Peacetime and wartime combined traffic infrastructure node importance dynamic evaluation method

Abstract

The invention discloses a peacetime and wartime combined traffic infrastructure node importance dynamic evaluation method. The method comprises the following steps: firstly, collecting basic information and traffic network information of traffic infrastructure nodes in a to-be-evaluated region, and constructing a three-level evaluation index system; constructing a single-layer undirected weighted sub-network and an overall multi-layer undirected weighted network under each traffic mode according to the distance between the edges of the nodes; determining the degree, betweenness and compactness of each node; according to a CRITIC method, objective weight values of all indexes in the three-level index system are determined; according to a D-AHP method, determining a subjective weight value; according to a game theory, determining a comprehensive weight after subjective and objective weight combination optimization; according to a TOPSIS method, determining a basic score and a usual importance degree sequence of each node; determining an efficiency score of each single-layer sub-network; determining the comprehensive efficiency of the whole multi-layer network; and determining a wartime importance degree sequence of each node according to the residual overall multi-layer network comprehensive efficiency after the nodes are dynamically deleted.

Description

Technical Field

[0001] The present invention belongs to the technical field of network node value assessment, and specifically is a method for dynamically assessing the importance of transportation infrastructure nodes in both peacetime and wartime. Background Art

[0002] In the era of multipolar international relations, wars are frequent and natural disasters are inevitable. As an important way to deliver combat power, materials, and supplies, the military transportation network is the logistics foundation and combat support in wartime, the basic guarantee of national defense and military strength, and directly affects the military transportation capacity and level of the theater. In wartime, the attack and destruction of certain transportation infrastructure nodes may cause the entire transportation network to be on the verge of paralysis, which in turn affects the overall command and control work. Therefore, the resilience of transportation infrastructure nodes, that is, the ability to resist attacks and the ability to recover after damage, is particularly important. High-value nodes have become the focus of attack and defense for both sides. Accurately identifying key nodes can provide strategic decision-making support and help us seize the initiative in combat.

[0003] The transportation network is a large and complex network consisting of transportation infrastructure nodes and transportation routes formed by different means of transportation running between nodes. The transportation infrastructure mainly includes fixed engineering facilities such as bridges, tunnels, airports, ports, railway stations, bus stations, etc. that provide transportation services for residents' travel and goods transportation.

[0004] The research on multi-layer complex networks is mainly divided into two aspects: one is to analyze the statistical characteristics of the network, including clustering coefficient, eigenvector, degree centrality, and betweenness centrality; the other is to model the complex network, evaluate the value and chain reaction of nodes in the network, and thus identify key nodes. For the specific evaluation methods and means of node importance in the network, the more mature ones are random walk betweenness evaluation and adjacency centrality evaluation. However, these two methods ignore the impact of nodes on the overall network during the evaluation process, and it is difficult to accurately reflect the connection value between different nodes, resulting in deviations in the final evaluation results. Moreover, for networks with a large number of nodes and layers, it is difficult to reflect the small gaps between nodes, and the accuracy of node evaluation is difficult to guarantee. Although the method based on multi-attribute decision-making proposed by Yu Hui et al. takes into account various qualitative and quantitative information in the comprehensive evaluation process, it cannot avoid possible randomness, subjective uncertainty of experts, and ambiguity in cognition, and the evaluation results are not objective enough. The evaluation method based on cosine similarity ignores the weight information between different nodes, resulting in a low degree of conformity between the evaluation model and the actual network. In addition, some scholars have used the node contraction method and the eigenvector method. However, the node contraction method needs to calculate the cohesion after each contraction, and the time complexity is very high; the eigenvector method has a very slow convergence process and may result in a periodic dead loop.

[0005] Most existing research results and directions focus on the evaluation of the value of traffic network nodes in peacetime, but there is less evaluation of node resilience in wartime, lacking comprehensive consideration of different situations in peacetime and wartime. The index systems and methods used by many scholars only reflect the inherent attributes of nodes or the impact on neighboring nodes, and do not adequately reflect their value and importance in large and complex transportation networks. In addition, the research on the identification of key nodes in complex networks is relatively mature in the field of single-layer networks, and the research on the connection between different types of nodes in multi-layer networks has just started, and the existing multi-layer network models cannot accurately reflect the actual traffic network conditions in the real world. Summary of the invention

[0006] The present invention proposes a method for evaluating and ranking the importance of nodes in a multi-layer complex transportation network taking into account the resilience and transportation capacity of wartime transportation infrastructure nodes to solve the above-mentioned existing problems.

[0007] The technical solution to achieve the purpose of the present invention is: a dynamic evaluation method for the importance of transportation infrastructure nodes in peacetime and wartime, comprising the following steps:

[0008] S1. Collect basic information of transportation infrastructure nodes and transportation network information in the area to be evaluated, and build a three-level evaluation index system;

[0009] S2. According to the distance between the edges of each node, construct a single-layer undirected weighted subnetwork and an overall multi-layer undirected weighted network for each mode of transportation;

[0010] S3, using D-number analytic hierarchy process to determine the subjective weight value;

[0011] S4, determine the degree, betweenness and closeness of each node as evaluation indicators;

[0012] S5. Use the objective weighting method to determine the objective weight value of each indicator in the three-level indicator system;

[0013] S6. According to game theory, determine the comprehensive weight after optimizing the combination of subjective and objective weights;

[0014] S7, use TOPSIS method to determine the basic score and usual importance ranking of each node;

[0015] S8. Determine the performance score of each single-layer sub-network according to the basic score of each node;

[0016] S9, determining the comprehensive performance of the overall multi-layer network according to the performance scores of each single-layer sub-network;

[0017] S10. Determine the wartime importance ranking of each node based on the overall multi-layer network comprehensive performance remaining after the nodes are dynamically deleted.

[0018] Preferably, in step S3, the specific steps of using the D-number hierarchical analysis method to determine the subjective weight value are:

[0019] S31. Compare the scores of each indicator by experts in combination with the status of each node and construct the preference relationship matrix R D ;

[0020] S32. Use the fusion formula of D number theory to get the real number matrix R c ;

[0021] Let D = {(b1,v1),(b2,ν2),…,(b n ,v n )}, then the fusion form is:

[0022] I(D) is the integration of D numbers, b i is the value of the subset B of the set Ω, and its corresponding special expression of D number is

[0023] v i , where v i >0, and

[0024] Therefore, based on S31, R C =I(R D );

[0025] S33. Construct matrix R C The probability matrix R P ;

[0026] Experts construct the preference basis of the evaluation object into a pair comparison matrix;

[0027]

[0028] G j >G i Represents G j G i is more important

[0029] S34, R P Perform triangular processing to obtain a triangular probability matrix

[0030] S35. Calculate R D The inconsistency coefficient of

[0031]

[0032] ID is the inconsistency coefficient, For the matrix The elements in , n represents the number of indicators for pairwise comparison;

[0033] S36. Obtain triangular matrix probability from the sorting results

[0034] S37. According to the triangular matrix probability Solve systems of equations;

[0035] set up

[0036] Construct the system of equations:

[0037] ω i represents the i-th weight, ω i ≥0, λ represents the credibility of the assessment, and the subjective weight values ​​of the indicators are calculated by solving the equation group.

[0038] Preferably, the calculation steps of the degree, betweenness and closeness of each node in step S4 are:

[0039] S41. According to the overall multi-layer undirected weighted network constructed in S2, the calculation formula of the node degree is:

[0040]

[0041] If node i is connected to node j, then δ ij =1, otherwise δ ij =0, M is the number of nodes in the overall multi-layer undirected weighted network;

[0042] S42. According to the overall multi-layer undirected weighted network constructed in S2, the calculation formula of the node betweenness is

[0043]

[0044] Among them, P jk is the number of shortest paths between node j and node k; P jk (i) represents the number of shortest paths between node j and node k passing through node i;

[0045] S43. According to the overall multi-layer undirected weighted network constructed in S2, the calculation formula of the node density is:

[0046]

[0047] Where d(i,j) represents the shortest path distance between node i and node j;

[0048] S44. The degree I1(i), betweenness I2(i) and closeness I3(i) of the node calculated in S41, S42 and S43 will be used as evaluation indicators in the evaluation indicator system and applied in the subsequent step S5.

[0049] Preferably, in step S5, the specific method of using the objective weighting method to determine the objective weight value of each indicator in the three-level indicator system is:

[0050] S51. In the three-level indicator system, there are m samples and n evaluation indicators, and the indicator matrix is ​​established:

[0051]

[0052] m is the number of samples, n is the number of evaluation indicators, x ij Take the value of the jth indicator of the i-th sample;

[0053] S52. Dimensionlessize the jth indicator in the i-th sample. For indicators whose values ​​are larger, the better, i.e., benefit-type indicators, we have:

[0054]

[0055] For indicators with smaller values, the better, that is, cost-type indicators, there are:

[0056]

[0057] In the calculation formulas of benefit-based indicators and cost-based indicators, max j (x ij ) is the maximum value of the jth index in the i-th row of samples; min j (x ij ) is the minimum value of the jth indicator in the i-th row of samples;

[0058] From this, the dimensionless evaluation matrix X′ can be calculated:

[0059]

[0060] S53. Calculate the mean of the index and standard deviation s j :

[0061]

[0062]

[0063] S54. Calculate the correlation coefficient. The specific formula is:

[0064]

[0065] ρij is the correlation coefficient between the i-th indicator and the j-th indicator. The closer it is to 1, the greater the correlation. i, j = 1, 2, ..., n, cov (y i y j ) is the covariance of the i-th index and the j-th index of the evaluation matrix X′, S i and S j are the standard deviations of the i-th and j-th indicators respectively;

[0066] S55. Calculate the amount of information. The specific formula is:

[0067]

[0068] E j is the amount of information of indicator j, The objective weight of indicator j is determined based on the difference between the jth indicator and other factors. The calculation formula is:

[0069]

[0070] Preferably, according to game theory, the specific method for determining the comprehensive weight after combined optimization of the subjective weighting method and the objective weighting method based on the D-number hierarchical analysis method in the present invention is:

[0071] S61. Assuming that there are L methods for weighting, there are L sets of weight vector engineering ωi={ω l1 ,ω l2 ,……,ω ln}, where l = 1, 2, ..., L; any combination of L weight combinations is where α l is the linear combination coefficient, α l >0,

[0072] S62. Based on the basic idea of ​​game theory combination weighting, the optimization strategy model is as follows:

[0073]

[0074] S63, transform the optimal solution formula of S62 into the following set of equations by optimizing the first-order derivative condition, and calculate the combination coefficient set {α1, α2, ..., α L}

[0075]

[0076] S64, calculating the combined weight;

[0077]

[0078] The comprehensive weight vector is

[0079]

[0080] Compared with the prior art, the present invention has the following significant advantages: 1. The multi-layer network construction model of the present invention takes into account the complexity of the actual transportation network, especially the characteristics of military transportation in the special geographical location of the Sino-Indian border. The network index evaluation system adopted fully considers the local attributes, global attributes and network connectivity of the nodes, and adopts a subjective and objective combination of index weighting methods to avoid single randomness and ambiguity, making the subsequent value evaluation results more accurate. 2. The node value evaluation and importance ranking method of the present invention integrates the static topological dimension and multiple rounds of dynamic evaluation based on the node deletion method, which improves the accuracy of the calculation results, reflects the changes in the nodes of the transportation infrastructure under the real combat scenario, and has a good reference for our strategic decision-making, helping to determine the best attack target for the enemy network, while also facilitating the defense construction and key management of our military network nodes.

[0081] The present invention is described in further detail below in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0082] Figure 1 This is the process of the comprehensive weighting method for the node indicator system established by the present invention.

[0083] Figure 2 It is a flow chart of calculating the importance of network nodes according to the present invention.

[0084] Figure 3 It is a flow chart of calculating the association degree between network nodes according to the present invention.

[0085] Figure 4 It is a weighted multi-layer complex topology network model constructed by the present invention.

[0086] Figure 5 It is a multi-layer complex network model of a traffic hub constructed by the present invention.

[0087] Figure 6 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0088] In order to make the technical principle of the present invention more clearly understood, the embodiments of the present invention are specifically and in detail described below with reference to the accompanying drawings.

[0089] The present invention is a node importance dynamic evaluation method that combines peacetime and wartime. The entire complex multi-layer transportation system network can be divided into a highway network, a railway network, and an aviation network. Based on the constructed complex network, the D-number hierarchical analysis method and the CRITIC method are used to assign subjective and objective weight values ​​to the node indicators respectively, and the TOPSIS method is used to determine the relative proximity of each node, and the fusion algorithm is used to determine the correlation between nodes. Figure 6 As shown, the specific steps are as follows:

[0090] S1. Collect basic information of transportation infrastructure nodes and transportation network information in the area to be evaluated, and build a three-level evaluation index system;

[0091] Specifically, transportation infrastructure such as bridges, tunnels, airports, ports, and stations are selected as nodes. The transportation network is composed of highway networks, railway networks, aviation and water transport networks. The transportation network is a multi-layer complex network. The internal nodes of each layer of sub-network are connected by edges, and the inter-layer connection relationships between different sub-networks are set through transfer nodes.

[0092] The three-level evaluation index system constructed is as follows: the first-level indicators are the types of transportation infrastructure; the second-level indicators are qualitative indicators such as node resilience or traffic capacity, network connectivity, etc.; the third-level indicators are related to the second-level indicators and quantifiable indicators such as the thickness of the protection layer, the number of lanes, and the degree of the node.

[0093] S2. According to the distance between the edges of each node, construct a single-layer undirected weighted subnetwork and an overall multi-layer undirected weighted network for each mode of transportation;

[0094] Furthermore, the single-layer undirected weighted sub-network is a road network, an aviation network and a railway network.

[0095] S3, determine the subjective weight value according to the D-AHP, which is the AHP improved by the D-AHP theory;

[0096] Assume that there exists a finite non-empty set Ω, which is related to the mapping D: D:Ω→[0,1] and satisfies Where B is a subset of Ω, is an empty set. It means that the number D contains complete information. If it is not equal to 1, it means that the information contained is incomplete.

[0097] Assume that there are samples to be evaluated G = {G1, G2, ..., G n}, and there is a matrix R that exists in the form of G×G, then there are related properties: μ R : G×G→[0, 1], where μ R is the fuzzy preference relation of matrix R. It can be expressed as

[0098]

[0099] This matrix satisfies:

[0100] 1)r ij ≥0;

[0101] 2)

[0102] 3)

[0103] Matrix element r ij Indicates that the evaluation experts believe that G j Relative to G i The importance of , its value range and its meaning are as follows:

[0104] r ij =μ R (G i ,G j ),

[0105] As r ij =0, then G j Obviously better than G i Important; if 0<r ij <0.5, then G j To some extent, G i important; such as r ij =0.5, then G j With G i No difference;

[0106] Since there may be uncertainty when experts make judgments, the scope of application of the preference relationship should be expanded through the D matrix.

[0107] In a further embodiment, the following steps are specifically included:

[0108] S31, compare the experts' scores on each indicator in pairs according to the status of each node, and construct the preference relationship matrix RD;

[0109] 1) Suppose there exists a set G = {G1, G2, ..., G n}, the D-number fuzzy preference relationship is: R D :G×G→D

[0110] The matrix form is:

[0111]

[0112] The matrix satisfies the following conditions:

[0113]

[0114]

[0115]

[0116]

[0117] Score the importance of solutions i and j for k experts, For support,

[0118] 2) In 1),

[0119] S32, using the fusion formula of D number theory, obtain the real number matrix Rc;

[0120] Let D={(b1,ν1),(b2,ν2),…,(b n ,ν n )}, then the fusion form is: I(D) is the integration of D numbers, b i is the value of the subset B of the set Ω, and its corresponding special expression for the D number is ν i , where ν i >0, and Therefore, based on S51, R C =I(R D )

[0121] S33. Construct matrix R C The probability matrix R P ;

[0122] The fuzzy preference relationship is represented by the symbol “>” i Comparison j The importance of is greater, and the preference of experts on the evaluation object is constructed into a pairwise comparison matrix;

[0123]

[0124] S34, R P Perform triangular processing to obtain a triangular probability matrix

[0125] S35. Calculate R D The coefficient of inconsistency;

[0126]

[0127] ID is the inconsistency coefficient, For the matrix The elements in , n represents the number of indicators for pairwise comparison;

[0128] Experts evaluate whether the coefficients are within the acceptable range, and then compare the matrix R P Sort the indicators.

[0129] S36. Obtain triangular matrix probability from the sorting results

[0130] S37, according to Solve systems of equations;

[0131] set up

[0132] Construct the system of equations:

[0133] ω i represents the i-th weight, ω i ≥0, λ represents the credibility of the assessment, and its value is related to the expert experience and the level of understanding of the problem background. Since the experts are experienced, λ=3 can be taken, and then the equation group is solved to calculate the indicator weight.

[0134] S4, determine the degree, betweenness and closeness of each node as evaluation indicators;

[0135] Specifically, according to the overall multi-layer undirected weighted network constructed in step S2, the degree of the node is calculated, and the specific calculation formula is:

[0136]

[0137] If node i is connected to node j, then δ ij =1, otherwise δ ij =0, M is the number of nodes in the overall multi-layer undirected weighted network;

[0138] According to the overall multi-layer undirected weighted network constructed in S2, the betweenness of the nodes is calculated. The specific calculation formula is:

[0139]

[0140] Among them, p jk is the number of shortest paths between node j and node k; p jk (i) represents the number of shortest paths between node j and node k passing through node i, and M is the number of nodes in the overall multi-layer undirected weighted network;

[0141] According to the overall multi-layer undirected weighted network constructed in S2, the density of the nodes is calculated. The specific calculation formula is:

[0142]

[0143] Where d(i,j) represents the shortest path distance between node i and node j.

[0144] S5. According to the above step S4, the node degree I1(i), betweenness I2(i) and closeness I3(i) will be used as evaluation indicators in the evaluation index system in this step and the subsequent steps. According to the objective weighting method (CRITIC), determine the objective weight value of each indicator in the three-level indicator system;

[0145] S51. In the three-level indicator system, there are m samples and n evaluation indicators, and the indicator matrix is ​​established:

[0146]

[0147] m is the number of samples, n is the number of evaluation indicators, x ij Take the value of the jth indicator of the i-th sample;

[0148] S52. Dimensionlessize the jth index in the i-th sample. For the index with the larger value, the better, we have:

[0149]

[0150] For indicators where the smaller the value, the better, there are:

[0151]

[0152] max j (x ij ) is the maximum value of the jth index in the i-th row of samples; min j (x ij ) is the minimum value of the jth indicator in the i-th row of samples;

[0153] From this, the dimensionless evaluation matrix X′ can be calculated:

[0154]

[0155] S53. Calculate the mean of the index and standard deviation s j :

[0156]

[0157]

[0158] S54. Calculate the correlation coefficient. The specific formula is:

[0159]

[0160] ρ ijis the correlation coefficient between the i-th indicator and the j-th indicator. The closer it is to 1, the greater the correlation. i,j=1,2,…,n,coυ(y i y j ) is the covariance of the i-th index and the j-th index of the evaluation matrix X′, S i and S j are the standard deviations of the i-th and j-th indicators respectively;

[0161] S55. Calculate the amount of information. The specific formula is:

[0162]

[0163] E j is the amount of information of indicator j, is the difference between the jth index and other factors,

[0164] Determine the objective weight of indicator j, and the calculation formula is:

[0165]

[0166] S6. According to game theory, determine the comprehensive weight after optimizing the combination of subjective and objective weights;

[0167] S61, determine the weight set combination; assuming that there are L methods for weighting, there are L sets of weight vector engineering ωi={ω l1 ,ω l2 ,……,ω ln}, where l = 1, 2, ..., L; any combination of L weight combinations is where α l is the linear combination coefficient, α l >0,

[0168] S62. Optimization strategy model: Based on the basic idea of ​​game theory combination weighting, the optimization strategy model is:

[0169]

[0170] S63. Find the optimal combination coefficient α l ; The optimal first-order derivative condition can be transformed into the following set of equations, and the combination coefficient set {α1, α2, …, α L}

[0171]

[0172] S64, calculating the combined weight;

[0173]

[0174] The comprehensive weight vector is

[0175]

[0176] S7. According to the TOPSIS method, determine the basic score and usual importance ranking of each node;

[0177] Suppose there are n nodes and m indicators in the network, and the jth indicator value of the i-th node is t ij , and arrange them into node index T.

[0178] S71. To enhance the comparability of the indicators, a normalized matrix T′ is first generated;

[0179] The element t in T′ ij 'for:

[0180]

[0181] S72. Calculate entropy value e j ;

[0182]

[0183] Then the entropy weight ω j for

[0184]

[0185] S73. The positive and negative ideal points in the definable vector space are:

[0186] and

[0187] Then the evaluation value of the jth ideal point is:

[0188]

[0189]

[0190] in

[0191] S74. Calculate the distance between the i-th node and the positive and negative ideal points:

[0192]

[0193]

[0194] S75. Further, the basic score of each node is:

[0195]

[0196] in, and are the distances between the i-th node and the positive and negative ideal points in the TOPSIS algorithm; BS i is the relative closeness between the ith evaluation object and the maximum value, that is, the basic score of the node.

[0197] By sorting the basic scores of each node, we can get the node importance evaluation results under normal conditions.

[0198] S8. Determine the performance score of each single-layer sub-network. The specific steps are as follows:

[0199] S81. Hyperdegree refers to the number of connections between a node and other subnet nodes; let the hyperdegree of the i-th node be Hyperdegree(i), and the efficiency score of node i be FS i ,but

[0200] FS i =Hyperdegree(i)×BS i

[0201] In the formula, BS i is the basic score of each node.

[0202] S82. The performance score of the k-th single-layer undirected weighted subnetwork is:

[0203]

[0204] S9. Determine the overall performance of the multi-layer network;

[0205] Furthermore, the comprehensive performance calculation steps of the whole multi-layer network in step S9 are:

[0206] S91, set G ab is the correlation between networks a and b. The number of edges in the overall multi-layer undirected weighted network is r. Traversing these edges, when the i-th edge connects the single-layer subnetworks a and b, it is recorded as Otherwise, record but Since the constructed network is an undirected network, G ab =G ba ;

[0207] S92, set J ab is the closeness of the two single-layer sub-networks a and b in the overall multi-layer undirected weighted network. Assume that the number of edges connecting the two sub-networks is r 1 The length of the i-th side is q i ,but Since the constructed network is an undirected network, J ab =J ba

[0208] S93, the overall comprehensive performance CE of the multi-layer undirected weighted network is

[0209] CE=FS C (J CM G CM +J CS G CS )+FS M (J MC G MC +J MS G MS )+FS S (J SC G SC +J SM G SM )

[0210] Among them, the first item is the comprehensive performance of the single-layer sub-network C in the overall multi-layer undirected weighted network: FS C represents the functional score within the sub-network, J CM G CM represents the connection score between the sub-network and the single-layer sub-network M, J CS G CS represents the connection score between the subnetwork and the single-layer subnetwork S; the second item is the comprehensive performance of the single-layer subnetwork M in the overall multi-layer undirected weighted network: FS M represents the functional score within the sub-network, J MC G MC represents the connection score between the sub-network and the single-layer sub-network C, J MS G MS represents the connection score between the sub-network and the single-layer sub-network S; the third item is the comprehensive performance of the single-layer sub-network S in the overall multi-layer undirected weighted network: FS S represents the functional score within the sub-network, J SC G SC represents the connection score between the sub-network and the single-layer sub-network C, J SM G SM Represents the connection score between this sub-network and the single-layer sub-network M.

[0211] S10. After obtaining the comprehensive network performance, the nodes to be analyzed and their edges are removed from the original network. According to the overall multi-layer network comprehensive performance remaining after the nodes are dynamically deleted, the wartime importance ranking of each node is determined, and the value of the node is measured by the degree of damage to the network function.

[0212] Furthermore, the steps for calculating the comprehensive performance of the entire multi-layer network remaining after the node is dynamically deleted in step S10 are:

[0213] S101, assuming that the total number of nodes is n, delete node i and its edges, and recalculate the comprehensive efficiency CE′ of the remaining overall multi-layer undirected weighted network, and i=(1,n);

[0214] S102. After each node and its edges are deleted, the remaining comprehensive performance CE′ of the overall multi-layer undirected weighted network is compared with the comprehensive performance CE of the original network, so as to obtain the wartime node importance evaluation result.

[0215] Let the system value of the node be SV (System value),

[0216]

[0217] The larger the SV, the higher the value of the node to be analyzed. By sorting the SV values ​​from large to small, the priority order of attacking military transportation super network targets based on node value can be generated.

[0218] The present invention constructs a multi-layer complex network consisting of road network, railway network, aviation and water transport network, adopts a scientific and reasonable transportation network evaluation index system and method, and identifies high-importance nodes and conducts multiple rounds of value evaluation.

Claims

1. A dynamic evaluation method for the importance of transportation infrastructure nodes in peacetime and wartime, characterized in that: The following steps are involved: S1. Collect basic information of transportation infrastructure nodes and transportation network information in the area to be evaluated, and build a three-level evaluation index system; S2. According to the distance between the edges of each node, construct a single-layer undirected weighted subnetwork and an overall multi-layer undirected weighted network for each mode of transportation; S3, using D-number analytic hierarchy process to determine the subjective weight value; S4, determine the degree, betweenness and closeness of each node as evaluation indicators; S5. Use the objective weighting method to determine the objective weight value of each indicator in the three-level indicator system; S6. According to game theory, determine the comprehensive weight after optimizing the combination of subjective and objective weights; S7, use TOPSIS method to determine the basic score and usual importance ranking of each node; S8. Determine the performance score of each single-layer sub-network according to the basic score of each node; S9, determining the comprehensive performance of the overall multi-layer network according to the performance scores of each single-layer sub-network; S10. Determine the wartime importance ranking of each node based on the overall multi-layer network comprehensive performance remaining after the nodes are dynamically deleted.

2. According to the dynamic evaluation method of traffic infrastructure node importance in peacetime and wartime combined according to claim 1, it is characterized in that: In step S3, the specific steps of using the D-number hierarchical analysis method to determine the subjective weight value are: S31. Compare the scores of each indicator by experts in combination with the status of each node and construct the preference relationship matrix R D ; S32. Use the fusion formula of D number theory to get the real number matrix R c ; Let D={(b1,ν1),(b2,ν2),…,(b n ,ν n )}, then the fusion form is: I(D) is the integration of D numbers, b i is the value of the subset B of the set Ω, and its corresponding special expression for the D number is ν i , where ν i >0, and Therefore, based on S31, R C =I(R D ); S33. Construct matrix R C The probability matrix R P ; Experts construct the preference basis of the evaluation object into a pair comparison matrix; Represents G j Comparison i The importance of S34, R P Perform triangular processing to obtain a triangular probability matrix S35. Calculate R D The inconsistency coefficient of ID is the inconsistency coefficient, For the matrix The elements in , n represents the number of indicators for pairwise comparison; S36. Obtain triangular matrix probability from the sorting results S37, according to the triangular matrix probability Solve systems of equations; set up Construct the system of equations: ω i represents the i-th weight, ω i ≥0, λ represents the credibility of the assessment, and the subjective weight values ​​of the indicators are calculated by solving the equation group.

3. The dynamic evaluation method for the importance of transportation infrastructure nodes combined with peacetime and wartime according to claim 1 is characterized in that: Step S4 The calculation steps of the degree, betweenness and closeness of each node are: S41. According to the overall multi-layer undirected weighted network constructed in S2, the calculation formula of the node degree is: If node i is connected to node j, then δ ij =1, otherwise δ ij =0, M is the number of nodes in the overall multi-layer undirected weighted network; S42. According to the overall multi-layer undirected weighted network constructed in S2, the calculation formula of the node betweenness is: Among them, P jk is the number of shortest paths between node j and node k; P jk (i) represents the number of shortest paths between node j and node k passing through node i; S43. According to the overall multi-layer undirected weighted network constructed in S2, the calculation formula of the node density is: Where d(i,j) represents the shortest path distance between node i and node j; S44. The degree I1(i), betweenness I2(i) and closeness I3(i) of the node calculated in S41, S42 and S43 will be used as evaluation indicators in the evaluation indicator system and applied in the subsequent step S5.

4. The dynamic evaluation method for the importance of transportation infrastructure nodes in peacetime and wartime according to claim 1 is characterized in that: In step S5, the specific method of using the objective weighting method to determine the objective weight value of each indicator in the three-level indicator system is: S51. In the three-level indicator system, there are m samples and n evaluation indicators, and the indicator matrix is ​​established: m is the number of samples, n is the number of evaluation indicators, x ij Take the value of the jth indicator of the i-th sample; S52. Dimensionlessize the jth indicator in the i-th sample. For indicators whose values ​​are larger, the better, i.e., benefit-type indicators, we have: For indicators with smaller values, the better, that is, cost-type indicators, there are: In the calculation formulas of benefit-based indicators and cost-based indicators, max j (x ij ) is the maximum value of the jth index in the i-th row of samples; min j (x ij ) is the minimum value of the jth indicator in the i-th row of samples; From this, the dimensionless evaluation matrix X′ can be calculated: S53. Calculate the mean of the index and standard deviation s j : S54. Calculate the correlation coefficient. The specific formula is: ρ ij is the correlation coefficient between the i-th indicator and the j-th indicator. The closer it is to 1, the greater the correlation. i,j=1,2,…,n,coυ(y i y j ) is the covariance of the i-th index and the j-th index of the evaluation matrix X′, S i and S j are the standard deviations of the i-th and j-th indicators respectively; S55. Calculate the amount of information. The specific formula is: E j is the amount of information of indicator j, The objective weight of indicator j is determined based on the difference between the jth indicator and other factors. The calculation formula is:

5. The dynamic evaluation method for the importance of transportation infrastructure nodes in peacetime and wartime according to claim 1 is characterized in that: According to game theory, the specific method for determining the comprehensive weight after combined optimization of the subjective weighting method and the objective weighting method based on the D-number hierarchical analysis method in the present invention is: S61. Assuming that there are L methods for weighting, there are L sets of weight vector engineering ωi={ω l1 ,ω l2 ,……,ω ln }, where l = 1, 2, ..., L; any combination of L weight combinations is where α l is the linear combination coefficient, α l >0, S62. Based on the basic idea of ​​game theory combination weighting, the optimization strategy model is as follows: S63, transform the optimal solution formula of S62 into the following set of equations by optimizing the first-order derivative condition, and calculate the combination coefficient set {α1, α2, ..., α L } S64, calculating the combined weight; The comprehensive weight vector is 6. The dynamic evaluation method for the importance of transportation infrastructure nodes in peacetime and wartime according to claim 1 is characterized in that: In step S7, the basic score and the usual importance ranking of each node are determined according to the TOPSIS method: S71, generate a normalized matrix T', the element t in T' ij 'for: S72. Calculate entropy value e j ; Then the entropy weight ω j for S73. Define the positive and negative ideal points in the vector space as: and Then the evaluation value of the jth ideal point is: in S74. Calculate the distance between the i-th node and the positive and negative ideal points: S75. The basic scores of each node are: in, and are the distances between the i-th node and the positive and negative ideal points in the TOPSIS algorithm; BS i is the relative closeness between the ith evaluation object and the maximum value, that is, the basic score of the node.

7. According to the dynamic evaluation method of traffic infrastructure node importance in peacetime and wartime according to claim 1, it is characterized in that: The performance score of each single-layer sub-network in step S8 is: S81. Hyperdegree refers to the number of connections between a node and other subnet nodes; let the hyperdegree of the i-th node be Hyperdegree(i), and the efficiency score of node i be FS i ,but FS i =Hyper degree(i)·BS i In the formula, BS i is the basic score of each node; S82. The performance score of the k-th single-layer undirected weighted subnetwork is:

8. The method for dynamically evaluating the importance of transportation infrastructure nodes in peacetime and wartime according to claim 1 is characterized in that: In step S9, the comprehensive performance CE of the entire multi-layer undirected weighted network is: S91, set G ab is the correlation between networks a and b. The number of edges in the overall multi-layer undirected weighted network is r. Traversing these edges, when the i-th edge connects the single-layer subnetworks a and b, it is recorded as Otherwise, record but Since the constructed network is an undirected network, G ab =G ba ; S92, set J ab is the closeness of the two single-layer sub-networks a and b in the overall multi-layer undirected weighted network; let the number of edges connecting the two sub-networks be r 1 The length of the i-th side is q i ,but Since the constructed network is an undirected network, J ab =J ba S93. The comprehensive performance CE of the overall multi-layer undirected weighted network is: CE=FS C (J CM G CM +J CS G CS )+FS M (J MC G MC +J MS G MS )+FS S (J SC G SC +J SM G SM ) Among them, J CM G CM represents the connection score between the sub-network and the single-layer sub-network M, J CS G CS Represents the connection score between the sub-network and the single-layer sub-network S; FS M represents the functional score within the sub-network, J MC G MC represents the connection score between the sub-network and the single-layer sub-network C, J MS G MS Represents the connection score between the sub-network and the single-layer sub-network S; FS S represents the functional score within the sub-network, J SC G SC represents the connection score between the sub-network and the single-layer sub-network C, J SM G SM Represents the connection score between this sub-network and the single-layer sub-network M.

9. The dynamic evaluation method for the importance of transportation infrastructure nodes in peacetime and wartime according to claim 1 is characterized in that: In step S10, the steps for calculating the comprehensive performance of the entire multi-layer network remaining after the nodes are dynamically deleted are: S101, assuming that the total number of nodes is n, delete node i and its edges, and recalculate the comprehensive efficiency CE′ of the remaining overall multi-layer undirected weighted network, and i=(1,n); S102, comparing the degree of decrease of the comprehensive efficiency CE′ of the remaining overall multi-layer undirected weighted network after each node and its edges are deleted relative to the comprehensive efficiency CE of the original network, so as to obtain the wartime node importance evaluation result; Let the system value of the node be SV: The larger the SV, the higher the value of the node to be analyzed. By sorting the SV values ​​from large to small, the strike priority order of the military transportation super network targets based on the node value can be generated.

Images