The height of each layer represents the similarity between enterprises or enterprise groups. The lower the node, the higher the similarity. By analyzing the dendrogram, the system terminal can select the appropriate clustering level according to actual needs. For example, the fault points in the dendrogram are obtained and segmented based on these faults, or the cutting selection can be based on the similarity of geographical distribution demand or replenishment characteristics. Finally, through this hierarchical clustering method, the system terminal can divide similar enterprises into multiple enterprise groups. The enterprises in each enterprise group have high consistency in geographical location and replenishment demand, which provides a solid foundation for subsequent joint replenishment, resource sharing and transportation route optimization. This method not only improves the efficiency of the replenishment process, but also optimizes transportation costs and improves the overall coordination ability of the supply chain.

[0021] Furthermore, the present application provides a method of performing high-frequency analysis on the plurality of historical replenishment record data sets to determine a plurality of high-frequency replenishment features, including:

[0022] Randomly select any historical replenishment record data set as a target record data set, and randomly select the first record data from the target record data set, wherein the first record data at least includes a first material type, a first replenishment quantity and a first replenishment cycle; perform deviation calculation on the first material type and other material types in the target record data set, and count the data volume whose deviation value is less than the deviation index, which is recorded as a first high-frequency coefficient; if the first high-frequency coefficient is greater than a predetermined coefficient threshold, perform deviation calculation on the first replenishment quantity and other replenishment quantities in the target record data set to obtain a second high-frequency coefficient; if the second high-frequency coefficient is greater than a predetermined coefficient threshold, perform deviation calculation on the first replenishment cycle and other replenishment cycles in the target record data set to obtain a third high-frequency coefficient; if the third high-frequency coefficient is greater than the predetermined coefficient threshold, add the first record data to the first high-frequency data set, perform mean operation on the first high-frequency data set, obtain a first high-frequency replenishment feature, and add it to the multiple high-frequency replenishment features.

[0023] Optionally, the system terminal randomly selects a historical replenishment record data set of an enterprise as a target record data set. The target record data set contains all replenishment information of the enterprise in the past period of time. In this target record data set, a replenishment record is randomly extracted as the first record data. This record at least includes key information such as the material type, replenishment quantity and replenishment cycle of this replenishment. Subsequently, the material type in this first record data is compared with the material type of other replenishment records in the target record data set, and the deviation value between them is calculated. This deviation refers to the degree of difference between different material types. If the data deviation of the material type of some records from the first record is less than the set deviation index, then these records will be counted together and counted to generate a first high-frequency coefficient. This coefficient reflects the frequency of the first material type in the entire replenishment record. If this first high-frequency coefficient is greater than the predetermined coefficient threshold, the deviation calculation will be performed on the replenishment quantity in the first record data and other replenishment quantities in the target data set. Similarly, the replenishment quantity records with a deviation value less than the deviation index will be counted together to generate a second high-frequency coefficient, indicating the high frequency of the replenishment quantity. When the second high-frequency coefficient also exceeds the predetermined coefficient threshold, the deviation calculation of the replenishment cycle in the first record data and other replenishment cycles in the target data set will continue to be performed to generate a third high-frequency coefficient, which represents the high frequency of the replenishment cycle. If the third high-frequency coefficient is also greater than the predetermined coefficient threshold, the first record data will be added to the first high-frequency data set. Repeat this process until the analysis of the record data in the target data set is completed, or the data in the first high-frequency data set reaches a preset amount. At this time, the first high-frequency data set is averaged to obtain the first high-frequency replenishment feature, which is added to multiple high-frequency replenishment features. These features can reflect the frequently occurring patterns in the company's replenishment behavior and are used for subsequent optimization and decision analysis.

[0024] Execute material demand forecasting for the plurality of enterprise groups to determine a plurality of demand material sets, wherein each demand material includes a demand node.

[0025] In one embodiment, the system terminal performs material demand forecasting for each of the multiple enterprise groups that have been grouped to determine the materials that each enterprise group may need in the future. The core goal of this process is to predict the quantity and type of materials required by each enterprise group in the future. The system terminal uses a pre-built first demand predictor to analyze the historical replenishment records of each enterprise group to predict future material demand. By analyzing multiple enterprise groups, multiple demand material sets can be generated. Each demand material set represents the demand for certain materials by an enterprise group in the future. The demand material set includes detailed information such as material type, required quantity, and expected replenishment time. This prediction can help enterprises plan in advance and ensure the efficiency and accuracy of the replenishment process. Each demand material set has a demand node, which represents the time point and location of the specific demand. Through these demand nodes, the demand for materials of the enterprise in time and space can be accurately located, which helps the subsequent replenishment optimization work, provides a basis for subsequent replenishment decisions, and ensures that each enterprise group can replenish on demand.

[0026] Furthermore, the present application provides a method for executing material demand forecasting of the plurality of enterprise groups and determining a plurality of required material sets, including:

[0027] A first enterprise group is randomly selected, and a correlation analysis of prediction indicators is performed on the first enterprise group to determine a first correlation factor; multiple first prediction branches are constructed according to multiple first hydropower enterprises in the first enterprise group, and a first demand predictor is integrated and built, and a prediction accuracy analysis is performed on the first demand predictor using a verification data set to determine a first prediction accuracy; a first correlation factor is acquired based on the first correlation factor, and the first correlation factor is input into the first demand predictor to output a first predicted demand material set; prediction compensation is performed on the first predicted demand material set according to the first prediction accuracy, and a first demand material set is output, which is added to the multiple demand material sets.

[0028] Preferably, the system terminal randomly selects an enterprise group as the first enterprise group, analyzes the historical replenishment records of the enterprise group, and identifies a series of prediction indicators associated with demand changes, including material type and demand, replenishment cycle, etc., as the first correlation factor. Subsequently, the system terminal constructs multiple first prediction branches based on the historical data of multiple first hydropower enterprises in the enterprise group. These prediction branches are demand prediction models established separately for each enterprise. These demand prediction models can be constructed based on time series models (such as ARIMA models), regression models or machine learning models. Taking the long short-term memory network (LSTM) as an example, the system terminal constructs time series data for the historical replenishment records of each enterprise. In order for LSTM to capture the impact of historical data on future demand, the historical replenishment data of each enterprise will be sorted into a time window suitable for LSTM model processing in chronological order. For example, the replenishment records of the past month are used to predict the demand for the next month. Each sample will contain 1 month of data, and the output of the model is the predicted value for the next month. The LSTM model uses memory units to save key information from the past and makes predictions based on this information. In order to achieve accurate prediction of enterprise replenishment demand, an LSTM model is constructed for each enterprise. The LSTM model uses three main mechanisms: input gate, forget gate, and output gate to determine which historical information should be retained and which should be forgotten, and then predict future demand through memory and calculation. Once the model architecture is designed, the terminal begins to train the LSTM model for each enterprise. During training, historical replenishment data is used as input, and the LSTM unit processes the data at these time steps. The forward propagation process generates preliminary predictions. To evaluate the prediction results, the model output is compared with the actual replenishment demand and the error is calculated. A commonly used error metric is the mean square error (MSE), which measures the deviation between the model prediction results and the actual data. After the error is calculated, the back propagation algorithm is used to adjust the parameters of the LSTM model. The unique gating mechanism of LSTM allows it to dynamically adjust the content of the memory unit according to the changes in historical data, which enables the model to better capture the long-term dependencies in replenishment demand. During the back propagation process, the weight and bias parameters of the model are optimized based on the error feedback, making future predictions more accurate. During the training process, the model will be iterated multiple times to gradually reduce the prediction error and improve the prediction performance of the model. When the model is trained, the validation data set is used to evaluate its prediction ability. This step is completed by inputting historical data in the validation data set into the model, generating predictions, and comparing these predictions with actual demand. The performance of the current model is evaluated by calculating the accuracy on the validation data set. If the accuracy is greater than or equal to the preset expectation, the current LSTM model is output as the first prediction branch, and the current accuracy is recorded as the first prediction accuracy. These first prediction branches are then integrated to form the first demand predictor.Through these prediction branches, the system terminal can predict the future material demand of the enterprise based on the latest historical replenishment data. After that, based on the determined first correlation factor, the data corresponding to the first correlation factor is collected from the historical replenishment records that are a time window away from the current time, and these data are input into the first demand predictor as the first correlation factor. The purpose of this step is to use the current actual data to generate the future material demand forecast of the enterprise group and obtain the first predicted demand material set, which contains the type and quantity of materials that each enterprise in the first enterprise group may need in the future. However, since there may be a certain error in the prediction branch, the system terminal will perform prediction compensation on the generated predicted demand material set according to the first prediction accuracy corresponding to the prediction branch, that is, multiply the first prediction accuracy by the prediction result of the corresponding branch. Through this correction step, the system terminal will adjust the prediction result to make it closer to the actual demand that may occur, output the first demand material set, and add it to the demand material set of multiple enterprise groups, providing more accurate data support for subsequent replenishment plans.

[0029] A plurality of historical supply records of a plurality of suppliers are collected, and a plurality of supply credibility levels are determined based on analysis of the plurality of historical supply records.

[0030] In one embodiment, the system terminal collects historical supply records from multiple suppliers. The supply records of each supplier include detailed information such as the time of supply, material type, supply quantity, delivery time, supply cycle, and contract performance. These data provide a basis for analyzing the supply performance of suppliers. Subsequently, these historical supply records are analyzed, focusing on two aspects: one is the supplier's contract performance in past supply, that is, whether the supply is completed on time and whether it meets the agreed quality and quantity; the other is the demand satisfaction, checking whether the supplier can stably meet the material needs of the enterprise. These factors will affect the supplier's supply performance. In order to quantify the supplier's supply performance, the system terminal will perform data analysis on the historical records of each supplier, and calculate a contract performance reliability coefficient based on the frequency of supplier supply delays, supply shortages, and the incidence of supply quality problems. At the same time, it will also analyze whether the supplier can stably provide materials in long-term cooperation based on its ability to meet the needs of the enterprise, and calculate a service level coefficient for it. These analysis results will be combined to form an overall supply credibility score, which represents the supplier's comprehensive performance in stable supply, contract fulfillment, and timely response to demand. The higher the credibility of a supplier, the more reliable its supply is, and in future cooperation, the enterprise can rely on it for material replenishment with greater peace of mind. Finally, by analyzing the historical supply data of multiple suppliers, the supply credibility of each supplier was determined. This credibility information will play an important role in the subsequent supplier screening process, helping enterprises select the most reliable suppliers to meet their material needs.

[0031] Furthermore, the present application provides a method for analyzing and determining multiple supply credibility levels based on the multiple historical supply records, including:

[0032] A first historical supply record is randomly selected, wherein the first historical supply record includes a first demand satisfaction data sequence and a first fulfillment data sequence; a gradient weight of a record node sequence is configured according to a chronological order, and the first demand satisfaction data sequence and the first fulfillment data sequence are calculated according to the gradient weight to determine a first service level coefficient and a first fulfillment reliability coefficient; a first supply credibility is determined according to the first service level coefficient and the first fulfillment reliability coefficient, and is added to the multiple supply credibility.

[0033] Preferably, the system terminal randomly selects a supply record from the supplier's historical supply records as the first historical supply record. This record contains a demand satisfaction data sequence and a contract performance data sequence. The demand satisfaction data sequence reflects whether the supplier meets the material needs of the enterprise in the supply, for example, whether the supply volume is sufficient and whether it is completed within the agreed time. The contract performance data sequence reflects the supplier's contract performance, such as whether it is delivered on time and whether it complies with the delivery conditions stipulated in the contract. Subsequently, a gradual weight is assigned to each node in this supply record according to the chronological order. This gradual weight is to make the records of the recent time more influential than the records of the distant time, because the recent supply performance may better reflect the current capabilities of the supplier. Through this weight distribution, the system terminal can balance the long-term trend and recent performance in the historical records, so that the recent supply situation occupies a larger proportion in the calculation. Afterwards, the demand satisfaction data sequence and the contract performance data sequence are weighted and calculated using these assigned gradual weights, so as to obtain the service level coefficient and the contract performance reliability coefficient respectively. The service level coefficient measures the supplier's performance in meeting the needs of the enterprise, that is, whether the supplier can meet the material needs of the enterprise on time and in quantity, while the performance reliability coefficient reflects the supplier's performance in fulfilling the contract and delivering on time. Then, based on these two coefficients, the supplier's supply credibility is obtained by average calculation. Supply credibility is a comprehensive indicator that represents the overall reliability of the supplier in the supply process and can help enterprises determine whether the supplier can provide a stable and reliable supply of materials in future cooperation. Finally, the system terminal will add this supply credibility result to the supply credibility list of multiple suppliers to provide a basis for subsequent supplier selection and evaluation.

[0034] Taking satisfaction of the multiple demand material sets as a constraint, suppliers are screened according to the multiple supply credibility levels to determine multiple supply mappings.

[0035] In one embodiment, the system terminal sets the generated multiple demand material sets as constraints that must be met during the replenishment process. Each demand material set represents the demand of different enterprises for a certain material, including demand quantity, time requirements, etc. These requirements will serve as the core constraints when the system terminal selects suppliers to ensure that the material needs of all enterprises can be met. Subsequently, the calculated supply credibility of each supplier is used as the screening basis. The supply credibility reflects the reliability and stability of the supplier in the past supply process. The higher the credibility of the supplier, the more capable it is to provide materials on time and in quantity. Therefore, the system terminal will calculate the supply loss of each supplier based on these supply credibility to find the most suitable supplier for each enterprise's demand materials. After the screening is completed, multiple supply mappings are generated. Supply mapping refers to the matching relationship between a specific supplier and a specific demand. Each supply mapping specifies which supplier will provide which material to which enterprise, as well as the specific supply quantity and time. These supply mappings ensure that the material needs of the enterprise can be met by the right supplier and provide a basis for subsequent replenishment and transportation arrangements.

[0036] Furthermore, the present application provides screening suppliers according to the multiple supply credibility and determining multiple supply mappings, including:

[0037] A plurality of supply loss degrees are determined based on the plurality of supply credibility calculations, wherein the sum of the supply credibility and the supply loss degree is 1; loss compensation is performed on the plurality of demand material sets based on the plurality of supply loss degrees to determine a plurality of compensation material sets; and supplier screening is performed using the plurality of compensation material sets to determine a plurality of supply mappings.

[0038] Preferably, the system terminal calculates the corresponding supply loss degree according to the supply credibility of each supplier. The supply loss degree reflects the potential risks or uncertainties that may occur in the supplier's supply process. Since the sum of the supply credibility and the supply loss degree is 1, the higher the credibility of the supplier, the lower the loss degree, and vice versa. This loss degree measures the possibility that the supplier fails to deliver on time, in quantity, or fails to meet the demand. Subsequently, the compensation calculation is performed on the enterprise's demand material set based on these supply loss degrees. Since some suppliers have a high supply loss degree, the system terminal will compensate for these potential risks to ensure that the enterprise's material needs can be guaranteed. By adding additional material quantities, suppliers with high loss degrees are supplemented to ensure that the enterprise's needs are met. Through this compensation, the system terminal can generate multiple compensation material sets, that is, the demand material quantities adjusted after compensation or multiple suppliers are selected to share risks. Afterwards, suppliers are screened again based on these compensation material sets. At this time, the supplier screening should not only consider the supplier's supply capacity and credibility, but also combine the material demand after compensation to ensure that all material needs can be met through a reasonable combination of suppliers. In this way, the system terminal finally determined multiple supply mappings, which clearly defined which materials specific suppliers would be responsible for supplying, and took into account potential risks in the supply chain, making the entire supply process more stable and reliable.

[0039] Based on multiple demand node sets and multiple supply mappings, a transportation path optimization analysis is performed to determine multiple optimal replenishment paths, and joint replenishment of hydropower enterprises is performed according to the multiple demand material sets and the multiple optimal replenishment paths.

[0040] In one embodiment, the system terminal optimizes and analyzes the transportation path based on the demand node set and the determined *supply mapping of multiple enterprises. The demand node set represents the material demand of the enterprise, including the specific location and demand time point of each enterprise, while the supply mapping clarifies which suppliers are responsible for providing materials to which enterprises. Based on this information, the path optimization is performed based on the regional road network diagram. In this process, the system terminal analyzes the geographical location relationship, transportation distance, transportation time and other factors between each supplier and the enterprise, and finds the transportation path with better transportation adaptability. Then, these better transportation paths are compensated by factors such as traffic conditions and weather factors to obtain the optimal replenishment path for each enterprise. These optimal replenishment paths indicate the best transportation route and specific replenishment plan from the supplier to the enterprise, ensuring that each enterprise can receive the required materials at the right time. According to these optimized optimal replenishment paths, hydropower enterprises can efficiently perform joint replenishment. This not only ensures that the material needs of each enterprise are met in a timely manner, but also minimizes the transportation costs and time costs, and improves the operational efficiency of the entire supply chain.

[0041] Furthermore, the present application provides a method for performing transportation path optimization analysis based on multiple demand node sets and multiple supply mappings to determine multiple optimal replenishment paths, including:

[0042] A first demand node set of a first demand material of a first enterprise group is selected, and multiple first enterprise locations of the first demand node set are obtained; the first demand material is input into the multiple supply mappings for matching, multiple first association mappings are determined, and multiple first supplier locations of the multiple first association mappings are determined; based on a regional road network diagram, with less than the first demand node as a constraint, a transportation path optimization analysis is performed according to the multiple first supplier locations and the multiple first enterprise locations, a first optimal replenishment path is determined, and the first optimal replenishment path is added to the multiple optimal replenishment paths.

[0043] Preferably, the system terminal selects the first demand node set related to the first demand material in the first enterprise group. Each demand node represents the time point and location of the enterprise's demand for materials. According to the first demand node set, all relevant first enterprise locations, that is, the specific geographical location coordinates of these enterprises, are obtained, indicating the location where each enterprise needs to receive materials. Subsequently, the first demand material is matched with the previously generated supply mapping. The supply mapping has clarified which suppliers are responsible for providing materials to these enterprises. By matching the supply mapping, the system terminal determines the first association mapping corresponding to the demand materials of the first enterprise group, which indicates which suppliers are responsible for providing the first demand materials to these enterprises. At the same time, the locations of these first suppliers, that is, the specific geographical location coordinates of the suppliers, are also extracted. Once the geographical locations of the suppliers and enterprises are determined, the system terminal performs transportation path optimization analysis based on the road network diagram in the region through transportation path simulation enumeration and fitness evaluation. In this process, a key constraint condition needs to be met, that is, the transportation time must be earlier than the demand node time of each enterprise, that is, the materials must be delivered before the enterprise needs them. In order to meet this condition, the system terminal will ensure that the optimal transportation path starting from the supplier is not only the shortest distance or the lowest cost, but also ensures that each enterprise can receive the materials before the specified demand node. During the route optimization process, the system terminal will also comprehensively consider multiple factors, such as traffic conditions, weather conditions, etc. to conduct compensation analysis to ensure that the transportation process is as efficient as possible. Through optimization analysis, the system terminal will generate the first optimal replenishment path for the first demand node set to ensure that the materials can reach each enterprise on time. This path not only meets the transportation time constraints, but also minimizes transportation costs and distances. Finally, this optimal path is added to the overall multiple optimal replenishment paths for subsequent joint replenishment execution.

[0044] Further, the present application provides a method for performing a transportation path optimization analysis based on the plurality of first supplier locations and the plurality of first enterprise locations to determine a first optimal replenishment path, including:

[0045] Based on the regional road network diagram, a transportation path simulation enumeration is performed according to the multiple first supplier locations and the multiple first enterprise locations to generate multiple simulated transportation plans; a fitness evaluation is performed on the multiple simulated transportation plans to determine multiple transportation fitnesses, wherein the transportation fitness is negatively correlated with the simulated path distance; the multiple simulated transportation plans are screened according to the multiple transportation fitnesses to determine multiple better simulated transportation plans; a compensation analysis is performed on the multiple better simulated transportation plans in combination with weather conditions and road information, and the first optimal replenishment path is output.

[0046] Optionally, the system terminal performs a simulation enumeration of the transportation path based on the regional road network diagram, using the determined first supplier location and the first enterprise location. This step is to generate a variety of possible transportation plans by simulating different transportation paths. Each simulation plan represents a different transportation route from the supplier to the enterprise, taking into account factors such as the distance between each supplier and the enterprise, and possible roads. Subsequently, the generated simulated transportation plans are evaluated for fitness. The purpose of fitness evaluation is to measure the pros and cons of each path. The system terminal will calculate the fitness of each path based on the distance of the simulated path. Because the fitness is negatively correlated with the distance of the simulated path, the system terminal will normalize the distances of all simulated paths according to the maximum and minimum method, and then use 1 to calculate the ratio of the normalized distance of each simulated transportation plan to obtain the transportation fitness of each simulated transportation plan. Once the fitness calculation is completed, the system terminal arranges these transport fitnesses in descending order, and extracts the transport fitnesses that meet the preset optimal number from the transport fitnesses in descending order based on the preset optimal number, and uses the simulated transport schemes corresponding to these transport fitnesses as the optimal simulated transport schemes. For example, if the preset optimal number is 3, the first three transport fitnesses will be extracted from the transport fitnesses in descending order. These optimal simulated transport schemes represent the best possible path choices in the transportation process, and they show a good balance in terms of distance and other important factors. However, path simulation is not enough to cope with various dynamic factors in actual transportation, so the system terminal will also conduct compensation analysis on these optimal schemes in combination with weather conditions and road information. In this stage, the system terminal will consider external factors such as actual weather forecasts, road closures, and traffic congestion, and further adjust and optimize the transportation plan, that is, adjust the travel time and transportation cost of each path according to weather conditions and road information. For example, if a path passes through a congested or construction section, it will be marked to increase the transportation time; if a path is not suitable for passage due to weather reasons (such as blizzards), it will be marked as unavailable. After that, the simulated transport plans with unavailable marks are eliminated from multiple optimal simulated transport plans, and the remaining simulated transport plans are identified to see if they have marks for increasing transport time. If so, the transport time of the simulated transport plan is adjusted according to the actual situation. Then, the optimal simulated transport plan with the shortest transport time is selected from the multiple adjusted optimal simulated transport plans as the first optimal replenishment route to ensure the safe and timely delivery of materials.

[0047] In summary, the embodiments of the present application have at least the following technical effects:

[0048] The embodiment of the present application queries historical replenishment records and performs similarity clustering on multiple hydropower enterprises, determines the enterprise group, and then predicts the material demand and determines the demand node of each enterprise. Subsequently, the historical supply records of multiple suppliers are collected and analyzed, the supply credibility of the suppliers is calculated, and suppliers are screened based on this to determine the supply mapping. After that, based on the enterprise demand node set and supply mapping, combined with the transportation path optimization analysis, the optimal replenishment path is generated. In order to further optimize the path selection, the transportation path is compensated and analyzed according to weather conditions and road information to ensure the reliability and efficiency of the selected transportation path under actual conditions. These technical effects jointly solve the technical problem of low replenishment efficiency caused by high demand complexity and difficulty in supplier screening in the joint replenishment process of hydropower enterprises, and realize the efficient allocation and path optimization of replenishment resources through historical data mining and supply and demand matching optimization, thereby improving the overall efficiency and reliability of joint replenishment.

[0049] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. The processes depicted in the accompanying drawings do not necessarily require the specific order and continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0050] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

[0051] This specification and drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. A dual-objective optimization method for joint replenishment of hydropower enterprises, characterized by: include: Access the information sharing platform to query historical replenishment records, perform similarity clustering on multiple hydropower enterprises based on historical replenishment records, and determine multiple enterprise groups; Execute material demand forecasting of the plurality of enterprise groups to determine a plurality of demand material sets, wherein each demand material includes a demand node; Collecting multiple historical supply records of multiple suppliers, and analyzing and determining multiple supply credibility levels according to the multiple historical supply records; Based on the constraints of satisfying the multiple demand material sets, screening suppliers according to the multiple supply credibility, and determining multiple supply mappings; Based on multiple demand node sets and multiple supply mappings, a transportation path optimization analysis is performed to determine multiple optimal replenishment paths, and joint replenishment of hydropower enterprises is performed according to the multiple demand material sets and the multiple optimal replenishment paths.

2. The dual-objective optimization method for joint replenishment of hydropower enterprises according to claim 1 is characterized in that: Based on historical replenishment records, multiple hydropower enterprises were clustered and multiple enterprise groups were identified, including: Acquire multiple historical replenishment record data sets of multiple hydropower enterprises, perform high-frequency analysis on the multiple historical replenishment record data sets, and determine multiple high-frequency replenishment features; A plurality of enterprise location coordinates are collected, and a plurality of hydropower enterprises are similarly clustered according to the plurality of enterprise location coordinates and a plurality of high-frequency replenishment features to determine a plurality of enterprise groups.

3. The dual-objective optimization method for joint replenishment of hydropower enterprises according to claim 2 is characterized in that: A high-frequency analysis is performed on the multiple historical replenishment record data sets to determine multiple high-frequency replenishment features, including: Randomly select any historical replenishment record data set as a target record data set, and randomly select first record data from the target record data set, wherein the first record data at least includes a first material type, a first replenishment quantity, and a first replenishment cycle; Calculate the deviation between the first material type and other material types in the target record data set, and count the amount of data whose deviation value is less than the deviation index, which is recorded as the first high-frequency coefficient; If the first high-frequency coefficient is greater than a predetermined coefficient threshold, performing deviation calculation on the first replenishment quantity and other replenishment quantities in the target record data set to obtain a second high-frequency coefficient; If the second high-frequency coefficient is greater than a predetermined coefficient threshold, performing deviation calculation on the first replenishment period and other replenishment periods in the target record data set to obtain a third high-frequency coefficient; If the third high-frequency coefficient is greater than a predetermined coefficient threshold, the first record data is added to a first high-frequency data set, and a mean operation is performed on the first high-frequency data set to obtain a first high-frequency replenishment feature, which is added to the multiple high-frequency replenishment features.

4. The dual-objective optimization method for joint replenishment of hydropower enterprises according to claim 1 is characterized in that: Executing material demand forecasts for the plurality of enterprise groups to determine a plurality of required material sets includes: Randomly selecting a first enterprise group, performing a prediction indicator correlation analysis on the first enterprise group, and determining a first correlation factor; Constructing a plurality of first prediction branches according to a plurality of first hydropower enterprises of the first enterprise group, integrating and building a first demand forecaster, and performing a prediction accuracy analysis on the first demand forecaster using a validation data set to determine a first prediction accuracy; Acquire a first correlation factor based on the first correlation factor collection, input the first correlation factor into the first demand forecaster, and output a first predicted demand material set; The first predicted demand material set is predicted and compensated according to the first prediction accuracy, and a first demand material set is output and added to the plurality of demand material sets.

5. The dual-objective optimization method for joint replenishment of hydropower enterprises according to claim 1 is characterized in that: Analyzing and determining multiple supply credibility levels based on the multiple historical supply records includes: Randomly selecting a first historical supply record, wherein the first historical supply record includes a first demand satisfaction data sequence and a first contract fulfillment data sequence; Configure the gradual weight of the record node sequence according to the chronological order, calculate the first demand satisfaction data sequence and the first contract fulfillment data sequence according to the gradual weight, and determine the first service level coefficient and the first contract fulfillment reliability coefficient; A first supply credibility is determined according to the first service level coefficient and the first fulfillment reliability coefficient, and is added to the plurality of supply credibility.

6. The dual-objective optimization method for joint replenishment of hydropower enterprises according to claim 1 is characterized in that: Supplier screening is performed according to the multiple supply credibility levels to determine multiple supply mappings, including: Calculate and determine a plurality of supply loss degrees according to the plurality of supply credibility degrees, wherein the sum of the supply credibility degree and the supply loss degree is 1; Performing loss compensation on the multiple demand material sets according to the multiple supply loss degrees to determine multiple compensation material sets; The multiple compensation material sets are used to screen suppliers and determine multiple supply mappings.

7. The dual-objective optimization method for joint replenishment of hydropower enterprises according to claim 1 is characterized in that: Based on multiple demand node sets and multiple supply mappings, transportation path optimization analysis is performed to determine multiple optimal replenishment paths, including: Select a first demand node set of a first demand material of a first enterprise group, and obtain multiple first enterprise locations of the first demand node set; Inputting the first demand material into the multiple supply mappings for matching, determining multiple first association mappings, and determining multiple first supplier positions of the multiple first association mappings; Based on the regional road network diagram, with the constraint of being less than the first demand node, a transportation path optimization analysis is performed according to the multiple first supplier locations and the multiple first enterprise locations, and a first optimal replenishment path is determined and added to the multiple optimal replenishment paths.

8. The dual-objective optimization method for joint replenishment of hydropower enterprises according to claim 7 is characterized in that: Performing a transportation path optimization analysis according to the plurality of first supplier locations and the plurality of first enterprise locations to determine a first optimal replenishment path includes: Based on the regional road network diagram, a transport path simulation enumeration is performed according to the plurality of first supplier locations and the plurality of first enterprise locations to generate a plurality of simulated transport plans; Performing fitness evaluation on the multiple simulated transportation schemes to determine multiple transportation fitnesses, wherein the transportation fitness is negatively correlated with the simulated path distance; Screening the multiple simulated transportation plans according to the multiple transportation fitnesses to determine multiple better simulated transportation plans; A compensation analysis is performed on the multiple optimal simulated transportation plans in combination with weather conditions and road information, and the first optimal replenishment path is output.

Images