Scheduling method and system for collaborative distribution of processing raw materials and finished products

Abstract

The invention relates to the technical field of logistics distribution, and provides a scheduling method and system for collaborative distribution of processing raw materials and finished products, and the method comprises the steps: collecting multi-dimensional data in the distribution of the processing raw materials and the finished products; the multi-dimensional data comprises logistics information and node information; decomposing the logistics information to obtain logistics feature information; the logistics feature information comprises trend feature information, periodic feature information and random feature information; constructing a logistics matrix containing logistics nodes based on the node information, and calculating node feature information; the node feature information comprises an average connection number and an aggregation degree; based on the logistics feature information and the node feature information, designing a scheduling strategy to carry out collaborative distribution scheduling on the processing raw materials and the finished products; therefore, the utilization rate of logistics resources is improved, the response capability to market demands is enhanced, the logistics cost is effectively reduced, the overall efficiency and benefits of logistics distribution are improved, and a more competitive solution is provided for the field of logistics distribution.

Description

Technical Field

[0001] The present invention relates to the technical field of logistics distribution. Specifically, it relates to a scheduling method and system for collaborative distribution of processing raw materials and finished products. Background Art

[0002] With the booming development of e-commerce and manufacturing, logistics distribution, as a key link connecting production and consumption, its efficiency and cost have a decisive impact on enterprise competitiveness and customer satisfaction. In the traditional logistics distribution mode, the distribution of processing raw materials and the distribution of finished products are often separated and planned and executed independently. Enterprises usually manage the procurement and transportation of raw materials and the sales distribution of finished products separately, adopting relatively simple inventory management strategies and fixed transportation route arrangements to meet the basic needs of production and sales. However, there are many significant defects in this traditional mode: Low resource utilization rate: Due to the lack of collaborative planning, the empty driving rate of transportation vehicles for processing raw materials and finished products is relatively high, and the warehouse space is also difficult to be reasonably allocated, resulting in serious waste of logistics resources and high transportation costs and warehouse costs. Slow market response: The independent distribution mode cannot comprehensively consider the dynamic changes of market demand in real time, and it is easy to cause inventory backlogs or shortages of processing raw materials and finished products. Lack of in-depth data-driven: The traditional mode only stays on the surface of logistics information processing, only paying attention to basic data such as transportation volume and inventory quantity, and fails to deeply explore features such as trends and cycles in logistics information, as well as information such as connection relationships and aggregation degrees of logistics nodes, making it difficult to achieve precise optimization of distribution scheduling.

[0003] In view of this, to overcome the defects of the above traditional logistics distribution mode, the present application proposes a scheduling method and system for collaborative distribution of processing raw materials and finished products. By collecting multi-dimensional data, deeply decomposing logistics feature information, constructing a logistics matrix to analyze node features, and then designing a scientific scheduling strategy, the collaborative distribution scheduling of processing raw materials and finished products is realized. This solution aims to improve the utilization rate of logistics resources, enhance the response ability to market demand, optimize distribution scheduling with the help of in-depth data analysis, effectively reduce logistics costs, improve the overall efficiency and effectiveness of logistics distribution, and provide a more competitive solution for the logistics distribution field. Summary of the Invention

[0004] The object of the present invention is to provide a scheduling method for coordinated distribution of processing raw materials and finished products, including: collecting multi-dimensional data of the distributed processing raw materials and finished products; the multi-dimensional data includes logistics information and node information; decomposing the logistics information to obtain logistics feature information; the logistics feature information includes trend feature information, periodic feature information, and random feature information; constructing a logistics matrix including logistics nodes based on the node information, and calculating node feature information; the node feature information includes the average connection number and the aggregation degree; designing a scheduling strategy based on the logistics feature information and the node feature information to perform coordinated distribution scheduling on the processing raw materials and the finished products.

[0005] Further, constructing a logistics matrix including logistics nodes based on the node information includes: classifying the nodes in the logistics network for distributing processing raw materials and finished products based on the node information to obtain multiple types of logistics nodes; determining the connection relationship between the logistics nodes based on the historical transportation routes; creating an adjacency matrix based on the logistics nodes and their connection relationships, and using the adjacency matrix as the logistics matrix; the adjacency matrix is a square matrix, the order is the total number of logistics nodes, and the matrix elements represent the connection states between the logistics nodes.

[0006] Further, calculating the node feature information includes: determining the node degree of each logistics node; the node degree represents the number of other logistics nodes directly connected to the logistics node; averaging the node degrees of all logistics nodes to obtain the average connection number; determining the edge degree between the logistics nodes; the edge degree represents the number of edges with connection relationships between the other logistics nodes connected to the logistics node; calculating the clustering coefficient of the logistics node based on the average connection number and the edge degree; averaging the clustering coefficients of all logistics nodes to obtain the aggregation degree.

[0007] Further, designing a scheduling strategy based on the logistics feature information and the node feature information to perform coordinated distribution scheduling on the processing raw materials and the finished products includes: constructing a scheduling objective function based on the node feature information and the distribution volume; constructing scheduling constraint conditions based on the logistics feature information; solving the scheduling objective function to obtain the inventory of processing raw materials and the inventory of finished products at the logistics nodes; optimizing the scheduling objective function through a scheduling optimization algorithm to obtain an optimized distribution volume; adjusting the optimized distribution volume based on the random feature information to obtain an adjusted distribution volume; substituting the adjusted distribution volume into the scheduling objective function to obtain a new scheduling objective function; repeating the operations of solving and optimizing the scheduling objective function until the change values of the inventory of processing raw materials and the inventory of finished products are respectively less than a preset raw material inventory threshold and a preset finished product inventory threshold, and using the final distribution volume, the inventory of processing raw materials, and the inventory of finished products as the scheduling strategy.

[0008] Further, the expression of the scheduling objective function is: ; ; ; ; Among them, min represents taking the minimum value; Z represents the value of the scheduling objective function; represents the first scheduling objective parameter; represents the second scheduling objective parameter; represents the third scheduling objective parameter; i and j represent different logistics node variables, representing the first logistics node variable and the second logistics node variable respectively; N represents the total number of logistics nodes; represents the influence degree of the aggregation degree on the basic transportation cost; D represents the aggregation degree; and respectively represent the basic transportation cost of the processed raw materials and the basic transportation cost of the finished products from logistics node i to logistics node j; and respectively represent the distribution volume of the processed raw materials and the distribution volume of the finished products from logistics node i to logistics node j; represents the influence degree of the average connection number on the basic transportation time; represents the average connection number; represents the basic transportation time from logistics node i to logistics node j; k represents the warehouse node variable; K represents the total number of warehouse nodes; and respectively represent the unit storage cost of the processed raw materials and the unit storage cost of the finished products of warehouse node k; and respectively represent the inventory of the processed raw materials and the inventory of the finished products of warehouse node k; The scheduling constraint conditions include: Demand matching constraint: ; ; Among them, and respectively represent the total distribution volume of the processed raw materials and the total distribution volume of the finished products from logistics node i to all connected logistics nodes j; and respectively represent the demand for the processed raw materials and the demand for the finished products of node i; and respectively represent the demand fluctuation volume of the processed raw materials and the demand fluctuation volume of the finished products of node i, and the demand fluctuation volume is related to the trend characteristic information and the cycle characteristic information; represents the conversion coefficient for processing the processed raw materials into finished products; S represents the set of logistics nodes supplying the processed raw materials; represents the total amount of processed raw materials transported from logistics node j to logistics node i; Transport capacity constraint: ; Among them, represents the maximum delivery volume that can be carried from logistics node i to logistics node j; Inventory capacity constraint: ; ; Among them, and respectively represent the maximum storage capacity of processed raw materials and the maximum storage capacity of finished products at warehouse node k; Non - negative constraint: ; Delivery time constraint: ; Among them, represents the start time of delivery; t represents the actual delivery time; represents the end time of delivery.

[0009] Furthermore, through the scheduling optimization algorithm, the scheduling objective function is optimized to obtain the optimized delivery volume, including: determining the initial delivery volume based on logistics feature information; calculating the value of the scheduling objective function using the initial delivery volume to provide an initial evaluation basis for particle swarm optimization; substituting the position of each particle into the scheduling objective function to calculate the fitness value; the fitness value represents the quality of the delivery combination; the position of each particle is represented as a set of delivery volume combinations; adjusting the exploration range of particles in different regions according to the average connection number and aggregation degree of logistics nodes; when the preset number of iterations is met or the fitness value converges, terminate the iteration, and use the delivery volume combination corresponding to the global optimal position at this time as the final delivery volume.

[0010] Furthermore, adjusting the exploration range of particles in different regions according to the average connection number and aggregation degree of logistics nodes is as follows: at logistics nodes where the average connection number is higher than the connection threshold, narrow the search range of particles; at logistics nodes where the aggregation degree is higher than the aggregation threshold, guide the particles to search along the local optimal solution at this logistics node.

[0011] Furthermore, the calculation formulas for updating the individual optimal position and global optimal position of particles through particle swarm optimization are: ; ; ; ; Among them, and respectively represent the velocities of the particle at times ti and ti+1 in dimension ; represents the original inertia weight of the particle; represents the self-cognition coefficient of the particle; represents the first random number in the interval [0, 1]; represents the particle in dimension of the historical best position; and respectively represent the positions of the particle at times ti and ti+1 in dimension ; represents the social cognition coefficient of the particle; represents the second random number in the interval [0, 1]; represents the global best position of all particles in dimension ; represents the average connection number; represents the connection threshold; D represents the aggregation degree; represents the aggregation threshold; represents the modified inertia weight of the particle; represents the modified self-cognition coefficient of the particle; represents the maximum value of the node degree; represents the maximum value of the edge degree.

[0012] Furthermore, the expression for adjusting the delivery volume is obtained as: ; ; ; Among them, and respectively represent the adjusted delivery volumes of the processed raw materials and the finished products; represents the adjustment coefficient dynamically adjusted according to the random feature information; and respectively represent the actual residual standard deviations of the processed raw materials and the finished products; and respectively represent the predicted residual standard deviations of the processed raw materials and the finished products; and respectively represent the minimum adjustment coefficient and the maximum adjustment coefficient; Fr, and respectively represent the random feature information, the maximum value of the random feature information, and the minimum value of the random feature information.

[0013] The present invention also provides a scheduling system for the collaborative distribution of processing raw materials and finished products by using the scheduling method described in any one of the above, including a collection module, a decomposition module, a calculation module, and a scheduling module; the collection module is used to collect multi-dimensional data in the distribution of processing raw materials and finished products; the multi-dimensional data includes logistics information and node information; the decomposition module is used to decompose the logistics information to obtain logistics feature information; the logistics feature information includes trend feature information, periodic feature information, and random feature information; the calculation module is used to construct a logistics matrix including logistics nodes based on the node information and calculate node feature information; the node feature information includes the average connection number and the aggregation degree; the scheduling module is used to design a scheduling strategy based on the logistics feature information and the node feature information to perform collaborative distribution scheduling on the processing raw materials and the finished products.

[0014] The technical solution of the embodiment of the present invention has at least the following advantages and beneficial effects: By decomposing the logistics information to obtain trend feature information, periodic feature information, and random feature information, the present invention can accurately explore the internal laws of logistics activities, and reveal the characteristics of logistics nodes from the network structure level through node feature information. By analyzing the logistics feature information and the node feature information, a scheduling strategy is designed to realize the optimization of the collaborative distribution scheduling of processing raw materials and finished products, which helps to improve the distribution efficiency, reduce the logistics cost, enhance the overall utilization efficiency of logistics resources, and strengthen the adaptability of the logistics system to complex business scenarios.

[0015] By classifying the nodes in the logistics network, the present invention can clarify the roles and functions of different types of nodes in logistics activities, making the scheduling more targeted. Determining the connection relationship between logistics nodes based on the historical transportation route ensures an accurate grasp of the actual operation mode of the logistics network. Creating an adjacency matrix as the logistics matrix to visually present the logistics nodes and their connection status in the form of a square matrix provides a clear model basis for analyzing the correlation degree between nodes and the rationality of transportation routes. This way makes the design of the scheduling strategy more in line with the real structure of the logistics network, helps to optimize the transportation route, reduce ineffective transportation, enhance the scientificity and effectiveness of the scheduling scheme, and improve the overall operation efficiency of the logistics network.

[0016] The present invention provides a key indicator for the importance and influence of a node in the logistics network through the node degree. By calculating the node feature information through this indicator, the core nodes and dense areas in the logistics network can be more accurately identified, providing a solid foundation for the subsequent scheduling strategy. For example, narrowing the search range near the nodes with a high average connection number can concentrate resources to improve the distribution efficiency in key areas; guiding particles to search along the local optimal solution for the nodes with a high aggregation degree can utilize the regional compactness to accelerate optimization. Thereby improving the efficiency and effectiveness of the overall logistics distribution, and enhancing the stability and reliability of the logistics system.

[0017] The present invention uses the particle swarm optimization algorithm to optimize the scheduling objective function to obtain the optimized distribution volume. By utilizing the efficient global search ability of this algorithm, it can quickly explore a better distribution plan in complex distribution scenarios, improving the quality and efficiency of the distribution plan. Moreover, based on the logistics feature information, the initial distribution volume is determined, making the initial solution closely conform to the laws and characteristics (such as trends, cycles, etc.) of the actual logistics business, providing a more reasonable and targeted starting point for the particle swarm optimization, effectively reducing the ineffective iterations of the algorithm, accelerating the algorithm convergence, and approaching the optimal solution faster.

[0018] The present invention can avoid the meaningless extensive search of particles in the node area with rich but complex connections by narrowing the particle search range at the logistics nodes where the average connection quantity is higher than the connection threshold, concentrating the computing resources on the effective area near the node, quickly locking in the high-quality distribution path or resource allocation method, improving the search efficiency, and is especially suitable for the precise scheduling of key nodes such as logistics hubs.

[0019] The present invention guides the particles to search along the local optimal solution at the logistics nodes where the aggregation degree is higher than the aggregation threshold. By utilizing the characteristics of close node connections and strong cooperation in the area where the node is located, the particles can accelerate the optimization process with the stability and representativeness of the local optimal solution. When dealing with areas with a high aggregation degree such as urban intensive distribution areas, it can quickly adapt to the demand changes in the area, improving the accuracy and efficiency of the distribution scheduling of processing raw materials and finished products in the local area, and better coping with the complex node structure and demand distribution in collaborative distribution. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 is an exemplary flowchart of a scheduling method for collaborative distribution of processing raw materials and finished products provided by the present invention; Figure 2 is an exemplary module diagram of a scheduling system for collaborative distribution of processing raw materials and finished products provided by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0021] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Usually, the components of the embodiments of the present invention described and illustrated in the drawings here can be arranged and designed in various different configurations.

[0022] Figure 1 is an exemplary flowchart of a scheduling method for collaborative distribution of processing raw materials and finished products provided by the present invention. As Figure 1As shown in the figure, the scheduling method for coordinated distribution of processing raw materials and finished products provided by the present invention includes the following contents: Step 110, collect multi-dimensional data in the distribution of processing raw materials and finished products; the multi-dimensional data includes logistics information and node information. The processing raw materials may include raw sugar, sweeteners and preservatives, and the finished products may include syrup. The multi-dimensional data may refer to various data related to the distribution of raw materials and finished products, and can be obtained by extracting the historical distribution data of the distribution of processing raw materials and finished products. The logistics information may refer to information related to logistics transportation, and the logistics information may include demand fluctuations, traffic timeliness, raw material inventory, finished product inventory, and processing equipment status, etc. The logistics information may be the daily logistics information. The node information may refer to information related to logistics nodes, and the logistics information may include factory coordinates, warehouse coordinates, demander coordinates, raw material supplier coordinates, transportation routes and transportation frequencies, etc.

[0023] Step 120, decompose the logistics information to obtain logistics feature information; the logistics feature information includes trend feature information, periodic feature information and random feature information. The logistics feature information refers to the systematic information that can be used to represent the logistics situation after processing the logistics information. The trend feature information can reflect the overall change direction of the logistics in a relatively long period. For example, the procurement volume of processing raw materials shows a stable annual growth trend as the market expands, and the distribution volume of finished products increases in the long term due to the improvement of brand influence, etc. The periodic feature information refers to the information that the logistics shows regular repeated fluctuations. For example, the peak demand for raw materials every week, and the congestion of finished product distribution during the morning and evening rush hours every day, etc. The random feature information refers to the irregular sudden fluctuations of the logistics. For example, the interruption of raw material transportation during a certain period due to sudden natural disasters, or the change of the finished product distribution route caused by temporary policy adjustments, etc.

[0024] In some embodiments, the logistics information is decomposed to obtain logistics feature information, including: segmenting the logistics information to extract the temporal features of each time period; the temporal features may include the mean, variance, slope, and periodic fluctuation amplitude of the logistics information, etc. Combining the logistics business scenario, analyze the prior rules related to the trends and cycles of logistics; the prior rules include promotion activity cycle rules, holiday transportation peak rules, seasonal supply and demand fluctuation rules, etc. Match the similarity between the current logistics data and the historical logistics data, compare the matching result of the similarity with the prior rules, and separate the trend feature information and cycle feature information that conform to the rules. Take the difference between the original logistics data and the separated trend feature information and cycle feature information as the random feature information. The temporal features are used to reflect the variation law and statistical characteristics of the logistics information at different time points. For example, the daily processing raw material purchase data and product shipment data within the past 30 days can be divided into one time period per day, the divided logistics data is processed, and arranged in chronological order to obtain the temporal features of each time period. The logistics business scenario may include processes such as storage, transportation, distribution of goods, and related order processing and supply chain collaboration. The prior rules refer to the regular rules related to logistics trends and cycles summarized based on past experience, industry rules, or domain knowledge before analyzing the logistics data. The current logistics data refers to various data related to the transportation of processing raw materials and products during the current time period. The historical logistics data refers to various data related to the transportation of processing raw materials and products during the historical time period. For example, the current logistics data is the cargo transportation volume data of a certain logistics company this week, and the historical logistics data is the cargo transportation volume data of each week in the past two years. Various distance methods can be used to calculate the similarity between the transportation volume data this week and the transportation volume data of each week in the past two years. It is found that the similarity between the transportation volume data this week and the transportation volume data of the week in a certain period last year is very high. Comparing with the prior rules (for example, the cycle rule in this period is that the transportation volume gradually increases in the first three days and first high then low in the last four days), the separated trend feature information is that the transportation volume this week shows a trend of first increasing and then decreasing.

[0025] Step 130, construct a logistics matrix including logistics nodes based on the node information, and calculate the node feature information; the node feature information includes the average connection number and the aggregation degree. The logistics nodes may include processing plants, transportation hubs, warehouses, demanders, distribution centers, raw material suppliers, etc. The logistics matrix can be used to represent the transportation logistics situation of processing raw materials and finished products. The node feature information can be used to represent the connectivity between logistics nodes. The average connection number can reflect the connection richness of the logistics nodes in the network. The aggregation degree can reflect the tightness of the area where the logistics nodes are located.

[0026] In some embodiments, constructing a logistics matrix including logistics nodes based on node information includes: classifying the nodes in the logistics network for distributing and processing raw materials and finished products based on the node information to obtain multiple types of logistics nodes. The multiple types of logistics nodes may include processing plant nodes, transportation hub nodes, warehouse nodes, demand merchant nodes, distribution center nodes, raw material merchant nodes, etc. Determining the connection relationships between the logistics nodes based on the historical transportation routes. The historical transportation routes may refer to the routes for historically transporting raw materials and products for processing. The connection relationships may refer to the connectivity between the logistics nodes. Creating an adjacency matrix based on the logistics nodes and their connection relationships, and using the adjacency matrix as the logistics matrix; the adjacency matrix is a square matrix, the order of which is the total number of logistics nodes, and the matrix elements represent the connection states between the logistics nodes.

[0027] In some embodiments, calculating node feature information includes: determining the node degree of each logistics node; the node degree represents the number of other logistics nodes directly connected to the logistics node. Calculating the average value of the node degrees of all logistics nodes to obtain the average connection number. Determining the edge degree between the logistics nodes; the edge degree represents the number of edges having connection relationships between the other logistics nodes connected to the logistics node. Calculating the clustering coefficient of the logistics node based on the average connection number and the edge degree. For the target logistics node (the node currently being processed), its clustering coefficient is equal to twice its edge degree divided by the product of the node degree and the node degree minus one, and when the node degree is less than 2, the clustering coefficient of the target logistics node is 0. Calculating the average value of the clustering coefficients of all logistics nodes to obtain the aggregation degree.

[0028] Step 140, designing a scheduling strategy to perform collaborative distribution scheduling on the processed raw materials and finished products based on the logistics feature information and the node feature information. The scheduling strategy may refer to the scheduling strategy for performing collaborative distribution on the processed raw materials and finished products. For example, adjusting the distribution volume of the processed raw materials and products, adjusting the quantity of the processed raw materials and products stored in the warehouse, etc.

[0029] In some embodiments, based on the logistics feature information and the node feature information, a scheduling strategy is designed for the collaborative distribution scheduling of processing raw materials and finished products, including: constructing a scheduling objective function based on the node feature information and the distribution volume. The distribution volume may include the quantity of processing raw materials transported from one node to another node and the quantity of products transported. Based on the logistics feature information, scheduling constraint conditions are constructed. The scheduling objective function is solved to obtain the inventory levels of processing raw materials and finished products at the logistics nodes. The inventory level of processing raw materials may refer to the quantity of processing raw materials stored in the warehouse node; the inventory level of finished products may refer to the quantity of finished products stored in the warehouse node. Through a scheduling optimization algorithm, the scheduling objective function is optimized to obtain an optimized distribution volume. The optimized distribution volume is adjusted based on the random feature information to obtain an adjusted distribution volume. The adjusted distribution volume is substituted into the scheduling objective function to obtain a new scheduling objective function. The operations of solving and optimizing the scheduling objective function are repeated until the change values of the inventory levels of processing raw materials and finished products are respectively less than a preset raw material inventory threshold and a preset finished product inventory threshold, and the final distribution volumes of processing raw materials, finished products, inventory levels of processing raw materials, and inventory levels of finished products are used as the scheduling strategy. The change values of the inventory levels of processing raw materials and finished products refer to the differences obtained by subtracting the inventory levels of processing raw materials and finished products obtained from the previous solution of the scheduling objective function from the inventory levels of processing raw materials and finished products obtained from the subsequent solution of the scheduling objective function. The preset raw material inventory threshold refers to the maximum value of the change amount of the inventory level of processing raw materials after two adjacent optimizations set in advance. The preset finished product inventory threshold refers to the maximum value of the change amount of the inventory level of finished products after two adjacent optimizations set in advance.

[0030] In some embodiments, the expression of the scheduling objective function is: ; ; ; ; where min represents taking the minimum value; Z represents the value of the scheduling objective function; represents the first scheduling objective parameter; represents the second scheduling objective parameter; represents the third scheduling objective parameter; i and j represent different logistics node variables, representing the first logistics node variable and the second logistics node variable respectively; N represents the total number of logistics nodes; represents the influence degree of the aggregation degree on the basic transportation cost; D represents the aggregation degree; and respectively represent the basic transportation cost of processing raw materials and the basic transportation cost of finished products from logistics node i to logistics node j; and respectively represent the distribution volume of processing raw materials and the distribution volume of finished products from logistics node i to logistics node j; represents the influence degree of the average connection quantity on the basic transportation time; represents the average connection quantity; represents the basic transportation time from logistics node i to logistics node j; k represents the warehouse node variable; K represents the total number of warehouse nodes; and respectively represent the unit storage cost of processing raw materials and the unit storage cost of finished products of warehouse node k; and respectively represent the inventory of processing raw materials and the inventory of finished products of warehouse node k.

[0031] The scheduling constraint conditions include: Demand matching constraint: ; ; Among them, and respectively represent the total distribution volume of processing raw materials and the total distribution volume of finished products from logistics node i to all connected logistics nodes j; and respectively represent the demand for processing raw materials and the demand for finished products of node i; and respectively represent the demand fluctuation volume of processing raw materials and the demand fluctuation volume of finished products of node i, and the demand fluctuation volume is related to the trend characteristic information and the cycle characteristic information; represents the conversion coefficient of processing raw materials into finished products; S represents the set of logistics nodes supplying processing raw materials; represents the total amount of processing raw materials transported from logistics node j to logistics node i.

[0032] In some embodiments, the expressions for the demand for processing raw materials and the demand for finished products of node i can be respectively: ; ; Among them, represents the trend component of processing raw materials of logistics node i in period t; represents the cycle component of processing raw materials of node i in period t; represents the trend component of finished products of logistics node i in period t; represents the cycle component of finished products of node i in period t.

[0033] In some embodiments, the expressions for the demand fluctuation volume of processing raw materials and the demand fluctuation volume of finished products of node i can be respectively: ; ; Among them, represents the trend component of the processed raw materials of logistics node i in period t - 1; represents the periodic component of the processed raw materials of node i in period t - 1; represents the trend component of the finished products of logistics node i in period t - 1; represents the periodic component of the finished products of node i in period t - 1.

[0034] Transport capacity constraint: ; Among them, represents the maximum delivery volume that can be carried from logistics node i to logistics node j.

[0035] Inventory capacity constraint: ; ; Among them, and respectively represent the maximum storage capacity of the processed raw materials and the maximum storage capacity of the finished products of warehouse node k.

[0036] Non - negative constraint: .

[0037] Delivery time constraint: ; Among them, represents the start time of delivery; t represents the actual delivery time; represents the end time of delivery.

[0038] In some embodiments, the scheduling optimization algorithm can be a particle swarm optimization algorithm. Through the scheduling algorithm, the scheduling objective function is optimized to obtain the optimized delivery volume, including: Based on the logistics characteristic information, determine the initial delivery volume. The initial delivery volume refers to the delivery volume before optimization. For example, by analyzing the logistics delivery data in the same or similar periods in the past, relevant indicators can be calculated to determine the initial delivery volume. Use the initial delivery volume to calculate the value of the scheduling objective function, providing an initial evaluation basis for the particle swarm optimization. For example, the initial storage volume can be determined through the initial delivery volume, and then the optimized delivery volume can be substituted into the scheduling objective function based on the initial storage volume to obtain the value of the objective function. Substitute the position of each particle into the scheduling objective function to calculate the fitness value; the fitness value represents the quality of the delivery combination; the position of each particle is represented as a set of delivery volume combinations. Adjust the exploration range of the particle in different regions according to the average connection number and aggregation degree of the logistics nodes. When the preset number of iterations is satisfied or the fitness value converges, terminate the iteration, and use the delivery volume combination corresponding to the global optimal position at this time as the final delivery volume. For example, there are 3 warehouses (Warehouse A, Warehouse B, Warehouse C), 2 factories (Factory X, Factory Y), and 2 product demand parties (Demand Party M, Demand Party N). The warehouses store both processing raw materials and products. Factories X and Y need processing raw materials for production, and Demand Parties M and N need products for sales or use. In the particle swarm optimization algorithm, each particle represents a combination of the delivery volume of processing raw materials from each warehouse to the factories and the delivery volume of products to the product demand parties. After multiple rounds of iterative search, the determined delivery volume combination corresponding to the global optimal position is: Processing raw material delivery: Warehouse A delivers 60 tons to Factory X and 50 tons to Factory Y; Warehouse B delivers 40 tons to Factory X and 30 tons to Factory Y; Warehouse C delivers 20 tons to Factory X and 25 tons to Factory Y. Product delivery: Warehouse A delivers 40 pieces to Demand Party M, Warehouse B delivers 30 pieces to Demand Party M, and Warehouse C delivers 50 pieces to Demand Party N.

[0039] In some embodiments, according to the average connection number and aggregation degree of logistics nodes, the exploration range of particles in different regions is adjusted as follows: at logistics nodes where the average connection number is higher than the connection threshold, the search range of particles is narrowed; at logistics nodes where the aggregation degree is higher than the aggregation threshold, particles are guided to search along the local optimal solution at this logistics node. The connection threshold is a reference value for measuring the average connection number of logistics nodes. When the average connection number of a logistics node exceeds this value, it indicates that this node has rich connections and an important position in the logistics network. The aggregation threshold is a reference value for judging the aggregation degree of logistics nodes. If the aggregation degree of a logistics node is higher than this value, it means that the nodes in the area where this node is located are closely connected and the structure is compact. For example, in a regional logistics network, through the analysis of historical data and network structure, the connection threshold is set to 5 and the aggregation threshold is set to 0.6. The average connection number of logistics node A is 6, then this node can be considered as a regional logistics hub, with multiple warehouses and transfer centers connected around it. In the particle swarm optimization search, originally, particles search for the optimal distribution path within a range with a radius of 10 kilometers centered on node A. Due to being higher than the connection threshold, the search range is narrowed, and the radius is adjusted to 5 kilometers to focus on exploring the efficient paths near node A. The aggregation degree of logistics node B is 0.7, indicating that this node is in a dense urban distribution area and the surrounding nodes are closely connected. When the particle search involves this area, particles are guided to search near the local optimal distribution path (such as a route with the shortest distribution time) found at node B, and fine-tuning is performed on the distribution sequence and vehicle scheduling parameters of this path, rather than blindly searching within a large range, so as to quickly optimize the distribution plan for this area.

[0040] In some embodiments, the calculation formulas for updating the individual optimal position and global optimal position of particles through particle swarm optimization are as follows: ; ; ; ; Among them, and respectively represent the velocities of particle at dimension at times ti and ti + 1; represents the original inertia weight of the particle; represents the self-cognition coefficient of the particle; represents the first random number in the interval [0, 1]; represents particle at dimension at its historical best position; and respectively represent particle At the dimension Position; Represents the social cognitive coefficient of the particle; Represents the second random number in the interval [0, 1]; Represents the global best position of all particles at the dimension ; Represents the average number of connections; Represents the connection threshold, with a value greater than 0 and less than the maximum value of the node degree; D represents the aggregation degree; Represents the aggregation threshold, with a value greater than 0 and less than the maximum value of the edge degree; Represents the modified inertia weight of the particle; Represents the modified self-cognitive coefficient of the particle; Represents the maximum value of the node degree; Represents the maximum value of the edge degree.

[0041] In some embodiments, the expression for obtaining the adjusted delivery quantity is: ; ; ; Wherein, And Respectively represent the adjusted delivery quantities of the processed raw materials and the finished products; Represents the adjustment coefficient dynamically adjusted according to the random feature information; And Respectively represent the actual residual standard deviations of the processed raw materials and the finished products; And Respectively represent the predicted residual standard deviations of the processed raw materials and the finished products; And Respectively represent the minimum adjustment coefficient and the maximum adjustment coefficient; Fr, And Respectively represent the random feature information, the maximum value of the random feature information, and the minimum value of the random feature information.

[0042] Figure 2 This is an exemplary module diagram of a scheduling system for collaborative distribution of processed raw materials and finished products provided by the present invention. As Figure 2 Shown, the scheduling system for collaborative distribution of processed raw materials and finished products provided by the present invention includes an acquisition module 210, a decomposition module 220, a calculation module 230, and a scheduling module 240.

[0043] The acquisition module is used to acquire multi-dimensional data in the distribution of processed raw materials and finished products; the multi-dimensional data includes logistics information and node information.

[0044] The decomposition module is used to decompose the logistics information to obtain logistics feature information; the logistics feature information includes trend feature information, periodic feature information, and random feature information.

[0045] The calculation module is used to construct a logistics matrix containing logistics nodes based on the node information and calculate the node feature information; the node feature information includes the average connection number and the aggregation degree.

[0046] The scheduling module is used to design a scheduling strategy to conduct collaborative distribution scheduling for the processing raw materials and finished products based on the logistics feature information and the node feature information.

[0047] The above are only the preferred embodiments of the present invention and are not used to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A scheduling method for coordinated distribution of processing raw materials and finished products, characterized in that, Including: Collecting multi-dimensional data in the raw materials and finished products for distribution and processing; The multi-dimensional data includes logistics information and node information; Decomposing the logistics information to obtain logistics feature information; the logistics feature information includes trend feature information, periodic feature information, and random feature information; Constructing a logistics matrix including logistics nodes based on the node information, and calculating node feature information; the node feature information includes the average connection number and the aggregation degree; Based on the logistics feature information and the node feature information, designing a scheduling strategy to perform collaborative distribution scheduling on the raw materials for processing and the finished products.

2. The scheduling method for coordinated distribution of processing raw materials and finished products according to claim 1, characterized in that, Constructing a logistics matrix including logistics nodes based on the node information, including: Based on the node information, classifying the nodes in the logistics network of the raw materials for distribution and processing and the finished products to obtain multiple types of logistics nodes; Based on the historical transportation routes, determining the connection relationships between the logistics nodes; Based on the logistics nodes and their connection relationships, creating an adjacency matrix, and using the adjacency matrix as the logistics matrix; the adjacency matrix is a square matrix, the order is the total number of nodes of the logistics nodes, and the matrix elements represent the connection status between the logistics nodes.

3. The scheduling method for coordinated distribution of processing raw materials and finished products according to claim 2, characterized in that, Calculating the node feature information, including: Determining the node degree of each logistics node; the node degree represents the number of other logistics nodes directly connected to the logistics node; Calculating the average of the node degrees of all logistics nodes to obtain the average connection number; Determining the edge degree between the logistics nodes; the edge degree represents the number of edges with connection relationships between the other logistics nodes connected to the logistics node; Based on the average connection number and the edge degree, calculating the clustering coefficient of the logistics nodes; Calculating the average of the clustering coefficients of all logistics nodes to obtain the aggregation degree.

4. The scheduling method for coordinated distribution of processing raw materials and finished products according to claim 1, characterized in that, Based on the logistics feature information and the node feature information, designing a scheduling strategy to perform collaborative distribution scheduling on the raw materials for processing and the finished products, including: Constructing a scheduling objective function based on the node feature information and the distribution volume; Constructing scheduling constraint conditions based on the logistics feature information; Solving the scheduling objective function to obtain the inventory levels of the raw materials for processing and the finished products at the logistics nodes; Optimizing the scheduling objective function through a scheduling optimization algorithm to obtain an optimized distribution volume; Adjusting the optimized distribution volume based on the random feature information to obtain an adjusted distribution volume; Substituting the adjusted distribution volume into the scheduling objective function to obtain a new scheduling objective function; Repeating the operations of solving and optimizing the scheduling objective function until the change values of the inventory levels of the raw materials for processing and the finished products are respectively less than a preset raw material inventory threshold and a preset finished product inventory threshold, and using the final distribution volume, the inventory level of the raw materials for processing, and the inventory level of the finished products as the scheduling strategy.

5. The scheduling method for coordinated distribution of processing raw materials and finished products according to claim 4, characterized in that, The expression of the scheduling objective function is: ； ; ； ; Among them, min represents taking the minimum value; Z represents the value of the scheduling objective function; represents the first scheduling objective parameter; represents the second scheduling objective parameter; represents the third scheduling objective parameter; i and j represent different logistics node variables, representing the first logistics node variable and the second logistics node variable respectively; N represents the total number of logistics nodes; represents the influence degree of the aggregation degree on the basic transportation cost; D represents the aggregation degree; and respectively represent the basic transportation cost of processing raw materials and the basic transportation cost of finished products from logistics node i to logistics node j; and respectively represent the distribution volume of processing raw materials and the distribution volume of finished products from logistics node i to logistics node j; represents the influence degree of the average connection number on the basic transportation time; represents the average connection number; represents the basic transportation time from logistics node i to logistics node j; k represents the warehouse node variable; K represents the total number of warehouse nodes; and respectively represent the unit storage cost of processing raw materials and the unit storage cost of finished products of warehouse node k; and respectively represent the inventory of processing raw materials and the inventory of finished products of warehouse node k; The scheduling constraint conditions include: Demand matching constraint: ； ； Among them, and respectively represent the total amount of processed raw material distribution and the total amount of finished product distribution to all connected logistics nodes j starting from logistics node i; and respectively represent the demand for processed raw materials and the demand for finished products at node i; and respectively represent the fluctuation amount of the demand for processed raw materials and the fluctuation amount of the demand for finished products at node i. The demand fluctuation amount is related to the trend characteristic information and the periodic characteristic information; represents the conversion coefficient for processing processed raw materials into finished products; S represents the set of logistics nodes supplying processed raw materials; represents the total amount of processed raw materials transported from logistics node j to logistics node i; Transportation capacity constraint: ； Among them, represents the maximum delivery volume that can be carried from logistics node i to logistics node j; Inventory capacity constraint: ； ； Among them, and respectively represent the maximum storage capacity of processing raw materials and the maximum storage capacity of finished products of the warehouse node k; Non-negativity constraint: ； Distribution time constraint: ； Among them, represents the start time of delivery; t represents the actual delivery time; represents the end time of delivery.

6. The scheduling method for coordinated distribution of processing raw materials and finished products according to claim 4, characterized in that, Optimizing the scheduling objective function through a scheduling optimization algorithm to obtain an optimized distribution volume, including: Determining an initial distribution volume based on the logistics feature information; Calculating the value of the scheduling objective function using the initial distribution volume to provide an initial evaluation basis for particle swarm optimization; Substitute the position of each particle into the scheduling objective function to calculate the fitness value; the fitness value represents the quality of the distribution combination; the position of each particle is represented as a set of distribution quantity combinations. Adjust the exploration range of particles in different regions according to the average connection number and aggregation degree of logistics nodes. When the preset number of iterations is met or the fitness value converges, terminate the iteration, and use the distribution quantity combination corresponding to the global optimal position at this time as the final distribution quantity.

7. The scheduling method for coordinated distribution of processing raw materials and finished products according to claim 6, characterized in that, Adjust the exploration range of particles in different regions according to the average connection number and aggregation degree of logistics nodes as follows: At logistics nodes where the average connection number is higher than the connection threshold, narrow the search range of particles. At logistics nodes where the aggregation degree is higher than the aggregation threshold, guide the particles to search along the local optimal solution at the logistics nodes.

8. The scheduling method for coordinated distribution of processing raw materials and finished products according to claim 6, characterized in that, The calculation formula for updating the individual optimal position and global optimal position of particles through particle swarm optimization is: ; ; ; ; Among them, and represent the velocities of the particle at times \(t_i\) and \(t_{i + 1}\) respectively in dimension ; represents the original inertia weight of the particle; represents the self - awareness coefficient of the particle; represents the first random number in the interval \([0, 1]\); represents the particle in dimension at its historical best position; and represent the positions of the particle at times \(t_i\) and \(t_{i + 1}\) respectively in dimension ; represents the social - awareness coefficient of the particle; represents the second random number in the interval \([0, 1]\); represents the global best position of all particles in dimension ; represents the average number of connections; represents the connection threshold; \(D\) represents the aggregation degree; represents the aggregation threshold; represents the modified inertia weight of the particle; represents the modified self - awareness coefficient of the particle; represents the maximum value of the node degree; represents the maximum value of the edge degree.

9. The scheduling method for coordinated distribution of processing raw materials and finished products according to claim 4, characterized in that, The expression for obtaining the adjusted distribution quantity is: ; ; ; Among them, and respectively represent the adjusted delivery quantities of the processing raw materials and the finished products; represents the adjustment coefficient dynamically adjusted according to the random feature information; and respectively represent the actual residual standard deviations of the processing raw materials and the finished products; and respectively represent the predicted residual standard deviations of the processing raw materials and the finished products; and respectively represent the minimum adjustment coefficient and the maximum adjustment coefficient; Fr, and respectively represent the random feature information, the maximum value of the random feature information, and the minimum value of the random feature information.

10. A scheduling system for the collaborative distribution of processing raw materials and finished products using the scheduling method according to any one of claims 1-9, characterized in that, It includes a collection module, a decomposition module, a calculation module, and a scheduling module. The collection module is used to collect multi-dimensional data in distribution processing raw materials and finished products; the multi-dimensional data includes logistics information and node information. The decomposition module is used to decompose the logistics information to obtain logistics feature information; the logistics feature information includes trend feature information, periodic feature information, and random feature information. The calculation module is used to construct a logistics matrix containing logistics nodes based on node information and calculate node feature information; the node feature information includes the average connection number and aggregation degree. The scheduling module is used to design a scheduling strategy to perform coordinated distribution scheduling on processing raw materials and finished products based on logistics feature information and node feature information.

Images