Container port cascade failure risk assessment method based on ship flow dynamic distribution

Abstract

The invention discloses a container port cascade failure risk assessment method based on ship flow dynamic allocation. The method comprises the following steps: extracting a route from mass container ship AIS trajectory data and constructing a directed weighted marine transportation network; generating an alternative alternative port set for the interrupted port based on network topology features and a spatial neighbor effect, and determining three optimal alternative ports of the interrupted port by integrating port multi-source marine transportation big data; a port cascade failure risk modeling method is constructed on the basis of a linear Moter-Lai model hypothesis; three port cascade failure risk assessment indexes in a sea transportation scene are constructed, the influence degree after port interruption in a container sea transportation network is assessed, and a dynamic cascade failure risk diffusion process and mechanism are explored in combination with a geographical visualization method. According to the method, the existing knowledge system of global dynamic risk assessment of the container transportation system is enriched, and an insight with important practical significance is provided for container cascade failure risk management and countermeasure research.

Description

Technical Field

[0001] The present invention relates to the field of container ocean transportation risks, and in particular to a container port cascading failure risk assessment method based on dynamic allocation of ship flows. Background Art

[0002] The container transportation system is a complex and interconnected network that plays a vital role in maintaining the stable and orderly operation of global trade relations. At the same time, ports are an important foundation for social and economic development and play an irreplaceable role in ensuring the stable operation of the global maritime supply chain. However, the current container transportation system is extremely vulnerable to international political events, extreme natural disasters and other emergencies.

[0003] In the prior art, the robustness assessment of shipping networks after port disruptions is mostly focused on static structural analysis, but the dynamic cascading failure risk propagation process of the disrupted ports is often ignored. Especially in the process of dynamic allocation of ship flows after port disruptions, relying solely on the topological structure characteristics of undirected / directed networks may lead to inaccurate and irrational dynamic allocation of ship flows, making it difficult to reflect the impact of port cascading failure risk propagation on the entire container shipping network. In addition, there is currently a lack of evaluation indicators for the propagation of cascading failure risks in container ports in shipping scenarios, resulting in a lack of accuracy and explainability when exploring the port cascading failure risk diffusion process and mechanism. Therefore, how to embed a reasonable ship flow dynamic allocation method in the cascading failure modeling process and accurately capture the port cascading failure risk diffusion process has become a technical problem that needs to be solved urgently. Summary of the invention

[0004] The present invention aims to solve the above-mentioned problems existing in the process of port cascading failure risk modeling under the background of port interruption in the prior art, and provides a container port cascading failure risk assessment method based on dynamic allocation of ship flow. The method extracts routes from massive container ship AIS trajectory data and constructs a directed weighted shipping network; generates a set of alternative replacement ports for interrupted ports based on network topology characteristics and spatial proximity effects, and determines the three optimal replacement ports for interrupted ports by integrating multi-source shipping big data including ports; constructs a port cascading failure risk modeling method based on the linear Motter-Lai model hypothesis; constructs three port cascading failure risk indicators under the shipping scenario, and evaluates the impact of port interruption in the container shipping network. Through the method of the present invention, it is possible to achieve effective simulation of the dynamic allocation of ship flows after port interruption, explore the dynamic cascading failure risk diffusion process and mechanism, enrich the existing knowledge system of dynamic risk assessment of global container transportation systems, and provide insights with important practical significance for container cascading failure risk management and countermeasure research.

[0005] In order to achieve the above-mentioned object of the invention, the technical solution adopted by the present invention is:

[0006] A container port cascading failure risk assessment method based on dynamic allocation of ship flows is provided, and the specific steps include:

[0007] S1. Use the shipping big data preprocessing method to eliminate noise data, clean the AIS data, and build a directed weighted container shipping network G based on massive AIS container ship trajectory data;

[0008] S2. Generate an alternative port set APCS for the interrupted port based on network topology characteristics and spatial proximity effect. By integrating multi-source shipping big data and based on the various differences between the interrupted port and each port in the alternative port set, a comprehensive decision-making method is used to determine the three best alternative ports for the interrupted port.

[0009] S3, distribute the ship flow load of the interrupted port to the three selected alternative ports in different proportions, make a linear assumption on the port operation capacity based on the Motter-Lai model, and construct a cascading failure modeling method;

[0010] S4. Construct multiple port cascading failure risk indicators in shipping scenarios, and evaluate the impact of port disruptions in the container shipping network on the shipping network based on the port cascading failure risk indicators, and explore the port cascading failure risk mechanism.

[0011] Furthermore, the specific method of step S1 includes the following sub-steps:

[0012] S1-1. Extract global container ship trajectories from massive AIS data, and extract container routes and frequencies based on ship MMSI numbers and arrival and departure times;

[0013] S1-2. Use complex network theory to construct a directed weighted container shipping network G, and use G = (N, E, W) to represent it. N represents the set of port nodes; E represents the set of route edges; ij =1 means that there is a directed edge from port i to port j in network G, E ij = 0 means that there is no directed edge from port i to port j in the network; W represents the weight set of route edges, W ij Corresponding to E ij , represents the frequency of routes from port i to port j.

[0014] Furthermore, the specific method of step S2 includes the following sub-steps:

[0015] S2-1. Construct the initial interruption port set based on the port's in-degree and out-degree values, and randomly initialize an interruption port as input; based on the network topology characteristics, use the shortest path algorithm to generate 3 alternative replacement ports for the interruption port; based on the spatial proximity effect, use r = 600KM as the search radius to generate 3 alternative replacement ports for the interruption port. Combine the above strategies to generate the alternative replacement port set APCS;

[0016] S2-2, extract all invalid routes passing through the interrupted port based on the network connectivity characteristics, and calculate the ship flow RL required to be allocated on each invalid route = {RL1, RL2, ..., RL m};

[0017] S2-3, based on the set of alternative ports for the interrupted ports and the ship flows required to be allocated for each invalid route, crawl multi-source shipping big data from the shipping platform SeaRates website. Calculate the shipping distance r between the interrupted port and each port in the set of alternative ports dist , load capacity difference r gap , Port scale comparison size In three aspects, all the alternative ports are ranked and the top three ports are selected as the best alternative ports.

[0018]

[0019] Furthermore, the specific method of step S3 includes the following sub-steps:

[0020] S3-1, according to the weighted out-degree S of the port in the initial shipping network i-in and weighted indegree S i-out To determine the initial operating capacity C of the port i (0):C i (0) = Max(S i-in, S i-out ). At the same time, a linear assumption is made for port operation capacity based on the Motter-Lai model: C i =(1+a)*C i (0);

[0021] S3-2, redistribute and balance the ship flow load of the interrupted port itself, and distribute it to the three selected alternative ports in a ratio of 3:2:1. After each round of ship flow distribution is completed, calculate the load increment ΔL of all ports j , based on the load capacity difference, determine whether the current port is overloaded. If overloaded: L i +ΔL i >C i , a new round of dynamic allocation of ship flows is triggered until the network collapses completely or no port is overloaded;

[0022] S3-3, record the interrupted ports newly caused by each port in each round of dynamic allocation of ship flows, and construct the interrupted port vector proVector, where proPort i represents the complete failure process of the i-th port.

[0023] proVector=concat(proPort1,proPort2,proPort3,…,proPort x )

[0024] Furthermore, the specific method of step S4 includes the following sub-steps:

[0025] S4-1. Use the number of new interrupted ports triggered by the initial interrupted port to evaluate the scale of port cascading failure risk CFS. The calculation formula is as follows: Where N represents the number of ports in the initial container shipping network G, i represents the initial interrupted port, N i It represents the number of remaining ports in the container shipping network G after the cascading risk caused by the interruption of port i spreads. The smaller the CFS value, the smaller the cascading failure risk of port i;

[0026] CFS=NN i

[0027] S4-2. Ship navigation efficiency is directly related to the spatiotemporal characteristics of the container shipping network. Therefore, we propose a new indicator SVER that takes into account the actual ship flow distribution and the sea distance to measure the impact of cascading failure risk on the container shipping network. The calculation formula is as follows. Where V′ represents the sum of the navigation efficiencies of all port pairs after the port disruption, and V represents the sum of the navigation efficiencies of all port pairs before the port disruption. Ship navigation efficiency refers to the reciprocal of the shortest path sea transportation distance between port pairs;

[0028] SVER=V′ / V

[0029] S4-3. When a port is interrupted, the dynamic allocation process of ship flow is likely to cause an increase or decrease in ship flow on other routes. Calculating the ship flow pressure FPSR of the route can help managers propose more detailed emergency management measures for the routes affected by the interrupted port. The calculation formula is as follows. ij represents the ship flow from port i to port j in the initial shipping network, ΔL ij It represents the load increment of the route from port i to port j after the ship flow distribution is completed. At the same time, according to the increase or decrease of the ship flow pressure FPSR of the route, the risk mechanism of port cascading failure can be explored in combination with geographic visualization.

[0030] FPSR Eij =(L ij +ΔLij ) / L ij .

[0031] The container port cascading failure risk assessment method based on dynamic allocation of ship flows in the embodiment of the present application has the following beneficial effects:

[0032] 1. By implementing a method for generating a set of alternative ports for interrupted ports taking into account multiple strategies on a directed weighted shipping network, the present invention can more accurately grasp the dynamic allocation process of ship flows after port interruption.

[0033] 2. The present invention significantly improves the understanding and control of the diffusion process and mechanism of port cascading failure risk by alternative port generation, dynamic ship flow allocation, and a cascading failure risk assessment model constructed in combination with the linear Motter-Lai hypothesis, enriching the existing knowledge system of dynamic risk modeling of container transportation systems.

[0034] 3. The present invention constructs multiple port cascading failure risk assessment indicators in shipping scenarios to effectively assess the impact of port disruptions in container shipping networks, which helps to understand and assess the spread of dynamic cascading failure risks from multiple perspectives and measure the vulnerability characteristics of marine transportation systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 The overall process diagram of this method is shown in FIG.

[0036] Figure 2 Constructing a flow chart for the directed weighted container shipping network implemented by the method;

[0037] Figure 3 Generate a flow chart for the replacement of interrupted ports for the implementation of this method;

[0038] Figure 4 Flowchart of the cascading failure modeling approach based on dynamic ship flow allocation implemented for this method. DETAILED DESCRIPTION

[0039] The specific implementation of the present invention is described below to facilitate the understanding of the present invention by those skilled in the art, but it should be clear that the described implementation is only a part of the implementation of the present invention, rather than all the implementations. The present invention is not limited to the scope of the specific implementation. For those skilled in the art, as long as various changes are within the spirit and scope of the present invention defined and determined by the attached claims, these changes are obvious, and all inventions and creations using the concept of the present invention are protected.

[0040] like Figure 1 As shown in the technical route, in one embodiment of the present invention, Figure 2 , 34 is an implementation example, which generally includes: building a directed weighted container shipping network based on massive AIS container ship trajectory data; determining the optimal alternative port for the interrupted port; dynamic allocation of ship flow and cascading failure modeling; evaluating the impact of port interruptions on the shipping network. The detailed steps include:

[0041] S1. Extract global container ship trajectories from massive AIS data, and extract container routes and frequencies based on ship MMSI numbers and arrival and departure times. Use complex network theory to construct a directed weighted container shipping network G, and use G = (N, E, W) to represent it. Among them, N represents the set of port nodes, and there are 1471 ports in the container shipping network G; E represents the set of route edges, and E represents the number of ports. ij =1 means that there is a directed edge from port i to port j in network G, E ij = 0 means that there is no directed edge from port i to port j in the network. There are 23,672 shipping routes in the container shipping network G. W represents the weight set of the route edges, W ij Corresponding to E ij , represents the frequency of routes from port i to port j, and its range is: [0,3538]. Figure 2 Taking port a to port b and port f as an example, E ab =1 and W ab =5, E af =0,W ab =0.

[0042] S21, construct the initial interrupt port set based on the port's inDegree and outDegree values. Figure 3 For example, since ports a, c, f, and g only have in-degree or out-degree, the final initial interruption port set is {b, d, e}. Then, we randomly initialize an interruption port b as input, and based on the network topology characteristics, use the shortest path algorithm to generate alternative replacement ports {d, e} for the interruption port. Based on the spatial proximity effect, with r = 600KM as the search radius, we generate alternative replacement ports {f, g} for the interruption port b, and combine the above strategies to generate the alternative replacement port set APCS = {d, e, f, g};

[0043] S22, extract all invalid routes passing through the interrupted ports based on network connectivity features. Figure 3 For example, the invalid routes passing through the interrupted port are: a→b→c. And calculate the ship flow RL1=SF required for each invalid route bc =4, and use it as the ship flow that needs to be dynamically allocated to interrupt port b;

[0044] S23. Based on the set of alternative ports for the interrupted ports and the ship flows required to be allocated for each invalid route, multi-source shipping big data is crawled from the shipping platform SeaRates website (https: / / www.searates.com / cn / maritime). The shipping distance r between the interrupted port and each port in the set of alternative ports is calculated. dist , load capacity difference r gap , Port scale comparison size In three aspects, all alternative ports are ranked and the top three ports are selected as the best alternative ports: {g, e, d}.

[0045]

[0046] S31, based on the weighted out-degree S of the port in the initial shipping network i-in and weighted indegree S i-out To determine the initial operating capacity C of the port i (0):C i (0) = Max(S i-in, S i-out ).by Figure 4 For example, Figure 4 As shown, C a (0) = 9, C d (0) = 3, and so on. At the same time, a linear assumption is made on the port operation capacity based on the Motter-Lai model:

[0047] C i =(1+a)*C i (0). The tolerance parameter α is: {0.1, 0.2, …, 1.0};

[0048] S32, redistribute and balance the ship flow load of the interrupted port itself and distribute it to the three selected alternative ports in a ratio of 3:2:1. Figure 4 For example, the ship flow load distribution of port a→b is distributed to the optimal alternative ports: {d, k, j}. After each round of ship flow distribution is completed, the load increment ΔL of all ports is calculated. j , based on the load capacity difference, determine whether the current port is overloaded. If overloaded: L i +ΔL i >C i , a new round of dynamic allocation of ship flows is triggered until the network collapses completely or no port is overloaded. Figure 4 For example, L d +ΔL d >C d , then port b is the newly created interruption port;

[0049] S33, record the interrupted ports newly caused by each port in each round of dynamic allocation of ship flows, and construct the interrupted port vector proVector, where proPort i represents the complete failure process of the i-th port.

[0050] proVector=concat(proPort1, proPort2, proPort3,..., proPortx).

[0051] S41. Use the number of new interrupted ports triggered by the initial interrupted port to evaluate the scale of port cascading failure risk CFS. The calculation formula is as follows: Where N represents the number of ports in the initial container shipping network G, i represents the initial interrupted port, N i It represents the number of remaining ports in the container shipping network G after the cascading risk caused by the interruption of port i spreads. The smaller the CFS value, the smaller the cascading failure risk of port i;

[0052] CFS=NN i

[0053] S42. Ship navigation efficiency is directly related to the spatiotemporal characteristics of the container shipping network. Therefore, we propose a new indicator SVER that takes into account the actual ship flow distribution and the sea distance to measure the impact of cascading failure risk on the container shipping network. The calculation formula is as follows. Where V′ represents the sum of the navigation efficiencies of all port pairs after the port disruption, and V represents the sum of the navigation efficiencies of all port pairs before the port disruption. Ship navigation efficiency refers to the reciprocal of the shortest path sea transportation distance between port pairs;

[0054] SVER=V′ / V

[0055] S43. When a port is interrupted, the dynamic allocation process of ship flow is likely to cause an increase or decrease in ship flow on other routes. Calculating the ship flow pressure FPSR of the route can help managers propose more detailed emergency management measures for the routes affected by the interrupted port. The calculation formula is as follows. ij represents the ship flow from port i to port j in the initial shipping network, ΔL ij It represents the load increment of the route from port i to port j after the ship flow distribution is completed. At the same time, according to the increase or decrease of the ship flow pressure FPSR of the route, the risk mechanism of port cascading failure can be explored in combination with geographic visualization.

[0056] FPSR Eij =(L ij +ΔL ij ) / L ij

[0057] In summary, the present invention provides an efficient solution for effectively simulating the dynamic allocation of ship flows after port disruptions and exploring the dynamic cascading failure risk diffusion process and mechanism through the construction of a directed weighted shipping network, the generation of optimal alternative ports, the construction of a multi-strategy port cascading failure risk modeling method, and the construction of a port cascading failure risk index in the process of port cascading failure risk modeling under the background of port disruptions. This method can enrich the existing knowledge system of dynamic risk assessment of the global container transportation system, provide support for the management and countermeasures of container cascading failure risks, and has important practical significance.

Claims

1. A container port cascading failure risk assessment method based on dynamic allocation of ship flows, characterized in that: The following steps are involved: S1. Use the shipping big data preprocessing method to eliminate noise data, clean the AIS data, and build a directed weighted container shipping network G based on massive AIS container ship trajectory data; S2. Generate an alternative port set APCS for the interrupted port based on network topology characteristics and spatial proximity effect. By integrating multi-source shipping big data and based on the various differences between the interrupted port and each port in the alternative port set, a comprehensive decision-making method is used to determine the three best alternative ports for the interrupted port. S3, distribute the ship flow load of the interrupted port to the three selected alternative ports in different proportions, make a linear assumption on the port operation capacity based on the Motter-Lai model, and construct a cascading failure modeling method; S4. Construct multiple port cascading failure risk indicators in shipping scenarios, and evaluate the impact of port disruptions in the container shipping network on the shipping network based on the port cascading failure risk indicators, and explore the port cascading failure risk mechanism.

2. The container port cascading failure risk assessment method based on dynamic allocation of ship flows according to claim 1 is characterized in that: The step S1 specifically includes the following steps: S1-1. Extract global container ship trajectories from massive AIS data, and extract container routes and frequencies based on ship MMSI numbers and arrival and departure times; S1-2. Use complex network theory to construct a directed weighted container shipping network G, and use G = (N, E, W) to represent it, where N represents the set of port nodes; E represents the set of route edges, and E represents the set of route edges. ij =1 means that there is a directed edge from port i to port j in network G, E ij = 0 means that there is no directed edge from port i to port j in the network. W represents the weight set of route edges. ij Corresponding to E ij , represents the frequency of routes from port i to port j.

3. The container port cascading failure risk assessment method based on dynamic allocation of ship flows according to claim 1 is characterized in that: The step S2 specifically includes the following steps: S2-1, construct the initial interruption port set based on the port's in-degree and out-degree values, and randomly initialize an interruption port as input. Based on the network topology characteristics, use the shortest path algorithm to generate three alternative replacement ports for the interruption port; Based on the spatial proximity effect, three alternative replacement ports are generated for the interrupted port with a search radius of r = 600 km, and the above strategies are combined to generate the set of alternative replacement ports APCS; S2-2, extract all invalid routes passing through the interrupted port based on the network connectivity characteristics, and calculate the ship flow RL required to be allocated on each invalid route = {RL1, RL2, ..., RL m }; S2-3, based on the set of alternative replacement ports for interrupted ports and the ship flows required to be allocated for each invalid route, crawl multi-source shipping big data from the shipping platform SeaRates website; Calculate the shipping distance r between the interruption port and each port in the alternative port set dist , load capacity difference r gap , Port scale comparison size In three aspects, all the alternative ports are ranked and the top three ports are selected as the best alternative ports.

4. The container port cascading failure risk assessment method based on dynamic allocation of ship flows according to claim 1 is characterized in that: The step S3 specifically includes the following steps: S3-1, according to the weighted out-degree S of the port in the initial shipping network i-in and weighted indegree S i-out To determine the initial operating capacity C of the port i (0):C i (0) = Max(S i-in ,S i-out ); At the same time, a linear assumption is made on the port operation capacity based on the Motter-Lai model: C i =(1+a)*C i (0); S3-2. Redistribute and balance the ship flow load of the interrupted port itself and distribute it to the three selected alternative ports in a ratio of 3:2:1; after each round of ship flow distribution is completed, calculate the load increment ΔL of all ports j , based on the load capacity difference, determine whether the current port is overloaded; if overloaded: L i +ΔL i >C i , a new round of dynamic allocation of ship flows is triggered until the network collapses completely or no port is overloaded; S3-3, record the interrupted ports newly caused by each port in each round of dynamic allocation of ship flows, and construct the interrupted port vector proVector, where proPort i represents the complete failure process of the i-th port; proVector=concat(proPort1,proPort2,proPort3,…,proPort x )。。 5. The container port cascading failure risk assessment method based on dynamic allocation of ship flows according to claim 1 is characterized in that: The step S4 specifically includes the following steps: S4-1. The scale of port cascading failure risk CFS is evaluated by using the number of new interrupted ports triggered by the initial interrupted port. The calculation formula is as follows: where N represents the number of ports in the initial container shipping network G, i represents the initial interrupted port, and N i It represents the number of remaining ports in the container shipping network G after the cascading risk caused by the interruption of port i spreads; the smaller the CFS value, the smaller the cascading failure risk of port i; CFS=N-N i S4-2. The ship navigation efficiency is directly related to the spatiotemporal characteristics of the container shipping network. Therefore, the present invention proposes a new indicator SVER that takes into account the actual ship flow distribution and the sea distance to measure the impact of cascading failure risk on the container shipping network. The calculation formula is as follows: V′ represents the sum of the navigation efficiencies of all port pairs after the port interruption, V represents the sum of the navigation efficiencies of all port pairs before the port interruption, and the ship navigation efficiency refers to the reciprocal of the shortest path sea distance between port pairs; SVER=V′ / V S4-3. When a port is interrupted, the dynamic allocation process of ship flow is likely to cause an increase or decrease in ship flow on other routes. Calculating the ship flow pressure FPSR of the route can help managers propose more detailed emergency management measures for the routes affected by the interrupted port. The calculation formula is as follows; where L ij represents the ship flow from port i to port j in the initial shipping network, ΔL ij It represents the load increment of the route from port i to port j after the ship flow distribution is completed; at the same time, according to the increase or decrease of the ship flow pressure FPSR of the route, the risk mechanism of port cascading failure can be explored in combination with geographic visualization; FPSR Eij =(L ij +ΔL ij ) / L ij 。

Images