Field warehouse site selection method in complex dynamic environment

Abstract

The invention belongs to the technical field of military material guarantee business, and particularly relates to a field warehouse site selection method in a complex dynamic environment. The method comprises the following steps: step 1, constructing a technical index evaluation system; 2, optimizing target design; and step 3, using a fast graph clustering algorithm to decide an optimal field warehouse site selection result. The method overcomes the defect that field warehouse site selection decision-making consideration factors are not systematic and are too one-sided, the complexity of a material supply network and the dynamic nature of a field material guarantee network are combined, the reasonability of a site selection scheme of a material supply network warehouse is evaluated from multiple angles, frontier technologies in the fields of artificial intelligence, operational research, intelligent optimization and the like are integrated, and the method has a wide application prospect. Different optimization decision preferences are designed, an intelligent field operation material guarantee network with high flexibility, flexibility, agility and survivability is constructed, rapid generation of a field operation warehouse site selection scheme under the confrontation condition is realized, and the material guarantee intelligence degree is effectively improved.

Description

Technical Field

[0001] The invention belongs to the technical field of military material support business, and in particular relates to a field warehouse site selection method under a complex dynamic environment. Background Art

[0002] Field warehouse site selection refers to the selection of warehouse addresses that meet combat missions in a complex material supply network in order to meet the storage and transportation needs of field materials in the military material support business scenario, taking into account a variety of influencing factors. It is necessary to comprehensively consider the impact of important factors such as safety, anti-destruction, and cost consumption, and select warehouse sites based on the possibility of connecting various transportation modes. Reasonable warehouse site selection can significantly improve the efficiency of field material support and enhance combat capabilities.

[0003] At present, site selection technologies mainly include site selection methods based on the strategy of minimizing transportation cost and site selection methods based on the strategy of maximizing safety.

[0004] In the location selection method based on the minimum transportation guarantee cost strategy, this method focuses on transportation cost factors, including transportation distance factors, transportation mode factors, transportation frequency factors, transportation route planning factors, and cargo loading and unloading cost factors. In the transportation distance factor, the shorter the distance between the field warehouse and the front-line demand point, the less transportation mileage, and the corresponding fuel consumption, vehicle wear and tear, and driver working hours costs are lower. Long-distance transportation not only increases direct transportation costs, but also may increase the risk of cargo damage and delays due to the long distance, thereby incurring additional costs. In the transportation mode factor, the costs of different modes of transportation (such as road transportation, rail transportation, water transportation, and air transportation) vary greatly. The location of the warehouse should consider locations that can make full use of cost-effective transportation methods. For example, freight stations close to railways are suitable for long-distance transportation of large quantities of goods, while those close to major highways are convenient for short-distance highway distribution. In the transportation frequency factor, if the warehouse location is not ideal, it may lead to frequent small-scale transportation, which will increase the organization and scheduling costs of transportation. On the contrary, a suitable location can achieve batch transportation, improve transportation efficiency, and reduce unit transportation costs. In the transportation route planning factor, the road conditions and traffic conditions around the warehouse will affect the planning and selection of distribution routes. Complex transportation networks and road conditions may lead to longer transportation time and higher transportation costs. The road capacity and the stability of the transportation network flow should be considered when selecting a site. Among the factors affecting the cost of cargo handling, the warehouse's loading and unloading facilities and site conditions will affect the efficiency of loading and unloading. If the warehouse location is not convenient for loading and unloading operations, it may lead to longer loading and unloading time, higher labor costs, and greater risk of cargo obstruction.

[0005] In the method of selecting the site for field warehouses based on the highest security strategy, this method focuses on factors such as the risk of enemy reconnaissance and attack, terrain cover, camouflage conditions, fortifications, balance between transportation convenience and confidentiality, coordination with friendly positions, and security of logistics support channels. In the factor of enemy reconnaissance and attack risk, the possibility of the selected location being discovered by enemy reconnaissance means and being attacked by enemy firepower is evaluated, and the enemy's key reconnaissance and attack areas should be avoided. In the factor of terrain cover, a location with natural barriers can provide a certain degree of concealment and protection, reducing the probability of being discovered and attacked by the enemy. In the factor of camouflage conditions, an environment that is easy to camouflage should be selected, such as a background similar to the surrounding terrain and available vegetation, so that the warehouse can be easily identified in appearance. In the factor of fortifications, effective trenches, bunkers, air defense facilities, etc. should be established around to enhance the defense capabilities of the warehouse. In the factor of balance between transportation convenience and confidentiality, it is necessary to ensure that materials can be transported and deployed in a timely manner, but the location cannot be exposed due to the obvious traffic lines. The factors of coordination with friendly positions should be considered in selecting a location that is convenient for mutual support and coordinated defense with friendly forces. The factors of logistics support channel security should be to ensure that the channels for warehouses to provide material replenishment and personnel reinforcement are relatively safe and not easily cut off by the enemy.

[0006] The technical solution of the present invention overcomes the shortcomings of the unsystematic and one-sided considerations in the site selection decision of field warehouses. It combines the complexity of the material supply network and the dynamics of the field material support network, evaluates the rationality of the site selection plan of the material supply network warehouse from multiple angles, integrates the cutting-edge technologies in the fields of artificial intelligence, operations research, intelligent optimization, etc., designs different optimization decision preferences, and constructs an intelligent field material support network with high flexibility, flexibility, agility and anti-destruction, so as to realize the rapid generation of field warehouse site selection plans under confrontation conditions and effectively improve the intelligence level of material support. Summary of the invention

[0007] 1. Technical issues to be resolved

[0008] The technical problem to be solved by the present invention is how to provide a method for selecting a site for a field warehouse in a complex dynamic environment.

[0009] (II) Technical solution

[0010] In order to solve the above technical problems, the present invention provides a field warehouse site selection method in a complex dynamic environment. The method generates storage site selection decision results under different optimization strategies according to combat mission information, enemy firepower information, field warehouse battlefield environment information, and available field warehouse deployment information. The basic idea of ​​the method is to first construct a three-level indicator evaluation system for field warehouse site selection, and then use a fast graph clustering algorithm and a combination of multiple site selection optimization strategies through the input of available field warehouse status information to decide the optimal field warehouse site selection result. The method flow is as follows: Figure 1 As shown;

[0011] The method comprises:

[0012] Step 1: Construction of technical indicator evaluation system;

[0013] Step 2: Optimize target design;

[0014] Step 3: Use a fast graph clustering algorithm to determine the optimal field warehouse location.

[0015] (III) Beneficial effects

[0016] Compared with the prior art, the technical solution of the present invention has the following advantages:

[0017] (1) Strong systematic index system: The analytic hierarchy process can decompose complex problems into multiple levels and factors for systematic and comprehensive analysis. The operation is relatively simple, easy to understand and apply, and does not require a large amount of complex calculations and advanced mathematical knowledge. It can comprehensively consider qualitative and quantitative factors to make decisions more reasonable. The flexibility allows decision makers to adjust and modify the judgment matrix according to actual conditions and personal experience. When dealing with multi-objective and multi-criteria decision-making problems, it can effectively determine the relative importance of each factor. Intuitively, it can be achieved through the relationship between hierarchical relationships and factors. Comprehensiveness: It covers many key factors, including but not limited to security conditions, transportation conditions, natural conditions, costs, etc., to comprehensively evaluate the advantages and disadvantages of site selection and avoid decision-making deviations caused by lack of factors. Scientificity: The selection and weight allocation of each index are based on in-depth research and data analysis, with a solid theoretical and practical foundation, and can accurately reflect the degree of influence on warehouse site selection. It can be flexibly adjusted and optimized according to the characteristics of the business process to meet various specific warehouse site selection needs.

[0018] (2) Highly adaptable algorithm design: This method uses the FGC fast graph clustering algorithm to calculate the field warehouse site selection data. It can process large-scale graph data in a short time, quickly obtain calculation results, and improve calculation efficiency. It has strong adaptability and can be applied to graph structures of various types and sizes, and can provide clustering results more accurately. It is easy to expand to larger data sets and more complex graph structures to meet the growing data processing needs. Easy to understand and explain: The results are intuitive and explainable. Combining the indicator system with the fast graph clustering algorithm has created a new field warehouse site selection decision-making method that breaks through the limitations of traditional methods.

[0019] (3) The logical design of the optimization target has good applicability: for the field warehouse material support business scenario, three optimization logic ideas are sorted out and designed: high comprehensive score, strong reconstruction ability, and low warehouse resource occupancy. It can better solve according to the user's preferences. Use advanced mathematical models and algorithms to accurately predict the effects and cost-effectiveness of different site selection schemes. The logical design of the optimization target has been carefully constructed to accurately capture the key elements and constraints in warehouse site selection, thereby providing a clear and definite direction for the model, and the site selection results are highly in line with actual needs, thereby accurately positioning. The sophisticated logic of the optimization target reduces unnecessary calculations and analysis, greatly improves the operating efficiency of the model, and provides decision makers with feasible site selection plans in a short time, accelerating the decision-making process, saving time and resources.

[0020] The innovation of the present invention is:

[0021] (1) An innovative and comprehensive framework of warehouse site selection indicators has been constructed, covering multiple key dimensions such as security conditions, transportation conditions, natural conditions, and costs, which can accurately evaluate the advantages and disadvantages of site selection.

[0022] (2) A unique warehouse site selection model construction process is proposed, which comprehensively applies mathematical algorithms and data analysis methods, fully considers various complex realistic factors, and improves the accuracy and reliability of site selection.

[0023] (3) A warehouse location decision strategy framework based on an indicator system and fast graph clustering was developed, which achieved efficient site selection analysis and was able to quickly screen out the optimal warehouse location solution. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 This is a flow chart of the field warehouse site selection method in a complex dynamic environment.

[0029] Figure 2 Design a flow chart for the optimization strategy with high comprehensive score.

[0030] Figure 3 Schematic diagram of the process flow for designing a strong optimization strategy for refactoring capabilities.

[0031] Figure 4 Design a flow chart for the optimization strategy to minimize warehouse resource utilization. DETAILED DESCRIPTION

[0032] In order to make the purpose, content, and advantages of the present invention more clear, the specific implementation methods of the present invention are further described in detail below in conjunction with the accompanying drawings and examples.

[0033] In order to solve the above technical problems, the present invention provides a field warehouse site selection method in a complex dynamic environment. The method generates storage site selection decision results under different optimization strategies according to combat mission information, enemy firepower information, field warehouse battlefield environment information, and available field warehouse deployment information. The basic idea of ​​the method is to first construct a three-level indicator evaluation system for field warehouse site selection, and then use a fast graph clustering algorithm and a combination of multiple site selection optimization strategies through the input of available field warehouse status information to decide the optimal field warehouse site selection result. The method flow is as follows: Figure 1 As shown;

[0034] The method comprises:

[0035] Step 1: Construction of technical indicator evaluation system;

[0036] Step 2: Optimize target design;

[0037] Step 3: Use a fast graph clustering algorithm to determine the optimal field warehouse location.

[0038] Among them, in step 1, the technical indicator evaluation system is constructed:

[0039] By constructing a specific field warehouse site selection evaluation index system, the abstract evaluation criteria are quantified; according to the index evaluation results, the field warehouse site selection plan can be improved; the site selection characteristics of the field material supply network warehouse under the rear loading condition and the role of the warehouse in military activities are comprehensively considered to construct a three-level evaluation index system;

[0040] Among them, the first-level indicators are divided into security conditions, transportation conditions, natural conditions, and costs;

[0041] The secondary indicators are divided into warehouse capacity, regional environment, road facilities, surrounding facilities, geological conditions, natural disasters, meteorological conditions, construction cost, and maintenance cost;

[0042] Each secondary indicator corresponds to several third-level indicators.

[0043] Among them, the three-level evaluation index system of field warehouse site selection data is shown in Table 1:

[0044] Table 1 The three-level indicator system structure for field warehouse site selection

[0045]

[0046]

[0047]

[0048]

[0049]

[0050]

[0051]

[0052]

[0053]

[0054] According to the three-level evaluation index system, the AHP hierarchical analysis method is adopted to evaluate the warehouse guarantee efficiency through the index system, providing a basis for the subsequent construction and reconstruction of the material supply network.

[0055] The analytic hierarchy process is a method used for multi-criteria decision analysis. It decomposes complex problems into multiple levels and determines the relative importance of each factor by comparing them two by two, thus providing a basis for decision-making. The construction process of the field warehouse location selection index based on the AHP analytic hierarchy process is as follows:

[0056] (1) Establishing a hierarchical model: Decompose the field warehouse location selection problem into target layer, criterion layer, and plan layer, and construct a hierarchical model;

[0057] (2) Constructing a judgment matrix: For each indicator element at the same level, determine their importance relative to a criterion at the previous level through pairwise comparison and express them with numerical values ​​to form a judgment matrix;

[0058] (3) Hierarchical single sorting: Calculate the eigenvector and maximum eigenvalue of the judgment matrix to obtain the weight ranking of each hierarchical element for a certain criterion in the previous layer;

[0059] (4) Consistency test: Check the consistency of the judgment matrix to determine whether it meets the consistency requirements; if not, the judgment matrix needs to be adjusted;

[0060] (5) Hierarchical total ranking: Based on the results of a single ranking, the weight ranking of each plan for a single goal is calculated to obtain the final decision result;

[0061] The AHP helps decision makers choose among multiple options or determine the importance of each element to the goal. It has the advantages of being simple, intuitive, easy to understand and apply, but it also has some limitations. For example, the construction of the judgment matrix may be affected by subjective factors, and the results of the consistency test may not be accurate enough. Therefore, when using the AHP, it is necessary to conduct reasonable application analysis based on actual conditions.

[0062] Wherein, in step 2, the optimization target design is:

[0063] After the three-level evaluation index system of field warehouse site selection is built, different optimization strategies are designed and combined with the field warehouse site selection business process to build a field warehouse site selection optimization function; and the field warehouse site selection results under different strategy biases are solved by combining fast graph clustering calculation; the optimization strategies include high comprehensive score, strong reconstruction ability, and low warehouse resource occupancy;

[0064] (1) Comprehensive scoring high optimization strategy design:

[0065] 1) Construct evaluation indicators and set key indicator weights;

[0066] 2) Use the AHP hierarchical analysis method to calculate and score the indicators;

[0067] 3) Sort all the scoring results and select the field warehouse site selection plan with the highest score;

[0068] (2) Design of optimization strategy with strong reconstruction capability - graph clustering:

[0069] 1) Sort out the key elements of reconstruction capability, including demand satisfaction rate and network load factors;

[0070] 2) Call the historical data of network structure changes;

[0071] 3) Dynamic link prediction based on similarity;

[0072] 4) Generate link prediction results;

[0073] (3) Design of the optimization strategy for the lowest warehouse resource occupancy rate, which meets the requirements and has the least number of requirements:

[0074] 1) Field warehouse resource occupancy constraints;

[0075] 2) Focus on computing task requirements;

[0076] 3) Analyze the capacity scope of field warehouses;

[0077] 4) Matching computing tasks with field warehouse capabilities;

[0078] 5) Generate a site selection plan for a field warehouse.

[0079] In step 3, a fast graph clustering algorithm is used to determine the optimal field warehouse location:

[0080] The purpose of using the fast graph clustering algorithm is to form subnets in the current network and quickly form multiple capability replacement clusters; as an improved spectral clustering algorithm, the fast graph clustering algorithm selects representative nodes to represent the original data, obtains the approximate feature vector of the data, and realizes the fusion of multiple approximate results through clustering integration to improve the robustness of graph clustering; the fast graph clustering algorithm is used in the following military scenarios to cluster warehouses of the same type; the overall framework of the fast graph clustering algorithm includes the following three main steps:

[0081] (1) Select key nodes and form a similarity matrix W between the key nodes and the original data;

[0082] (2) Construct a bipartite graph and obtain the approximate eigenvectors of the data through singular value decomposition;

[0083] (3) Integrate multiple approximate feature vectors.

[0084] In the fast graph clustering algorithm, the first step is to select key nodes; key nodes refer to the most representative field warehouses at the node position in the field material supply and support network; for large-scale data, due to the huge amount of data when performing spectral clustering, the time complexity of the algorithm is too high when calculating eigenvalues ​​and eigenvectors; therefore, a key node selection method is used; through this method, the data is screened, the most representative key nodes are selected, a new adjacency matrix is ​​formed, and then the eigenvalues ​​and eigenvectors are calculated, thereby reducing the time complexity;

[0085] After selecting the key nodes according to the key node evaluation index, the similarity matrix of the relationship between the key nodes and the original data can be obtained, forming a bipartite graph of the key nodes and the original data; according to the spectral graph theory, the bipartite graph is normalized and segmented, and the singular value decomposition (SVD) is used instead of the eigendecomposition to reduce the time complexity and obtain the approximate eigenvector of the data;

[0086] Since the formation of the matrix has certain uncertainty and will be affected by the initial K-means clustering division, the clustering integration operation is used on the matrix to solve the problem of unstable clustering results. Then, K-means clustering is performed on each row of the obtained fusion matrix to obtain the final clustering result. The clustering calculation method is as follows:

[0087] Use G = (V, E) to represent the graph, where V represents the vertex set and E represents the edge set. Given a data set containing n nodes x i Corresponding to each data point in the vertex set V; spectral clustering construction matrix W = {w ij} ij=1,2,…,n is used to reveal the similarity relationship between data points. The row sum (or column sum) of the matrix W is obtained to obtain the degree matrix D∈R n×n , the Laplacian matrix is ​​obtained by the following definition: L = DW, where W is calculated by the formula:

[0088]

[0089] Among them, σ is the kernel parameter, which is generally set to the variance of the data; the essence of spectral clustering is to convert the clustering problem into a graph segmentation problem, and its purpose is to minimize the following objective function:

[0090]

[0091] Among them, A 1 ,A 2 ,…,A k are multiple disjoint subsets that are divided; however, the above objective function cannot directly obtain the optimal solution, so the Laplacian matrix is ​​used to convert the objective function into:

[0092] min tr ( H T LH ) s H T H=I (3)

[0093] Where H = [h 1 ,h 2 ,…,h k ] is the indicator vector; spectral clustering solves the first k eigenvectors of the Laplacian matrix, and finally the field warehouse location clustering result is obtained through the K-means algorithm.

[0094] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for selecting a field warehouse site in a complex dynamic environment, characterized in that: The method generates warehouse site selection decision results under different optimization strategies according to combat mission information, enemy firepower information, field warehouse battlefield environment information, and available field warehouse deployment information; the basic idea of ​​the method is to first construct a three-level indicator evaluation system for field warehouse site selection, and then use a fast graph clustering algorithm and a combination of multiple site selection optimization strategies through the input of available field warehouse status information to decide the optimal field warehouse site selection result.

2. The method for selecting a site for a field warehouse in a complex dynamic environment as claimed in claim 1, characterized in that: The method comprises: Step 1: Construction of technical indicator evaluation system; Step 2: Optimize target design; Step 3: Use a fast graph clustering algorithm to determine the optimal field warehouse location.

3. The method for selecting a site for a field warehouse in a complex dynamic environment as claimed in claim 2, characterized in that: In step 1, the technical indicator evaluation system is constructed: By constructing a specific field warehouse site selection evaluation index system, the abstract evaluation criteria are quantified; according to the index evaluation results, the field warehouse site selection plan can be improved; the site selection characteristics of the field material supply network warehouse under the rear loading condition and the role of the warehouse in military activities are comprehensively considered to construct a three-level evaluation index system; Among them, the first-level indicators are divided into security conditions, transportation conditions, natural conditions, and costs; The secondary indicators are divided into warehouse capacity, regional environment, road facilities, surrounding facilities, geological conditions, natural disasters, meteorological conditions, construction cost, and maintenance cost; Each secondary indicator corresponds to several tertiary indicators.

4. The method for selecting a site for a field warehouse in a complex dynamic environment as claimed in claim 3 is characterized in that: The three-level evaluation index system for field warehouse site selection data is shown in Table 1: Table 1 The three-level indicator system structure for field warehouse site selection 5. The method for selecting a site for a field warehouse in a complex dynamic environment as claimed in claim 4 is characterized in that: According to the three-level evaluation index system, the AHP hierarchical analysis method is adopted to evaluate the warehouse guarantee efficiency through the index system, providing a basis for the subsequent construction and reconstruction of the material supply network.

6. The method for selecting a site for a field warehouse in a complex dynamic environment as claimed in claim 5, characterized in that: The construction process of field warehouse location index based on AHP is as follows: (1) Establishing a hierarchical model: Decompose the field warehouse location selection problem into target layer, criterion layer, and plan layer, and construct a hierarchical model; (2) Constructing a judgment matrix: For each indicator element at the same level, determine their importance relative to a criterion at the previous level through pairwise comparison and express them with numerical values ​​to form a judgment matrix; (3) Hierarchical single sorting: Calculate the eigenvector and maximum eigenvalue of the judgment matrix to obtain the weight ranking of each hierarchical element for a certain criterion in the previous layer; (4) Consistency test: Check the consistency of the judgment matrix to determine whether it meets the consistency requirements; if not, the judgment matrix needs to be adjusted; (5) Hierarchical total ranking: Based on the results of a single ranking, the weight ranking of each plan for a single objective is calculated to obtain the final decision result.

7. The method for selecting a site for a field warehouse in a complex dynamic environment as claimed in claim 6, characterized in that: In step 2, the optimization target design is: After the three-level evaluation index system of field warehouse site selection is completed, the field warehouse site selection optimization function is constructed by designing different optimization strategies and combining the field warehouse site selection business process; and the field warehouse site selection results under different strategy biases are solved by combining fast graph clustering algorithm; The optimization strategies include high comprehensive score, strong reconstruction capability, and low warehouse resource usage; (1) Comprehensive scoring high optimization strategy design: 1) Construct evaluation indicators and set key indicator weights; 2) Use the AHP hierarchical analysis method to calculate and score the indicators; 3) Sort all the scoring results and select the field warehouse site selection plan with the highest score; (2) Design of optimization strategy with strong reconstruction capability - graph clustering: 1) Sort out the key elements of reconstruction capability, including demand satisfaction rate and network load factors; 2) Call the historical data of network structure changes; 3) Dynamic link prediction based on similarity; 4) Generate link prediction results; (3) Design of optimization strategy for minimum warehouse resource occupancy rate: 1) Field warehouse resource occupancy constraints; 2) Focus on computing task requirements; 3) Analyze the capacity scope of field warehouses; 4) Matching computing tasks with field warehouse capabilities; 5) Generate a site selection plan for a field warehouse.

8. The method for selecting a site for a field warehouse in a complex dynamic environment as claimed in claim 7, characterized in that: In step 3, a fast graph clustering algorithm is used to determine the optimal field warehouse location: The purpose of using the fast graph clustering algorithm is to form subnets in the current network and quickly form multiple capability replacement clusters; as an improved spectral clustering algorithm, the fast graph clustering algorithm selects representative nodes to represent the original data, obtains the approximate feature vector of the data, and realizes the fusion of multiple approximate results through clustering integration to improve the robustness of graph clustering; the fast graph clustering algorithm is used in the following military scenarios to cluster warehouses of the same type; the overall framework of the fast graph clustering algorithm includes the following three main steps: (1) Select key nodes and form a similarity matrix W between the key nodes and the original data; (2) Construct a bipartite graph and obtain the approximate eigenvectors of the data through singular value decomposition; (3) Integrate multiple approximate feature vectors.

9. The method for selecting a site for a field warehouse in a complex dynamic environment as claimed in claim 8, characterized in that: In the fast graph clustering algorithm, the first step is to select key nodes; key nodes refer to the most representative field warehouses at the node position in the field material supply and support network; for large-scale data, due to the huge amount of data when performing spectral clustering, the time complexity of the algorithm is too high when calculating eigenvalues ​​and eigenvectors; therefore, a key node selection method is used; through this method, the data is screened, the most representative key nodes are selected, a new adjacency matrix is ​​formed, and then the eigenvalues ​​and eigenvectors are calculated, thereby reducing the time complexity; After selecting the key nodes according to the key node evaluation index, the similarity matrix of the relationship between the key nodes and the original data can be obtained, forming a bipartite graph of the key nodes and the original data; according to the spectral graph theory, the bipartite graph is normalized and segmented, and singular value decomposition is used instead of eigendecomposition to reduce time complexity, and the approximate eigenvector of the data is obtained; Since the formation of the matrix has certain uncertainty and will be affected by the initial K-means clustering division, the clustering integration operation is used on the matrix to solve the problem of unstable clustering results. Then, K-means clustering is performed on each row of the obtained fusion matrix to obtain the final clustering result. The clustering calculation method is as follows: Use G = (V, E) to represent the graph, where V represents the vertex set and E represents the edge set. Given a data set containing n nodes x i Corresponding to each data point in the vertex set V; spectral clustering construction matrix W = {w ij } ij =1,2,…,n is used to reveal the similarity relationship between data points, and the row sum of matrix W is obtained to obtain the degree matrix D∈R n×n , the Laplacian matrix is ​​obtained by the following definition: L = DW, where W is calculated by the formula: Among them, σ is the kernel parameter, which is generally set to the variance of the data; the essence of spectral clustering is to convert the clustering problem into a graph segmentation problem, and its purpose is to minimize the following objective function: Among them, A1, A2, …, A k are multiple disjoint subsets that are divided; however, the above objective function cannot directly obtain the optimal solution, so the Laplacian matrix is ​​used to convert the objective function into: mintr ( H T LH ) s.t.H T H=I (3) Where H = [h1,h2,…,h k ] is the indicator vector; spectral clustering solves the first k eigenvectors of the Laplacian matrix, and finally the field warehouse location clustering result is obtained through the K-means algorithm.

10. The method for selecting a site for a field warehouse in a complex dynamic environment as claimed in claim 9, characterized in that: The method overcomes the shortcomings of the unsystematic and one-sided considerations in the site selection decision of field warehouses. Combining the complexity of the material supply network and the dynamics of the field material support network, the rationality of the site selection plan of the material supply network warehouse is evaluated from multiple angles. It integrates the cutting-edge technologies in the fields of artificial intelligence, operations research, and intelligent optimization, designs different optimization decision preferences, and constructs an intelligent field material support network with high flexibility, flexibility, agility, and anti-destruction. It realizes the rapid generation of field warehouse site selection plans under confrontation conditions and effectively improves the intelligence level of material support.

Images