Knowledge graph-based intelligent decision support system for intelligent grass and animal husbandry logistics

Abstract

The invention discloses an intelligent grass and animal husbandry logistics intelligent decision support system based on a knowledge graph, and relates to the technical field of logistics management. Through a data driving mode, the risk value of the transmission path is quantified by innovatively combining the reliability index and the collaboration degree value, so that a real-time, efficient and intelligent optimization scheme is provided for a grass and animal husbandry logistics system. The system not only can dynamically update the importance coefficient of the supply chain node, but also can accurately calculate the reliability index of the node by deeply analyzing the historical transportation data of each node, and constructs a cooperative relationship matrix by using the number of common suppliers and the historical cooperation frequency so as to evaluate the cooperation closeness between the nodes. Based on the comprehensive indexes, the system can quantify the risk value of each potential path in real time, and it is ensured that quick response is made in a complex and changeable logistics environment.

Description

Technical Field

[0001] The present invention relates to the field of logistics management technology, and in particular to a knowledge graph-based intelligent decision support system for grassland and animal husbandry logistics. Background Art

[0002] As an important part of the agricultural economy, the grassland and animal husbandry industry's logistics system involves every link from production and processing to sales, including raw material procurement, product transportation, warehousing management, and final distribution to consumers. With the diversification of market demand and the increasing complexity of the supply chain, how to efficiently manage these links has become one of the key challenges for the sustainable development of the grassland and animal husbandry industry. Traditional logistics management systems often rely on static data analysis and empirical judgment, which are inadequate in the face of dynamically changing market environments. For example, existing logistics decision support systems often lack the ability to effectively integrate and analyze real-time data, making it difficult to fully understand the actual operation of each node in the supply chain and their interrelationships. In addition, traditional systems are inefficient when processing complex multi-source heterogeneous data, and are unable to quickly respond to emergencies or make accurate predictions.

[0003] Specifically, the existing logistics system has significant deficiencies in the following aspects: First, in terms of data processing, traditional systems fail to fully utilize modern information technology means, such as big data analysis and artificial intelligence algorithms, resulting in delayed information updates and insufficient decision-making basis; second, in route planning, most systems only perform simple calculations based on basic distance and cost factors, which limits the flexibility and accuracy of route selection; third, in terms of risk management, current methods mostly rely on historical experience and fixed rules, lack dynamic adjustment mechanisms, and cannot effectively deal with the risks brought about by uncertainty; finally, at the collaborative cooperation level, traditional systems lack effective tools to quantify and optimize the cooperative relationship between nodes, resulting in inefficient resource allocation and scheduling.

[0004] These challenges combine to make it difficult for existing logistics systems to meet the demands of efficient management, particularly in the grassland and animal husbandry logistics environment, which demands high precision and rapid response. Therefore, developing an intelligent decision-making support system that integrates multiple advanced technologies and methods to enhance the transparency, flexibility, and responsiveness of the entire logistics network has become a critical issue that needs to be addressed within the industry. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent grassland and animal husbandry logistics intelligent decision support system based on knowledge graphs to solve the problem that traditional systems lack effective tools to quantify and optimize the cooperative relationship between nodes, resulting in low efficiency in resource allocation and scheduling.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a knowledge graph-based intelligent decision support system for grassland and animal husbandry logistics, including the following modules:

[0007] Knowledge graph construction module: used to build a knowledge graph for the grassland and animal husbandry logistics system. The knowledge graph includes supply chain nodes and transmission paths.

[0008] Supply chain node weight determination module: used to determine the initial distance weight and importance coefficient for each supply chain node, and obtain the supply chain node weight based on the initial distance weight and importance coefficient;

[0009] Reliability index determination module: used to determine the comprehensive index, and obtain the reliability index based on the comprehensive index and the supply chain node weight;

[0010] Collaborative relationship matrix construction module: used to construct the collaborative relationship matrix between supply chain nodes and update it to the knowledge graph;

[0011] Transmission path risk quantification module: used to quantify the risk of the transmission path into a risk value based on the reliability index and the synergy relationship matrix;

[0012] Logistics optimization module: used to construct the objective function based on the risk value and calculate the objective function value of each transmission path, and obtain the logistics optimization plan based on the objective function value.

[0013] In another embodiment, the supply chain nodes include production nodes, warehousing nodes, transportation nodes and consumption nodes.

[0014] In another embodiment, the supply chain node weight determination module includes the following submodules;

[0015] Node distance determination submodule: used to determine the node distance between the current supply chain node and other supply chain nodes;

[0016] Node average distance determination submodule: used to determine the node average distance based on the node distances between all supply chain nodes and other supply chain nodes;

[0017] Initial distance weight determination submodule; used to determine the initial distance weight according to the node distance and the average node distance;

[0018] Importance coefficient determination submodule: used to determine the importance coefficient for each supply chain node;

[0019] Supply chain node weight determination submodule: used to obtain the supply chain node weight based on the initial distance weight and importance coefficient.

[0020] In another embodiment, the importance coefficient determination submodule includes the following units:

[0021] Basic parameter determination unit: used to obtain the basic parameters of each supply chain node, including connectivity and centrality;

[0022] Key behavior determination unit: used to obtain historical behavior data of each supply chain node and determine the key behavior for each supply chain node based on the historical behavior data;

[0023] Time series data acquisition unit: used to determine the historical time interval and obtain the time series data of key behavior data within the historical time interval;

[0024] Frequency domain conversion unit: used to convert time series data into spectrum distribution;

[0025] Spectrum extraction unit; used to extract the main frequency component and the amplitude corresponding to the main frequency component from the spectrum distribution;

[0026] Impact factor calculation unit: used to calculate the impact factor according to the main frequency component and the amplitude corresponding to the main frequency component;

[0027] Importance coefficient calculation unit: used to calculate the importance coefficient based on the impact factor, connectivity and centrality.

[0028] In another embodiment, the behavior corresponding to the maximum value of the historical behavior data is used as the key behavior.

[0029] In another embodiment, the reliability index determination module includes the following submodules:

[0030] Historical logistics data retrieval submodule: used to retrieve historical logistics data of each supply chain node;

[0031] Time difference series construction submodule: It is used to count the time difference between the actual arrival time and the planned arrival time of each transportation in the historical logistics data of each supply chain node within the historical time interval, and integrate all the time differences in the historical logistics data of each supply chain node within the historical time interval into a time difference series;

[0032] Statistical analysis submodule: used to calculate the standard deviation and deviation of time difference in time difference series;

[0033] Comprehensive index calculation submodule: used to calculate the comprehensive index based on the standard deviation of the time difference and the deviation of the time difference;

[0034] Reliability index calculation submodule: used to calculate the reliability index based on the comprehensive index and supply chain node weights.

[0035] In another embodiment, the collaborative relationship matrix construction module includes the following submodules:

[0036] Common supplier quantity determination submodule: used to determine the number of common suppliers between the current supply chain node and other supply chain nodes;

[0037] Historical cooperation frequency statistics submodule: used to count the historical cooperation frequency between the current supply chain node and other supply chain nodes;

[0038] Collaboration degree calculation submodule: used to calculate the collaboration degree value based on the number of common suppliers and historical cooperation frequency;

[0039] Update submodule: used to update the collaboration value to the knowledge graph.

[0040] In another embodiment, the reliability index of each supply chain node on each transmission path is extracted, and the average reliability index of the transmission path is calculated; the average degree of coordination of the transmission path is calculated based on the coordination degree value; and the risk value is calculated based on the average reliability index and the average degree of coordination.

[0041] In another embodiment, the objective function value of each transmission path is calculated, the transmission path corresponding to the maximum objective function value or the minimum objective function value is used as the optimal transmission path, and a logistics optimization plan is formulated based on the optimal transmission path.

[0042] The technical effects and advantages of the present invention are as follows:

[0043] This invention uses a data-driven approach, innovatively combining reliability indexes and collaboration metrics to quantify the risk of transmission paths. This provides a real-time, efficient, and intelligent optimization solution for grassland and animal husbandry logistics systems. This system not only dynamically updates the importance coefficients of supply chain nodes but also accurately calculates their reliability index through in-depth analysis of each node's historical transportation data. It also constructs a collaboration matrix using the number of shared suppliers and historical collaboration frequency to assess the closeness of collaboration between nodes. Based on these comprehensive metrics, the system can quantify the risk of each potential path in real time, ensuring rapid response in complex and volatile logistics environments. The system demonstrates significant advantages in resource scheduling: it can flexibly adjust resource allocation based on real-time data, optimizing vehicle scheduling, cargo distribution, and route selection, thereby reducing delays and costs and improving overall operational efficiency. Furthermore, by quickly identifying and responding to potential risks, the system can promptly adjust logistics plans to ensure that goods arrive safely and on time. This real-time and flexible approach significantly enhances the adaptability and competitiveness of grassland and animal husbandry logistics systems, improving customer satisfaction and providing strong support for sustainable business development.

[0044] The present invention dynamically updates node weights by comprehensively considering the importance coefficient of each node, thereby achieving refined management of the entire supply chain network. Specifically, the method first uses time series data of key behaviors to extract the main frequency components and their amplitudes as influencing factors of the time dimension, and combines the connection degree and centrality to calculate the node importance coefficient. Subsequently, for each node, the reliability score is calculated by statistically analyzing the difference between the actual arrival time and the planned arrival time in its historical transportation data, combined with the weight factor, and a collaborative relationship matrix is ​​established to quantify the closeness of cooperation between nodes. On this basis, a risk value is calculated for each potential path. This risk value comprehensively considers the reliability score and average degree of collaboration of all nodes on the path. Finally, an objective function is defined that combines the path risk value and cost. Based on the requirements of the specific application scenario, the method of maximizing or minimizing the objective function is selected to solve the optimal transmission path. This method not only extracts valuable information from multi-source heterogeneous data, but also provides scientific and intelligent decision support through data-driven approach, ensuring that the optimal path that is both safe and economical can be found efficiently and accurately in a complex and changing logistics environment, thereby significantly improving the overall operational efficiency and service quality of the grassland and animal husbandry logistics system. This comprehensive and sophisticated data processing and optimization strategy makes the system more adaptable and intelligent, and expands the scope of application of beneficial effects. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 A structural diagram of the intelligent grassland and animal husbandry logistics intelligent decision support system based on knowledge graph provided in this embodiment. DETAILED DESCRIPTION

[0046] The following will be combined with the accompanying drawings to clarify the technical solutions in the embodiments of the present invention; although it is clear that the embodiments described are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0047] like Figure 1 As shown, this embodiment discloses an intelligent decision support system for grassland and animal husbandry logistics based on knowledge graph, including the following modules:

[0048] Knowledge graph construction module: used to build a knowledge graph for the grassland and animal husbandry logistics system. The knowledge graph includes supply chain nodes and transmission paths.

[0049] Supply chain nodes include production nodes, warehousing nodes, transportation nodes, and consumption nodes. Specifically, the knowledge graph mainly includes supply chain node data, transmission path data, environmental data, etc.

[0050] Supply chain node data: Production node: pasture location, production information, livestock type, feed requirements, etc.

[0051] Warehouse node: warehouse address, storage capacity, inventory, cargo types, temperature control capabilities, etc.

[0052] Transportation nodes: distribution center address, number of vehicles, load capacity, type of transportation tools, etc.

[0053] Consumption node: the location, demand, time window requirements, etc. of retailers or end customers.

[0054] Transmission path data: distance between nodes, road conditions, tolls (such as tolls), traffic restrictions, etc.

[0055] External environmental data: real-time weather data (such as rainfall, snowfall, wind speed, etc.).

[0056] Traffic flow data (e.g., congestion levels, accident reports).

[0057] Geographic information (such as altitude, terrain complexity).

[0058] Preprocess the above data:

[0059] Data cleaning: remove duplicate data, fill missing values, and correct erroneous records.

[0060] Data standardization: Unifying data from different sources into the same format and units. For example, distance units are standardized into kilometers and weight units are standardized into tons.

[0061] Data fusion: Linking heterogeneous data from multiple sources through unique identifiers (such as node IDs). For example, matching yield information from a ranch with inventory management data from a storage node.

[0062] Based on the pre-processed data, a knowledge graph is constructed, with node entities and path entities as graph entities and graph relationships determined. For example:

[0063] Node entity:

[0064] Production node: Attributes include “node ID”, “geographic location”, “livestock type”, “feed requirement”, “production capacity”, etc.

[0065] Warehouse node: Attributes include "node ID", "geographic location", "storage capacity", "inventory quantity", "goods type", "temperature control capability", etc.

[0066] Transport node: Attributes include "node ID", "geographic location", "number of vehicles", "load capacity", "transportation tool type", etc.

[0067] Consumption node: Attributes include "node ID", "geographic location", "demand quantity", "time window requirement", etc.

[0068] Path entity: Attributes include "path ID", "starting node ID", "ending node ID", "distance", "road conditions", "toll", "real-time traffic conditions", etc.

[0069] Relationship between nodes: "Supply relationship": means that the production node provides raw materials to the storage node.

[0070] "Transshipment relationship": indicates the transfer of goods between storage nodes and transportation nodes.

[0071] “Consumption relationship”: indicates that the transportation node delivers goods to the consumption node.

[0072] Path relationship: "Connection relationship": indicates that there is a transmission path between two nodes, including attributes such as distance and road conditions.

[0073] Supply chain node weight determination module: used to determine the initial distance weight and importance coefficient for each supply chain node, and obtain the supply chain node weight based on the initial distance weight and importance coefficient.

[0074] The supply chain node weight determination module includes the following submodules:

[0075] Node distance determination submodule: used to determine the node distance between the current supply chain node and other supply chain nodes.

[0076] Node distance refers to the physical distance between two nodes, which is directly obtained from the transmission path between the two nodes.

[0077] Node average distance determination submodule: used to determine the node average distance based on the node distance between the supply chain node and other supply chain nodes.

[0078] The average node distance means: calling the physical distance between the current node and any other node in the knowledge graph, and then calculating the average distance of all physical distances.

[0079] Initial distance weight determination submodule; used to determine the initial distance weight based on the node distance and the average node distance.

[0080]

[0081] ω ij is the initial distance weight between the i-th supply chain node and the j-th supply chain node, d ij is the node distance between the i-th supply chain node and the j-th supply chain node, log is the logarithmic function, and avg(d) is the average node distance.

[0082] The strength of the connection between nodes is quantified by comprehensively considering the physical distance between nodes and the overall average distance. Specifically, the d ij is the node distance between the i-th supply chain node and the j-th supply chain node, avg(d) is the average node distance, and the average distance is adjusted by the logarithmic function log(avg(d)+1) to ensure that when the distance between nodes is large, the weight will not become too small due to the slow growth of the logarithmic function, thereby maintaining the relative stability of the weight.

[0083] For example, the following is a graphic example of the initial distance weights calculated based on different nodes:

[0084]

[0085]

[0086] From the above table, we can see that even if d ij From 10 to 1000, the weight ω ij The magnitude of the change remains relatively gentle, without significant fluctuations. This demonstrates that the logarithmic function effectively balances the rate of change of weights with increasing distance, ensuring relative weight stability. This design is particularly well-suited for logistics optimization scenarios, such as in grassland and animal husbandry supply chains. When the distances between certain nodes are large, the system can still reasonably allocate weights, avoiding neglecting the importance of these paths due to excessive distances, thereby enabling more scientific route planning and resource scheduling.

[0087] Importance coefficient determination submodule: used to determine the importance coefficient for each supply chain node.

[0088] The importance coefficient determination submodule includes the following units:

[0089] Basic parameter determination unit: used to obtain the basic parameters of each supply chain node, including connectivity and centrality.

[0090] The degree of connection refers to the number of supply chain nodes j that are directly connected to the current supply chain node i, which can be directly obtained from the knowledge graph. The centrality can be obtained by the degree of connection of other supply chain nodes.

[0091]

[0092] C i is the centrality of the i-th supply chain node, D j is the connection degree of the j-th supply chain node, and n is the total number of supply chain nodes in the knowledge graph.

[0093] This formula calculates the centrality of a node by adding up the inverse of the connectivity degrees of all other nodes connected to the node. This means that if a node is connected to many nodes with low connectivity (i.e., these nodes have fewer connections with other nodes), then the centrality of the node will be higher because it plays a key role in these nodes. Conversely, if the node is connected to many nodes with high connectivity (i.e., these nodes have more connections with other nodes), then the centrality of the node will be relatively low because these nodes themselves already have high centrality. This calculation method is particularly suitable for logistics network optimization, helping to identify which nodes are key transportation hubs or bottlenecks, thereby providing an important basis for path planning and resource scheduling.

[0094] Key behavior determination unit: used to obtain historical behavior data of each supply chain node and determine the key behavior for each supply chain node based on the historical behavior data;

[0095] In another embodiment, the behavior corresponding to the maximum value of the historical behavior data is used as the key behavior.

[0096] Specifically, historical behavior data may include:

[0097] Cargo loading and unloading times: Record the number of times cargo is transferred from one node to another.

[0098] Transport frequency: Statistics on the frequency of transport activities in which a node participates within a period of time.

[0099] Inventory turnover rate: calculates the number of times inventory enters and exits per unit time, reflecting the activity level of the storage node.

[0100] Order processing time: records the time required from receiving an order to completing delivery.

[0101] Number of delays: Counts the number of times the goods were not delivered on time.

[0102] Extract historical behavior data from the company's internal ERP (Enterprise Resource Planning) system, WMS (Warehouse Management System), TMS (Transportation Management System) and other databases, and perform cleaning and standardization processing.

[0103] Example: Assume that there is the following historical behavior data about a storage node: Number of cargo loading and unloading times: [10, 15, 20, 18, 22]

[0104] Transport frequency: [3,4,6,7,5]

[0105] Inventory turnover ratio: [2.5, 3.0, 3.5, 4.0, 4.5]

[0106] Order processing time: [24 hours, 22 hours, 20 hours, 19 hours, 18 hours]

[0107] Delay times: [2,1,0,0,1]

[0108] Analyze each behavior data separately to find the maximum value and its corresponding behavior. Number of cargo loading and unloading: The maximum value is 22 times, indicating that cargo loading and unloading activities were the most frequent during this period.

[0109] Transport frequency: The maximum value is 7 times, which means that the node participated in the most transport activities during this period.

[0110] Inventory turnover rate: The maximum value is 4.5 times / cycle, indicating that the inventory turnover rate is the fastest.

[0111] Order processing time: The minimum value is 18 hours, which means the order processing efficiency is the highest.

[0112] Delay times: The minimum value is 0, indicating no delay and the best service quality.

[0113] Based on business needs, the most important behaviors are selected as Key Behaviors. For example, when pursuing efficient logistics management, prioritize behaviors that improve overall operational efficiency. If the goal is to improve cargo handling capacity, "number of cargo loading and unloading times" might be a Key Behavior. If the focus is on reducing delays to improve customer satisfaction, "number of delays" becomes a Key Behavior. When optimizing transportation processes, "shipment frequency" is selected as a Key Behavior.

[0114] Assuming the goal is to identify key behaviors that significantly affect logistics efficiency, we will focus on "number of cargo loading and unloading times", "transportation frequency" and "order processing time". After analysis, we found that:

[0115] The maximum number of cargo loading and unloading times occurs in the fifth cycle, which is 22 times;

[0116] The maximum value of transport frequency also occurs in the fifth cycle, which is 7 times;

[0117] The shortest order processing time (i.e. the optimal case) occurs in the fifth cycle, which is 18 hours.

[0118] Based on the above information, we can conclude that this particular warehousing node exhibits the highest logistics efficiency during the fifth cycle, as both the number of cargo handling and transportation frequency reach their peaks, while the order processing time is also the shortest. Therefore, the performance of "cargo handling," "transportation frequency," and "order processing time" can be considered key behaviors for this node.

[0119] Time series data acquisition unit: used to determine the historical time interval and obtain the time series data of key behavior data within the historical time interval.

[0120] The appropriate historical time interval can be chosen according to business needs, for example:

[0121] Short-term analysis: data from the past 30 days, used to capture recent behavior patterns.

[0122] Medium-term analysis: data from the past 6 months, used to identify seasonal fluctuations.

[0123] Long-term analysis: data from the past 1 year, used to study long-term trends.

[0124] Finally, obtain the time series data of the key behavior data within this historical time interval.

[0125] Frequency domain conversion unit: used to convert time series data into frequency spectrum distribution.

[0126] The frequency domain conversion unit essentially converts the original time series data from the time domain to the frequency domain by converting the time series data into the frequency spectrum distribution. This process is usually completed using the Fast Fourier Transform (FFT) or other similar frequency domain transformation methods. The data of the frequency spectrum distribution represents the energy distribution of the original signal at different frequencies.

[0127] Specifically, the obtained frequency spectrum distribution is a data set containing multiple frequency components and their corresponding amplitude (or power) values. Each frequency component represents a potential periodic pattern, and the corresponding amplitude (or power) reflects the importance or intensity of that frequency in the original time series. For example, in a simplified case, the frequency spectrum distribution is:

[0128] Frequency 1 (every week): amplitude 5

[0129] Frequency 2 (every two weeks): amplitude 10

[0130] Frequency 3 (every four weeks): amplitude 15

[0131] Frequency 4 (every eight weeks): amplitude 3

[0132] Frequency 1 (every week): means that a certain key behavior (such as the number of goods loading and unloading, transportation frequency, etc.) may occur once a week. Amplitude 5 represents the importance or intensity of this periodic behavior.

[0133] Frequency 2 (every two weeks): indicates that the key behavior may occur once every two weeks, and the amplitude 10 is greater than the amplitude 5, indicating that this periodic behavior is more significant.

[0134] Frequency 3 (every four weeks): indicates that the behavior repeats every four weeks, and in this example it has the largest amplitude (15), which may be the most important periodic pattern.

[0135] Frequency 4 (every eight weeks): This indicates that there is a behavior pattern that occurs only once every eight weeks, but its amplitude (3) is relatively small, indicating that its importance is relatively low.

[0136] Therefore, spectral distribution essentially provides a method for observing and analyzing hidden periodicities and trends in time series data. By analyzing spectral distribution, dominant periodic patterns (i.e., primary frequency components) can be identified, which is important for understanding behavioral patterns, predicting future trends, and formulating corresponding strategies. For example, in logistics and supply chain management, this can help identify cyclical fluctuations in inventory levels, transportation demand, and other aspects, thereby optimizing resource allocation and scheduling plans.

[0137] The spectrum extraction unit is used to extract the main frequency component and the amplitude corresponding to the main frequency component from the spectrum distribution.

[0138] The dominant frequency component is the frequency component with a significant amplitude in the spectrum distribution. In the above example, the frequency with the largest amplitude is "every four weeks" (amplitude 15), followed by "every two weeks" (amplitude 10). Therefore, the dominant frequency component and the corresponding amplitude of the dominant frequency component are:

[0139] Main frequency 1: every four weeks, amplitude 15;

[0140] Main frequency 2: every two weeks, amplitude 10;

[0141] Main frequency 3: weekly, amplitude 5;

[0142] Main frequency 4: every eight weeks, amplitude 3.

[0143] Main Frequency 1: Every four weeks, with an amplitude of 15, indicates a cyclical pattern in the time series data that repeats every four weeks. In other words, key behaviors (such as cargo handling times, transportation frequency, etc.) may show significant changes or peaks every four weeks. An amplitude of 15 indicates that this four-week cyclical behavior has the greatest impact on the overall data. This is the most significant cyclical pattern among all frequency components. In a supply chain or logistics system, this may correspond to the following situations:

[0144] Monthly replenishment cycle: For example, some storage nodes may conduct large-scale inventory replenishment every four weeks.

[0145] Production planning cycle: The harvesting cycle of a pasture or the processing cycle of pasture products may be once a month.

[0146] Fluctuations in consumer demand: The demand for certain products may be related to consumer purchasing habits, such as fixed monthly purchasing peaks.

[0147] This periodic pattern is the most important feature and should be the core focus of resource allocation and operational optimization.

[0148] Primary Frequency 2: Every two weeks, amplitude 10, indicates that there is a periodic pattern that repeats every two weeks in the time series data. That is, the key behavior changes regularly every two weeks. The amplitude of 10 indicates that this two-week periodic behavior has a relatively high impact on the overall data, but is lower than the primary frequency 1 (every four weeks). In a supply chain or logistics system, this could correspond to the following scenarios:

[0149] Bi-weekly scheduling plans: For example, some transportation activities may be scheduled every two weeks.

[0150] Inventory adjustments: Some nodes may check and adjust inventory levels every two weeks.

[0151] Promotional activities: There may be a small-scale promotional activity at the consumer end every two weeks, driving transportation and handling needs.

[0152] This periodic pattern is less important than the primary frequency 1, but still needs to be focused on, especially in short-term operational planning.

[0153] Primary Frequency 3: Weekly, amplitude 5, indicates that there is a periodic pattern that repeats every week in the time series data. That is, the key behavior changes regularly every week.

[0154] Amplitude interpretation:

[0155] The amplitude of 5 indicates that this weekly periodic behavior has a relatively low impact on the overall data, but still has some significance.

[0156] In a supply chain or logistics system, this could correspond to the following scenarios:

[0157] Daily operational activities: For example, regular inventory checks, shipments, or receiving goods at fixed times every week.

[0158] Weekday vs. weekend differences: The behavior of some nodes may be influenced by different demand patterns on weekdays and weekends.

[0159] Small replenishments: Some nodes may conduct small replenishments every week to meet short-term demand.

[0160] This periodic pattern, although less important, still needs to be considered in daily operations.

[0161] A major frequency of 4 (every eight weeks) with an amplitude of 3 indicates a recurring pattern in the time series data. This means that the key behavior exhibits a regular pattern every eight weeks. An amplitude of 3 indicates that this eight-week periodicity has minimal impact on the overall data. This is a minor cyclical pattern.

[0162] In a supply chain or logistics system, this might correspond to the following:

[0163] Quarterly maintenance schedules: For example, some equipment or facilities may require comprehensive maintenance every eight weeks.

[0164] Special event driven: For example, certain special holidays or promotions may only occur every two months.

[0165] Long-term trends: This could be a reflection of some long-term seasonal fluctuations or changes in market demand.

[0166] Although smaller in magnitude, this cyclical pattern is still worth paying attention to, especially in medium- to long-term planning.

[0167] Impact factor calculation unit: used to calculate the impact factor according to the main frequency component and the amplitude corresponding to the main frequency component;

[0168]

[0169] T i is the influencing factor of the i-th supply chain node, A i is the main frequency component of the i-th supply chain node, f i is the amplitude of the i-th supply chain node, γ1 is the adjustment coefficient of the main frequency component, γ2 is the adjustment coefficient of the amplitude, γ1+γ2=1, γ1=0.8, γ2=0.2, max(·) is the maximum value operation, max(A) is the maximum value of the main frequency component, and max(f) is the maximum amplitude.

[0170] The maximum main frequency value refers to the maximum main frequency component value selected from the time series data of all historical behavior data, and the maximum amplitude refers to the maximum amplitude value selected from the time series data of all historical behavior data. It refers to the standardization of the main frequency components. It refers to the normalization of amplitude.

[0171] The quantification method not only identifies nodes with key behavior patterns, but also provides a scientific basis for node weight calculation, ensuring more reasonable and effective resource allocation. For example, in a grass-based livestock logistics system, if a node exhibits strong periodicity and high activity intensity, its influence factor is higher, indicating that it plays a more critical role in the supply chain network and should receive more attention and resource investment. Therefore, the influence factor calculated using this formula helps optimize node weights, improve overall logistics efficiency, reduce operating costs, and enhance the responsiveness and flexibility of the supply chain.

[0172] The importance coefficient calculation unit is used to calculate the importance coefficient based on the influence factor, the connection degree, and the centrality.

[0173]

[0174] c i is the importance coefficient of the ith supply chain node, D i is the connection degree of the ith supply chain node, C i is the centrality of the ith supply chain node. max(D) is the maximum connection degree, max(C) is the maximum centrality, γ3 is the adjustment coefficient of the connection degree, and γ4 is the adjustment coefficient of the centrality.

[0175] The maximum connection degree is the maximum connection degree selected from all node connection degrees, and the maximum centrality is the maximum centrality selected from all node centralities. This part is the standardization processing of the connection degree, This part is the standardization processing of the centrality. γ3 + γ4 = 1, γ3 = 0.8, and γ4 = 0.2.

[0176] The formula calculates the influence factor based on time series by considering the connection degree and centrality of the supply chain nodes and performing standardization processing, which has a significant effect on subsequent node weight calculation. Specifically, by normalizing the connection degree and centrality of each node and using adjustment coefficients for weighting, the relative importance of each node in the network can be quantified. This quantification method not only identifies nodes with key connection characteristics and central positions, but also provides a scientific basis for node weight calculation, ensuring more reasonable and effective resource allocation. For example, in a grass-based livestock logistics system, if a node exhibits high connection degree and centrality, its influence factor is higher, indicating that it plays a more critical role in the supply chain network and should receive more attention and resource investment. Therefore, the influence factor calculated using this formula helps optimize node weights, improve overall logistics efficiency, reduce operating costs, and enhance the responsiveness and flexibility of the supply chain.

[0177] Supply chain node weight determination submodule: used to obtain the supply chain node weight according to the initial distance weight and the importance coefficient.

[0178] W j = ω ij × c i , W j is the i th supply chain node weight.

[0179] For example, the following is an illustrative example of the supply chain node weight calculated based on different nodes:

[0180] <![CDATA[ω ij ]]> <![CDATA[T i ]]> <![CDATA[c i ]]> <![CDATA[W j ]]> 2.4 1.32 0.76 1.82 4.16 1.7 0.59 2.45 5 2 0.5 2.5 6.1 2.32 0.43 2.63 9.01 2.48 0.4 3.6 11.71 2.7 0.37 4.33

[0181] As can be seen from the table, ω ij After adjustment by c i , the final supply chain node weight has changed significantly. These differences show that by introducing the importance coefficient for adjustment, the actual importance of each node in the supply chain can be more accurately reflected.

[0182] Reliability index determination module: used to determine the comprehensive index, and obtain the reliability index according to the comprehensive index and the supply chain node weight.

[0183] The reliability index determination module includes the following submodules:

[0184] Historical logistics data retrieval submodule: used to retrieve historical logistics data of each supply chain node.

[0185] Time difference sequence construction submodule: used to calculate the time difference between the actual arrival time and the planned arrival time of each transportation of the historical logistics data of each supply chain node within the historical time interval, and integrate all the time differences of the historical logistics data of each supply chain node within the historical time interval into a time difference sequence.

[0186] The consideration of the effect of time difference is to quantitatively evaluate the on-time delivery performance of each link of the supply chain, and to reveal the potential efficiency problems and bottlenecks. Through analysis of the time difference sequence, enterprises can not only identify which nodes or paths frequently appear delayed, but also further explore the root causes of these delays, such as traffic conditions, weather influence, unreasonable operation process, etc. This provides a scientific basis for optimizing the logistics network, improving the transportation plan, and improving the response speed and service quality of the overall supply chain.

[0187] Statistical analysis submodule: used to calculate the standard deviation of the time difference in the time difference sequence and the deviation of the time difference.

[0188] First, calculate the mean value of the time difference sequence

[0189] r is the total number of transportation times of the current supply chain node, Δt z The time difference between the actual arrival time and the planned arrival time of the zth shipment.

[0190] Standard deviation σ i (Δt z ) reflects the degree of fluctuation of the time difference series, and the formula is:

[0191]

[0192] Deviation skew(Δt z ) can also be called skewness, which is used to characterize the symmetry of the distribution of time difference series. The formula is:

[0193]

[0194] Comprehensive index calculation submodule: used to calculate the comprehensive index based on the standard deviation of the time difference and the deviation of the time difference, and update it to the knowledge graph;

[0195]

[0196] I i is the comprehensive index of the i-th supply chain node, σ i (Δt z ) is the standard deviation of the time difference of the i-th supply chain node, skew(Δt z ) is the deviation of the time difference of the i-th supply chain node.

[0197] Standard deviation reflects the volatility of the time difference series, while skewness describes the symmetry of the time difference distribution. By combining these two indicators, a comprehensive index can more comprehensively assess the temporal stability of supply chain nodes. Specifically, standard deviation measures the dispersion of time differences, while skewness reveals the asymmetry of the time difference distribution, which helps identify potential anomalies. This comprehensive assessment method not only quantifies the reliability of each node but also provides more accurate basic data for the subsequent calculation of the reliability index. Compared with relying on a single indicator, the comprehensive index can more effectively reflect the actual performance of supply chain nodes, thereby playing a greater role in resource allocation, risk management, and optimization decisions.

[0198] Reliability index calculation submodule: used to calculate the reliability index based on the comprehensive index and supply chain node weights.

[0199] R i =I i ×W j ; R i is the reliability index of the i-th supply chain node.

[0200] The composite index takes into account the standard deviation and skewness of time differences, reflecting the node's temporal stability. Supply chain node weights are based on the node's importance and key role in the network. By multiplying these two metrics, the reliability index not only quantifies the node's temporal stability but also considers its importance within the network, providing a scientific basis for subsequent resource allocation, risk management, and optimization decisions. This comprehensive assessment method can more accurately identify critical nodes that have a significant impact on the overall supply chain, helping to improve the efficiency and responsiveness of the entire supply chain.

[0201] Finally, the calculated reliability index is updated to the knowledge graph.

[0202] Collaborative relationship matrix construction module: used to construct the collaborative relationship matrix between supply chain nodes and update it to the knowledge graph.

[0203] The collaborative relationship matrix building module includes the following sub-modules:

[0204] Common supplier quantity determination submodule: used to determine the number of common suppliers between the current supply chain node and other supply chain nodes.

[0205] Historical cooperation frequency statistics submodule: used to count the historical cooperation frequency between the current supply chain node and other supply chain nodes.

[0206] Supplier information is recorded in the company's internal supplier management system (such as the ERP system), including the corresponding relationship between each supply chain node and its direct supplier. Therefore, the number of common suppliers and the historical frequency of cooperation can be directly obtained.

[0207] Collaboration degree calculation submodule: used to calculate the collaboration degree value based on the number of common suppliers and historical cooperation frequency;

[0208]

[0209] M[i][j] is the degree of coordination between the i-th supply chain node and the j-th supply chain node, S ij is the number of common suppliers between the i-th supply chain node and the j-th supply chain node, F ijis the historical frequency of collaboration between the i-th supply chain node and the j-th supply chain node, max(S) is the maximum number of common suppliers, max(F) is the maximum historical frequency of collaboration, α is the adjustment coefficient for the number of common suppliers, and β is the adjustment coefficient for the historical frequency of collaboration. The maximum number of common suppliers refers to the maximum number of common suppliers recorded in the company's internal supplier management system (such as the ERP system), and the maximum historical frequency of collaboration refers to the maximum historical frequency of collaboration recorded in the company's internal supplier management system (such as the ERP system). α + β = 1, α = 0.4, β = 0.6. The calculated collaboration values ​​are integrated into a collaboration matrix.

[0210] The significance of calculating the degree of synergy value by considering the historical frequency of cooperation and the number of common suppliers is that it can comprehensively evaluate the closeness of the cooperative relationship between supply chain nodes. By quantifying these indicators, the degree of synergy value not only reflects the degree of sharing between nodes at the supplier level (the number of common suppliers), but also reflects the frequency of their interaction in historical transactions (historical frequency of cooperation). This comprehensive evaluation method helps to identify which nodes have the closest cooperative relationship, thereby providing an important basis for subsequent risk assessment. The degree of synergy value is of great significance for the calculation of subsequent risk values ​​because it can reveal potential risk points and critical paths in the supply chain network. Specifically, nodes with higher degree of synergy values ​​have a greater impact on the overall supply chain, so more attention needs to be paid during risk assessment to ensure the stability and reliability of these key nodes.

[0211] Update submodule: used to update the collaboration value to the knowledge graph.

[0212] Transmission path risk quantification module: used to quantify the risk of the transmission path into a risk value based on the reliability index and the synergy relationship matrix.

[0213] Extract the reliability index of each supply chain node on each transmission path and calculate the mean reliability index of the transmission path; calculate the mean degree of coordination of the transmission path based on the coordination degree value; calculate the risk value based on the mean reliability index and the mean degree of coordination.

[0214]

[0215] avg(M[i][j]) is the average value of the degree of coordination between the i-th supply chain node and the j-th supply chain node, L k is the length of the kth transmission path.

[0216] Among them, L kThis refers to the sum of the distances to all nodes along a path, normalized by the path length. Using the square root of the path length as the denominator adjusts the scale of the mean degree of coordination to be related to path length. This means that the mean degree of coordination will be smaller over longer paths, reflecting the increased difficulty of coordination over longer paths.

[0217] The synergy values ​​between all pairs of nodes are summed, and the square root of each synergy value is taken. The square root is taken to reduce the impact of extreme values, making the synergy value smoother and more accurately reflecting the overall synergy situation.

[0218] V k =exp(-avg(R i ))+sin(avg(M[i][j]));

[0219] V k is the risk value of the kth transmission path, avg(R i ) is the mean reliability index of the i-th supply chain node, exp is the exponential function, and sin is the sine function.

[0220] This formula can comprehensively evaluate the risk of the transmission path by combining the reliability index and the coordination value. The reliability index reflects the stability of the node, while the coordination value measures the closeness of cooperation between nodes. The exponential function exp(-avg(R i )) emphasizes the importance of reliability. The more reliable a node is, the less impact it has on the risk value. The sine function sin(avg(M[i][j])) introduces periodicity and volatility. The higher the degree of coordination, the closer the value of the sine function is to 1, thus offsetting the impact of the reliability index to a certain extent. This indicates that even if some nodes are not highly reliable, the overall risk can be reduced if the degree of coordination between them is high.

[0221] Logistics optimization module: used to construct the objective function based on the risk value and calculate the objective function value of each transmission path, and obtain the logistics optimization plan based on the objective function value.

[0222] Calculate the objective function value of each transmission path, take the transmission path corresponding to the maximum objective function value or the minimum objective function value as the optimal transmission path, and formulate a logistics optimization plan based on the optimal transmission path.

[0223] In another embodiment, the objective function is:

[0224] P=δ×tanh(V k )+ε×log(Q k +1);

[0225] P is the objective function, Q k is the cost of the k-th transmission path, tanh is the hyperbolic tangent function, and δ and ε are weight coefficients respectively.

[0226] The path cost reflects the total cost of transporting by the chosen path. These costs may include but are not limited to:

[0227] Fuel cost: based on the distance traveled and fuel consumption rate of the vehicle;

[0228] Tolls: Different road sections may have different toll rates;

[0229] Time cost: Considering the timeliness of the goods, quick arrival at the destination may reduce storage and other related costs;

[0230] Vehicle maintenance costs: Long-distance transport may result in higher vehicle wear and tear and maintenance needs.

[0231] In practical applications, the cost of each path can be estimated using historical data, real-time traffic information, and the rate tables of specific transportation companies.

[0232] This formula is obtained by the hyperbolic tangent function tanh(V k The combination of a tangent function and a logarithmic function effectively balances the risk and cost of transmission paths. The hyperbolic tangent function maps the risk value to the interval (-1, 1), compressing the impact of high risk values ​​and amplifying the impact of low risk values, thereby preventing extreme risk values ​​from excessively influencing the overall assessment. The logarithmic function is used to process path costs. By converting cost values ​​into logarithmic form, it can reduce the dominant influence of high-cost paths on the objective function while preserving information about cost differences.

[0233] The objective function is used to calculate the objective function value for each transmission path, and the transmission path corresponding to the maximum or minimum value of the objective function is selected as the optimal transmission path. At the same time, the optimal path should satisfy the constraints.

[0234] In another embodiment, the optimal path required for different application scenarios is not unique. This embodiment preferably adopts different methods to obtain the optimal transmission path and logistics optimization plan for different scenarios: 1. Find the optimal path that minimizes the objective function:

[0235] Application scenario: Cost and risk control are the main focus. In this case, the goal is to achieve both low risk and low cost. This scenario is suitable for cost-sensitive business scenarios that require a certain level of security and reliability. For example:

[0236] Cold chain logistics: For companies transporting perishable goods, ensuring delivery within a specified timeframe to maintain freshness (reducing the risk of delays) while minimizing transportation costs (such as fuel and tolls). Using an objective function that comprehensively considers both risk (such as the possibility of delay or damage) and cost, and finding the path that minimizes this objective function as the optimal cooking path, can effectively balance these two requirements.

[0237] Implementation details: Strengthen the in-transit monitoring system to ensure real-time monitoring of cargo status and timely address issues that may affect freshness.

[0238] High-value goods: When transporting high-value, relatively durable goods, while time isn't the most critical factor, any loss can result in significant financial losses. Therefore, mitigating risks (such as theft and damage) and controlling costs during transportation are paramount.

[0239] Implementation details: Regularly evaluate and update the risk management plan to ensure it can address various potential security threats.

[0240] In the objective function, the weight coefficients δ and ε tend to emphasize cost control, ε = 0.8, while paying moderate attention to risk, δ = 0.2.

[0241] 2. Find the optimal path to maximize the objective function

[0242] When the main goal of an enterprise is to improve service quality and operational efficiency rather than simply save costs, maximizing the objective function value may be a better choice. This situation usually occurs in the following scenarios:

[0243] Express delivery services: For companies offering fast delivery services, speed and service quality often outweigh cost. They prioritize the fastest route to the destination, even if this means higher transportation costs. In this case, the objective function might prioritize minimizing delays (i.e., reducing risk), with weights adjusted appropriately to reflect this.

[0244] Implementation details: Train employees to improve service levels and establish a quick response mechanism to resolve customer issues.

[0245] Emergency supply transportation: When responding to natural disasters or other emergencies, rapid response and timely delivery of supplies are crucial. In these situations, businesses or organizations may be willing to pay higher shipping costs in exchange for faster delivery times and a higher success rate.

[0246] Implementation details: Establish emergency material reserves to shorten response time; strictly monitor all links in the transportation process to ensure the safe arrival of materials.

[0247] In this case, the weight coefficient δ will be set larger, δ = 0.8, to highlight the importance of low risk (high efficiency, punctuality); ε = 0.2.

[0248] Although the two application scenarios (minimizing and maximizing the objective function value) differ in their optimization directions, their constraints are not exactly the same. While some basic constraints may apply to all logistics routing optimization problems, the specific constraints and their priorities may vary depending on different business needs. The following are common constraints in these two application scenarios and their differences:

[0249] Common basic constraints

[0250] Regardless of the application scenario, the following basic constraints are usually required:

[0251] Maximum driving distance: Limit the maximum distance of a single transport to avoid driver fatigue or exceeding the vehicle's endurance.

[0252] Time window: Ensure that the goods can be delivered to the recipient within the specified time period.

[0253] Weight and volume restrictions: Ensure that the transport vehicle is not overloaded or exceeded.

[0254] Path connectivity and legality: Ensures that the selected path is practical, such as avoiding impassable roads or areas.

[0255] Node capacity limit: For example, the storage capacity limit of the storage node to prevent it from exceeding its processing capacity.

[0256] Specific constraints in different application scenarios

[0257] 1. Scenario focusing on cost and risk control

[0258] In this case, the focus is on reducing overall costs and risks, so there are specific constraints or adjustments as follows:

[0259] Strict cost budgeting: Pre-set cost caps must be strictly adhered to, with no overspending allowed. This means more meticulously calculating the costs of each path and strictly restricting or excluding paths that exceed the budget.

[0260] Avoid high-risk areas: For those sections of roads with higher safety risks (such as areas prone to natural disasters and areas with poor public security), these routes should be avoided as much as possible, even if they seem to be better in other aspects (such as short distance and low cost).

[0261] Additional security measures required: To further reduce risks, it may be necessary to add some additional security measures, such as hiring security guards to escort high-value goods, which will also affect the final cost and route selection.

[0262] 2. Scenario focusing on improving efficiency and service quality

[0263] When the primary focus shifts to improving efficiency and service quality, the associated constraints and priorities change:

[0264] Expedited Service Options: Allows for faster, but often more expensive, service options, such as priority on highways or selecting nighttime transit times to reduce the impact of traffic congestion.

[0265] Priority handling mechanism: For urgent materials or high-priority orders, a special priority handling mechanism can be set up to ensure that these goods can receive the fastest processing speed and the shortest transportation time, even if it means sacrificing a certain degree of economy.

[0266] Enhanced flexibility: When faced with emergencies (such as temporary traffic control or severe weather), the system needs to be able to quickly adjust the route in order to respond quickly and find alternative solutions to ensure that the goods arrive at the destination on time.

[0267] While the two application scenarios share some fundamental constraints, their core objectives differ, necessitating specific constraints or adjustments to support their respective optimization strategies during implementation. Scenarios prioritizing cost and risk control tend to be more conservative and cautious, focusing on detail and precision; whereas scenarios prioritizing efficiency and service quality improvement are more flexible and proactive, willing to accept higher costs in exchange for a better service experience. Understanding these differences helps better tailor and apply corresponding optimization strategies.

[0268] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. The intelligent decision support system for grassland and animal husbandry logistics based on knowledge graph is characterized by: Includes the following modules: Knowledge graph construction module: used to build a knowledge graph for the grassland and animal husbandry logistics system. The knowledge graph includes supply chain nodes and transmission paths. Supply chain node weight determination module: used to determine the initial distance weight and importance coefficient for each supply chain node, and obtain the supply chain node weight based on the initial distance weight and importance coefficient; Reliability index determination module: used to determine the comprehensive index, and obtain the reliability index based on the comprehensive index and the supply chain node weight; Collaborative relationship matrix construction module: used to construct the collaborative relationship matrix between supply chain nodes and update it to the knowledge graph; Transmission path risk quantification module: used to quantify the risk of the transmission path into a risk value based on the reliability index and the synergy relationship matrix; Logistics optimization module: used to construct the objective function based on the risk value and calculate the objective function value of each transmission path, and obtain the logistics optimization plan based on the objective function value.

2. The intelligent decision support system for grassland and animal husbandry logistics based on knowledge graph according to claim 1 is characterized by: Supply chain nodes include production nodes, warehousing nodes, transportation nodes and consumption nodes.

3. The knowledge graph-based intelligent decision support system for grassland and animal husbandry logistics according to claim 2 is characterized by: The supply chain node weight determination module includes the following submodules; Node distance determination submodule: used to determine the node distance between the current supply chain node and other supply chain nodes; Node average distance determination submodule: used to determine the node average distance based on the node distances between all supply chain nodes and other supply chain nodes; Initial distance weight determination submodule; used to determine the initial distance weight according to the node distance and the average node distance; Importance coefficient determination submodule: used to determine the importance coefficient for each supply chain node; Supply chain node weight determination submodule: used to obtain the supply chain node weight based on the initial distance weight and importance coefficient.

4. The knowledge graph-based intelligent decision support system for grassland and animal husbandry logistics according to claim 3 is characterized by: The importance coefficient determination submodule includes the following units: Basic parameter determination unit: used to obtain the basic parameters of each supply chain node, including connectivity and centrality; Key behavior determination unit: used to obtain historical behavior data of each supply chain node and determine the key behavior for each supply chain node based on the historical behavior data; Time series data acquisition unit: used to determine the historical time interval and obtain the time series data of key behavior data within the historical time interval; Frequency domain conversion unit: used to convert time series data into spectrum distribution; Spectrum extraction unit; used to extract the main frequency component and the amplitude corresponding to the main frequency component from the spectrum distribution; Impact factor calculation unit: used to calculate the impact factor according to the main frequency component and the amplitude corresponding to the main frequency component; Importance coefficient calculation unit: used to calculate the importance coefficient based on the impact factor, connectivity and centrality.

5. The knowledge graph-based intelligent decision support system for grassland and animal husbandry logistics according to claim 4 is characterized by: The behavior corresponding to the maximum value of historical behavior data is taken as the key behavior.

6. The knowledge graph-based intelligent decision support system for grassland and animal husbandry logistics according to claim 5 is characterized by: The reliability index determination module includes the following submodules: Historical logistics data retrieval submodule: used to retrieve historical logistics data of each supply chain node; Time difference series construction submodule: It is used to count the time difference between the actual arrival time and the planned arrival time of each transportation in the historical logistics data of each supply chain node within the historical time interval, and integrate all the time differences in the historical logistics data of each supply chain node within the historical time interval into a time difference series; Statistical analysis submodule: used to calculate the standard deviation and deviation of time difference in time difference series; Comprehensive index calculation submodule: used to calculate the comprehensive index based on the standard deviation of the time difference and the deviation of the time difference; Reliability index calculation submodule: used to calculate the reliability index based on the comprehensive index and supply chain node weights.

7. The knowledge graph-based intelligent decision support system for grassland and animal husbandry logistics according to claim 6 is characterized by: The collaborative relationship matrix building module includes the following sub-modules: Common supplier quantity determination submodule: used to determine the number of common suppliers between the current supply chain node and other supply chain nodes; Historical cooperation frequency statistics submodule: used to count the historical cooperation frequency between the current supply chain node and other supply chain nodes; Collaboration degree calculation submodule: used to calculate the collaboration degree value based on the number of common suppliers and historical cooperation frequency; Update submodule: used to update the collaboration value to the knowledge graph.

8. The knowledge graph-based intelligent decision support system for grassland and animal husbandry logistics according to claim 7 is characterized by: Extract the reliability index of each supply chain node on each transmission path and calculate the mean reliability index of the transmission path; The mean value of the degree of coordination of the transmission path is calculated based on the degree of coordination value; the risk value is calculated based on the mean value of the reliability index and the mean value of the degree of coordination.

9. The knowledge graph-based intelligent decision support system for grassland and animal husbandry logistics according to claim 8 is characterized by: Calculate the objective function value of each transmission path, take the transmission path corresponding to the maximum objective function value or the minimum objective function value as the optimal transmission path, and formulate a logistics optimization plan based on the optimal transmission path.

Images