A climate-adaptive recovery approach for container shipping networks integrating resilience theory

Abstract

The present invention is applicable to the field of shipping technology and provides a method for climate-adaptive recovery of container shipping networks that integrates resilience theory. The method comprises: acquiring liner schedule data, vessel dynamic data, route adjustment data, port geographic coordinate data, and multi-scale geo-fence data, and introducing a time dimension parameter to construct a spatiotemporal-weighted container shipping dynamic network; quantifying the strength of cascading failures in the spatiotemporal-weighted container shipping dynamic network, and then calculating the comprehensive efficiency index of the spatiotemporal-weighted container shipping dynamic network; quantifying the resilience index of the spatiotemporal-weighted container shipping dynamic network under extreme weather events; and evaluating the cost-effectiveness of climate-adaptability enhancement strategies for the spatiotemporal-weighted container shipping dynamic network. The present invention can quantitatively analyze the impact mechanism of extreme weather events on network efficiency and comprehensively evaluate the network's resilience and adaptability under extreme weather events.

Description

Technical Field

[0001] The present invention relates to the field of shipping technology, and in particular to a climate adaptability recovery method for a container shipping network integrating resilience theory. Background Art

[0002] With the intensification of global climate change, the impact of extreme weather events (such as hurricanes, storm surges, and floods) on container shipping networks is becoming increasingly significant. Existing technologies mainly focus on the following aspects: 1. The impact of extreme weather events on ports and waterways: Existing research mainly focuses on the direct impact of extreme weather events on individual ports or waterways, such as damage to port facilities and blockage of waterways. 2. Vulnerability analysis of container shipping networks: Some studies have analyzed the vulnerability of container shipping networks through complex network theory, pointing out that the robustness of container shipping networks is relatively weak, and small-scale attacks or natural disasters may cause large-scale failures of the network. 3. Application of resilience theory: The application of resilience theory in the field of transportation is gradually increasing, especially in urban transportation networks and road networks. However, research on the resilience of container shipping networks is still in its infancy, and most existing research focuses on the theoretical level, lacking specific quantitative evaluation and recovery strategies.

[0003] Existing research has shown that container ports are crucial infrastructure in my country and important hubs in its integrated transportation system. Research on climate-adaptive recovery strategies for container shipping networks, integrating resilience theory, at my country's coastal container ports is of great significance for enhancing the resilience of port infrastructure and implementing green development requirements.

[0004] At present, (1) there is a lack of quantitative analysis of the impact mechanism of extreme weather events: most existing studies remain at the qualitative analysis level, lacking quantitative analysis of how extreme weather events affect the efficiency of container shipping networks, especially insufficient research on network cascade effects. (2) The shipping network resilience assessment system is imperfect: existing studies lack systematic and multi-dimensional assessments of the resilience of container shipping networks, and fail to fully consider the network's recovery capacity and adaptability under extreme weather events. (3) The implementation of recovery strategies is poor: most recovery strategies proposed in existing studies remain at the theoretical level, lack practical operability and economic feasibility, and are difficult to connect with existing port and waterway planning. Therefore, it is necessary to provide a climate-adaptive recovery method for container shipping networks that integrates resilience theory to address the above problems. Summary of the Invention

[0005] In view of the shortcomings of the existing technology, the purpose of the present invention is to provide a climate adaptability recovery method for container shipping networks that integrates resilience theory to solve the problems existing in the above-mentioned background technology.

[0006] The present invention is implemented as follows: a method for climate adaptive recovery of container shipping networks that integrates resilience theory, the method comprising the following steps:

[0007] Obtain liner schedule data, vessel dynamic data, route adjustment data, port geographic coordinate data, and multi-scale geo-fence data, and introduce time dimension parameters to construct a spatiotemporal weighted container shipping dynamic network;

[0008] quantifying the strength of cascading failures in the spatiotemporal weighted container shipping dynamic network based on a cascading failure propagation model, and then calculating a comprehensive efficiency index of the spatiotemporal weighted container shipping dynamic network based on transportation timeliness, operational economy, and network robustness;

[0009] Based on the resilience triangle theory, the resilience index of the spatiotemporal weighted container shipping dynamic network under extreme weather events is quantified;

[0010] Evaluate the cost-effectiveness of climate suitability improvement strategies for the proposed spatiotemporally weighted dynamic container shipping network.

[0011] As a further solution of the present invention: the liner plan data includes the route code, starting and ending ports and the name of the shipping company, and the ship dynamic data includes the latitude and longitude coordinates, speed, heading, current status and next port of call.

[0012] As a further solution of the present invention, the step of constructing a spatiotemporal weighted container shipping dynamic network further includes:

[0013] (1) First, the traditional topology definition is performed:

[0014] Node collection: ,in, For the indivual ;

[0015] Edge collection: ,in, for The starting point, for The end point;

[0016] Static weight: ;

[0017] (2) Secondly, expand the time dimension:

[0018] The first step is to discretize the time layer and divide the continuous time axis into equally spaced time periods:

[0019]

[0020] in, is the number of days;

[0021] The second step is to build a three-dimensional space-time network , where each spatiotemporal node is ;

[0022] The third step is to define the spatiotemporal edge weight as a multidimensional vector :

[0023]

[0024] in, ; for During the period Actual sailing time; for During the period liner frequency; for During the period Available capacity;

[0025] (3) Again, perform dynamic parameter quantization:

[0026] The first step is to calculate During the period Actual sailing time :

[0027]

[0028] in, For distance, It is the undisturbed speed; for Wind and wave level coefficient during the period; for Actual route cost index during the period; 、 is the calibration parameter, for The maximum route cost index within the time period;

[0029] The second step is to calculate During the period Frequencies of liner services :

[0030]

[0031] in, for Annual benchmark flights; is the seasonal fluctuation amplitude; for During the period Temporary increase or decrease of shifts;

[0032] The third step is to calculate During the period Available capacity :

[0033]

[0034] in, for Number of ships deployed during the time period; For ships The actual utilization rate of the voyage; For ships Rated load capacity; for Ship schedule reliability coefficient within the time period;

[0035] Based on the above steps, the time-space weighted container shipping dynamic network is constructed, wherein the time-space weighted container shipping dynamic network is a physical layer network.

[0036] As a further solution of the present invention, the quantification of the strength of the cascading failure of the spatiotemporal weighted container shipping dynamic network based on the cascading failure propagation model further includes:

[0037] (1) First, after the construction of the spatiotemporal weighted container shipping dynamic network is completed, the logical layer network is further constructed. , where the logical node It is a set of cargo flow paths, representing the transportation path of cargo from the starting point to the end point; logical edge is a set of transport capacity scheduling relationships, representing the transport capacity allocation or scheduling relationship between paths;

[0038] (2) Secondly, define the bidirectional impact factor between nodes :

[0039]

[0040] When the physical node When the logical node fails, The load will be Proportional interruption;

[0041] (3) Again, construct a cascading failure propagation model:

[0042] First define the initial load: the load of the physical node is the port weekly throughput , the load of the logical node is the capacity demand of the path ;

[0043] When the node When a node fails, its load is distributed according to the capacity ratio of adjacent nodes:

[0044]

[0045] in, is a node Newly added loads; is the allocation preference coefficient, when σ>1, the nodes with high degree are preferred; when σ<1, the nodes with low degree are preferred; The capacity transfer efficiency is the ratio of the load of the failed node transferred to the adjacent nodes; Is a failed node The original load; is a node The degree of Power, representing a node influence; is the degree of all adjacent nodes the sum of powers;

[0046] Wherein, in constructing the cascading failure propagation model, it also includes:

[0047] Construct the propagation dynamics equation:

[0048]

[0049] in, is the elasticity coefficient of the physical facility; For physical node p The failure state at the moment; Logical Node In time load; Logical Node In time load; Logical Node In time load; For physical nodes In time load; For physical nodes In time load;

[0050] (4) Finally, based on the cascading failure propagation model, the strength of the cascading failure of the spatiotemporal weighted container shipping dynamic network is quantified, revealing the cascading effect of the spatiotemporal weighted container shipping dynamic network.

[0051] As a further solution of the present invention: after constructing the spatiotemporal weighted container shipping dynamic network, the method further includes:

[0052] Construct linear programming models;

[0053] Performing data cleaning and entity alignment on the spatiotemporal weighted container shipping dynamic network through the linear programming model;

[0054] Wherein, the construction of the linear programming model further includes:

[0055] Define the decision variables:

[0056] Objective function:

[0057] Constraints:

[0058] in, Assemble for routes; for gather; For ships Rated load capacity; For routes Total capacity demand; The loading and unloading time for a single vessel is , For the efficiency of the quay crane, is the number of quay cranes; For ships The maximum total available laytime, For ships The maximum number of available shifts.

[0059] As a further solution of the present invention, the calculation of the comprehensive efficiency index of the time-space weighted container shipping dynamic network based on transportation timeliness, operational economy, and network robustness further includes:

[0060] The first step is to determine the boundaries of the spatiotemporal weighted container shipping dynamic network, and determine the number of nodes, routes, and cargo flows covered in the evaluation area;

[0061] The second step is to establish a three-tier indicator system including transportation timeliness, operational economy and network robustness;

[0062] (1) Transport timeliness indicators include average speed ratio and on-time arrival rate Two sub-indicators:

[0063] Average speed ratio The calculation formula is:

[0064]

[0065] in, For routes design speed; For routes Actual average speed; For routes any vessel on board; For routes The collection of ships on board; Assemble for the ship China Shipbuilding the total number of

[0066] On-time arrival rate The calculation formula is:

[0067]

[0068] in, For ships Scheduled arrival time; For ships The actual arrival time of mustering for ships; is the allowed time deviation threshold; is an indicator function, which takes 1 when the condition is met and 0 otherwise;

[0069] (2) Operational economic indicators include unit TEU cost and capacity utilization Two sub-indicators:

[0070] Unit TEU cost The calculation formula is:

[0071]

[0072] in, For fuel costs, For ships Total distance traveled, For ships The average speed of is the main engine fuel consumption rate, is the unit price of fuel;

[0073] in, For port fees, , For the port Basic berthing fee, It is a port The total number of containers, For the port Single box loading and unloading fee, mustering for the port;

[0074] in, For canal tolls, is the overall length of the ship, For the width of the ship, is the deadweight tonnage;

[0075] in, For other costs;

[0076] Capacity Utilization The calculation formula is:

[0077]

[0078] in, For ships The actual loading box volume, For ships Maximum loading capacity, mustering for ships;

[0079] (3) Network robustness indicators include node betweenness centrality and edge redundancy index Two sub-indicators:

[0080] Node betweenness centrality The calculation formula is:

[0081]

[0082] in, is the total number of shortest paths from node s to node t; is the number of shortest paths passing through node v; where the summation range is for all node pairs sum;

[0083] Node betweenness centrality Perform normalization:

[0084] Wherein, V is the total number of nodes in the spatiotemporal weighted container shipping dynamic network; is the normalized node betweenness centrality, ranging from [0, 1];

[0085] Edge redundancy index The calculation formula is:

[0086]

[0087] in, For the path collection , the number of paths for which there are alternative paths; is the total number of edges in the spatiotemporal weighted container shipping dynamic network; For the side , the proportion of paths with alternative paths;

[0088] The definition of alternative paths is: first, for each edge (u→v), find the set of all paths that use this edge ; For each path Then check whether there is an alternative path p' that satisfies: p' does not pass through edge e, and the weight of p' is ≤ 1.2 × the weight of the original path; finally, count the proportion of paths with alternative paths;

[0089] The third step is to calculate the comprehensive efficiency index of the time-space weighted container shipping dynamic network based on the three-layer index system; wherein the comprehensive efficiency index of the time-space weighted container shipping dynamic network is The calculation formula is:

[0090]

[0091] in, are weights respectively.

[0092] As a further solution of the present invention, the quantification of the resilience index of the spatiotemporal weighted container shipping dynamic network under extreme weather events based on the resilience triangle theory further includes:

[0093] The calculation formula of the resilience index Re of the spatiotemporal weighted container shipping dynamic network under extreme weather events is as follows:

[0094]

[0095] Where, between 0 and 1; represents the starting time when the spatiotemporal weighted container shipping dynamic network is affected by extreme weather events, represents the time when the spatiotemporal weighted container shipping dynamic network returns to its original state, Indicates time Global network efficiency; represents the initial network efficiency;

[0096] Among them, if , indicating that the spatiotemporal weighted container shipping dynamic network is not affected by extreme weather events, and the performance of the spatiotemporal weighted container shipping dynamic network remains original;

[0097] Among them, if , indicating that the extreme weather event has a negative impact on the spatiotemporal weighted container shipping dynamic network, and a recovery strategy needs to be adopted to restore the performance of the spatiotemporal weighted container shipping dynamic network;

[0098] Among them, if , indicating that the spatiotemporal weighted container shipping dynamic network is severely damaged by the destructive event and will take a long time to recover.

[0099] As a further solution of the present invention, the cost-effectiveness of the climate suitability improvement strategy for the spatiotemporal weighted container shipping dynamic network is further comprised of:

[0100] Construct an improved greedy algorithm based on a multi-constrained marginal cost optimization model;

[0101] Evaluating the cost-effectiveness of different climate suitability improvement strategies for the spatiotemporal weighted container shipping dynamic network based on the improved greedy algorithm to determine the climate suitability improvement strategy with the best cost-effectiveness;

[0102] A climate adaptive recovery strategy for the spatiotemporal weighted container shipping dynamic network is formulated based on a dynamic adaptive policy path enhancement framework to ensure the rapid recovery of the spatiotemporal weighted container shipping dynamic network under extreme weather events.

[0103] As a further solution of the present invention: the improved greedy algorithm based on the multi-constrained marginal cost optimization model further includes:

[0104] First, define a four-dimensional cost vector: C = [economic cost, carbon emissions, recovery time, climate vulnerability];

[0105] Secondly, construct the multi-constrained marginal cost optimization model :

[0106]

[0107] in, For the the implementation intensity of various climate suitability improvement strategies; For the port Climate Vulnerability Index; is the climate risk sensitivity coefficient; the economic costs of implementing climate suitability improvement strategies; Carbon emissions resulting from the implementation of climate suitability improvement strategies; The recovery time of ports or shipping routes after climate disasters; is the climate vulnerability threshold; 、 、 is the weight coefficient; To strengthen the implementation of the strategy Find partial derivatives;

[0108] The multi-constrained marginal cost optimization model is dynamically updated using the NSGA-II multi-objective optimization framework. The weight coefficient in 、 、 .

[0109] As a further solution of the present invention: the dynamic adaptive policy path enhancement framework specifically includes:

[0110] (1) Constructing a system for quantifying system failure signals:

[0111] ATP calculation model:

[0112]

[0113] in, For the port Real-time wind speed; For the port The highest wind speed in history; For the port cumulative rainfall; port The maximum cumulative rainfall in history; For the port hinterland economic equivalent; is the sum of the economic equivalents of all port hinterlands; is the total number of ports;

[0114] (2) Strategy tree construction based on deep reinforcement learning:

[0115] State space: ATP level, network connectivity, and ship availability;

[0116] Action space: {reroute, temporary call, activation of alternate port, initiation of land transport connection};

[0117] Reward function: ;

[0118] in, and is the weight coefficient; Cost changes after implementation of climate-resilient restoration strategies; Temporal changes after implementation of climate-adaptive restoration strategies; is the reward coefficient; The restoration benefits after the implementation of climate-adaptive restoration strategies.

[0119] Compared with the prior art, the present invention has the following beneficial effects:

[0120] (1) Quantitative analysis of the impact mechanism of extreme weather events: By constructing the topological structure and efficiency evaluation method of the container shipping network, the impact mechanism of extreme weather events on network efficiency was quantitatively analyzed, especially the study of the network cascade effect;

[0121] (2) Multi-dimensional evaluation system: A multi-dimensional container shipping network resilience evaluation system is proposed to comprehensively evaluate the network's resilience and adaptability to extreme weather events;

[0122] (3) Economically feasible recovery strategy: Based on the greedy strategy of marginal cost minimization and the dynamic adaptive policy path, an economically feasible container shipping network recovery strategy is proposed, which can effectively improve the network's risk resistance and recovery speed. BRIEF DESCRIPTION OF THE DRAWINGS

[0123] Figure 1 This is a flow chart of a method for climate adaptability recovery of container shipping networks that integrates resilience theory, provided by the present invention.

[0124] Figure 2 This is a structural diagram of the dynamic adaptive policy path enhancement framework provided by the present invention. DETAILED DESCRIPTION

[0125] In order to make the purpose, technical solutions and advantages of the present invention clearer, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0126] The specific implementation of the present invention is described in detail below with reference to specific embodiments.

[0127] like Figure 1 As shown, an embodiment of the present invention provides a method for climate adaptive recovery of container shipping networks integrating resilience theory, the method comprising the following steps:

[0128] Step S1: Acquire liner schedule data, vessel dynamic data, route adjustment data, port geographic coordinate data, and multi-scale geo-fence data, and introduce time dimension parameters to construct a spatiotemporal weighted container shipping dynamic network;

[0129] In an embodiment of the present invention, the liner plan data includes a route code, starting and ending ports, and a shipping company name, and the ship dynamic data includes latitude and longitude coordinates, speed, heading, current status (underway / anchored / berthed), and next port of call.

[0130] It should be noted that (1) liner schedule data and ship dynamic data are structured data. The present invention captures liner schedule data and ship dynamic data from shipping databases such as Alphaliner in real time through an API interface; (2) route adjustment data are unstructured data. The present invention uses natural language processing (NLP) technology to parse route adjustment text information in port announcements and shipping company notices. The sources of unstructured data include port official website announcements, shipping company emails, social media notifications, etc. The above unstructured data are cleaned based on Python; (3) port geographic coordinate data and multi-scale geographic fence data are spatial data. The present invention can integrate port geographic coordinates and multi-scale geographic fence data through a GIS platform (ArcGIS / QGIS).

[0131] Based on the above structured data, unstructured data and spatial data, a time-space weighted container shipping dynamic network is modeled. That is, on the basis of the traditional static topology, time dimension parameters (such as ship scheduling intervals and seasonal capacity fluctuations) are introduced to construct a time-space weighted container shipping dynamic network (Space-Time Network). The edge weights include sailing time, liner frequency and capacity dynamic change factors.

[0132] Preferably, the step of constructing a spatiotemporal weighted container shipping dynamic network further includes:

[0133] (1) First, the traditional topology definition is performed:

[0134] Node collection: ,in, For the indivual ;

[0135] Edge collection: ,in, for The starting point, for The end point;

[0136] Static weight: ;

[0137] (2) Secondly, expand the time dimension:

[0138] The first step is to discretize the time layer and divide the continuous time axis into equally spaced time periods (such as by week):

[0139]

[0140] in, is the number of days;

[0141] The second step is to build a three-dimensional space-time network , where each spatiotemporal node is ;

[0142] The third step is to set the dynamic edge weight function:

[0143] Define spatiotemporal edge weights as multidimensional vectors :

[0144]

[0145] in, ; for During the period Actual sailing time (including weather conditions); for During the period Frequency of liner services (sails / week); for During the period Available capacity (TEU / week);

[0146] (3) Again, perform dynamic parameter quantization:

[0147] The first step is to calculate During the period Actual sailing time :

[0148]

[0149] in, For distance, is the undisturbed speed (knots); for Wind and wave level coefficient within the time period (value ranges from 0 to 1); for Actual route cost index during the period; 、 For calibration parameters (recommended , ), for The maximum route cost index during the time period;

[0150] The second step is to consider the seasonal adjustment and emergency scheduling of liner frequency and calculate During the period Frequencies of liner services :

[0151]

[0152] in, for Annual benchmark flights; is the seasonal fluctuation amplitude (usually 0.2~0.5); for During the period Temporary increase or decrease of flights (such as suspension of service during the freezing period);

[0153] The third step is to calculate During the period Available capacity :

[0154]

[0155] in, for Number of ships deployed during the time period; For ships The actual utilization rate of the voyage; For ships Rated load capacity; for The reliability coefficient of the ship schedule within the time period (the value ranges from 0 to 1);

[0156] Based on the above steps, the time-space weighted container shipping dynamic network is constructed, wherein the time-space weighted container shipping dynamic network is a physical layer network.

[0157] While existing research is largely based on static topology analysis, this paper utilizes time slicing technology to model the dynamic evolution of networks, capturing the instantaneous impact of events such as temporary suspensions and seasonal adjustments. This spatiotemporal weighted container shipping dynamic network significantly reduces robustness testing errors and more closely reflects actual operational scenarios than traditional static models.

[0158] Specifically, after constructing the spatiotemporal weighted container shipping dynamic network, the method further includes:

[0159] A Port Knowledge Graph (PortKG) was constructed, embedding multilingual port aliases, historical names, and geographic coordinate features to resolve naming conflicts such as "Shenzhen Port / Shenzhen Port / Shekou Port." A linear programming model was designed to verify the throughput-capacity relationship. The objective function was weekly port throughput ≥ ∑(weekly route frequency × single-vessel capacity), with constraints including terminal operation time windows and loading and unloading efficiency thresholds. Specifically:

[0160] Construct linear programming models;

[0161] Performing data cleaning and entity alignment on the spatiotemporal weighted container shipping dynamic network through the linear programming model;

[0162] Wherein, the construction of the linear programming model further includes:

[0163] Define the decision variables:

[0164] Objective function: (Maximize total capacity)

[0165] Constraints:

[0166] in, Assemble for routes; for gather; For ships Rated container capacity, i.e. the ship The number of standard containers that can be loaded; For routes Total capacity demand; The loading and unloading time for a single vessel is , For quay crane efficiency, for example, is the number of quay cranes; For ships The maximum total available laytime, For ships The maximum number of available shifts.

[0167] Step S2: quantifying the strength of cascading failures in the spatiotemporal weighted container shipping dynamic network based on a cascading failure propagation model, and then calculating a comprehensive efficiency index of the spatiotemporal weighted container shipping dynamic network based on transportation timeliness, operational economy, and network robustness;

[0168] Preferably, the quantifying the strength of the cascading failure of the spatiotemporal weighted container shipping dynamic network based on the cascading failure propagation model further includes:

[0169] (1) First, after the construction of the spatiotemporal weighted container shipping dynamic network is completed, the logical layer network is further constructed. , where the logical node It is a set of cargo flow paths, representing the transportation path of cargo from the starting point to the end point; logical edge is a set of transport capacity scheduling relationships, representing the transport capacity allocation or scheduling relationship between paths;

[0170] (2) Secondly, define the bidirectional impact factor between nodes :

[0171]

[0172] When the physical node When the logical node fails, The load will be Proportional interruption; for example, the total cargo volume of a route is 1000 TEU, and the port 300 TEUs handled, then ( path) = 30%. If the port If it fails, the load on this path will be reduced by 30% (i.e. 300 TEU).

[0173] (3) Again, construct a cascading failure propagation model:

[0174] First define the initial load: the load of the physical node is the port weekly throughput , the load of the logical node is the capacity demand of the path ;

[0175] When the node When a node fails, its load is distributed according to the capacity ratio of adjacent nodes:

[0176]

[0177] in, is a node Newly added load (unit: TEU / week); is the allocation preference coefficient (when it is 1, it is evenly distributed according to degree). When σ>1, nodes with high degree are preferred; when σ<1, nodes with low degree are preferred. is the capacity transfer efficiency (usually 0.6–0.8), which indicates the proportion of load from a failed node transferred to adjacent nodes; Is a failed node Original load (unit: TEU / week); is a node The degree (number of connections) of Power, representing a node influence; is the degree of all adjacent nodes the sum of powers;

[0178] Yes, node It can be a physical node or a logical node, and needs to be distinguished based on the failure cause:

[0179] (1) Physical node failure (such as port paralysis): The load is distributed to adjacent physical nodes or logical nodes.

[0180] (2) Logical node failure (such as path interruption): the load is distributed to the alternative logical node.

[0181] Wherein, in constructing the cascading failure propagation model, it also includes:

[0182] Construct the propagation dynamics equation:

[0183]

[0184] in, is the elasticity coefficient of the physical facility; For physical node p The failure status at the moment (0 for failure, or 1 for not failure); Logical Node In time Load (unit: TEU / week); Logical Node In time Load (unit: TEU / week); Logical Node In time Load (unit: TEU / week); For physical nodes In time Load (unit: TEU / week); For physical nodes In time Load (unit: TEU / week);

[0185] (4) Finally, based on the cascading failure propagation model, the strength of the cascading failure of the spatiotemporal weighted container shipping dynamic network is quantified, revealing the cascading effect of the spatiotemporal weighted container shipping dynamic network.

[0186] Preferably, the calculation of the comprehensive efficiency index of the time-space weighted container shipping dynamic network based on transport timeliness, operational economy and network robustness further includes:

[0187] The first step is to determine the boundaries of the spatiotemporal weighted container shipping dynamic network, and determine the number of nodes, routes, and cargo flows covered in the evaluation area;

[0188] The second step is to establish a three-tier indicator system including transportation timeliness, operational economy and network robustness;

[0189] (1) Transport timeliness indicators include average speed ratio and on-time arrival rate Two sub-indicators:

[0190] Average speed ratio The calculation formula is:

[0191]

[0192] in, For routes Design speed (knots); For routes Actual average ship speed (calculated from AIS data); For routes any vessel on board; For routes The collection of ships on board; Assemble for the ship China Shipbuilding the total number of

[0193] On-time arrival rate The calculation formula is:

[0194]

[0195] in, For ships Scheduled arrival time (UTC); For ships The actual arrival time of the port (based on the time of completion of berthing); mustering for ships; is the allowed time deviation threshold (usually 6 hours); is an indicator function, which takes 1 when the condition is met and 0 otherwise;

[0196] (2) Operational economic indicators include unit TEU cost and capacity utilization Two sub-indicators:

[0197] Unit TEU cost The calculation formula is:

[0198]

[0199] in, For fuel costs, For ships Total distance travelled (nautical miles), For ships The average speed of is the main engine fuel consumption rate (tons / day), which is related to the ship's deadweight tonnage; is the unit price of fuel (yuan / ton);

[0200] in, For port fees, , For the port Basic berthing fee, It is a port The total number of containers, For the port Single box loading and unloading fee, mustering for the port;

[0201] in, Canal tolls (charged according to ship size), is the total length of the ship (m), is the ship width (m), is the deadweight tonnage;

[0202] in, For other costs;

[0203] Capacity Utilization The calculation formula is:

[0204]

[0205] in, For ships The actual loading box volume, For ships Maximum loading capacity, mustering for ships;

[0206] (3) Network robustness indicators include node betweenness centrality and edge redundancy index Two sub-indicators:

[0207] Node betweenness centrality The calculation formula is:

[0208]

[0209] in, is the total number of shortest paths from node s to node t; is the number of shortest paths passing through node v; where the summation range is for all node pairs Sum (i.e. and are not equal to v);

[0210] Node betweenness centrality Perform normalization:

[0211] Wherein, V is the total number of nodes in the spatiotemporal weighted container shipping dynamic network; is the normalized node betweenness centrality, ranging from [0, 1];

[0212] Edge redundancy index The calculation formula is:

[0213]

[0214] Among them, the edge redundancy index The range is between [0, 1]; For the path collection , the number of paths for which there are alternative paths; is the total number of edges in the spatiotemporal weighted container shipping dynamic network; For the side , the proportion of paths with alternative paths;

[0215] The definition of alternative paths is: first, for each edge (u→v), find the set of all paths that use this edge ; For each path Then check whether there is an alternative path p' that satisfies: p' does not pass through edge e, and the weight of p' is ≤ 1.2 × the weight of the original path (allowing a 20% cost increase); finally, count the proportion of paths with alternative paths;

[0216] The third step is to calculate the comprehensive efficiency index of the time-space weighted container shipping dynamic network based on the three-layer index system; wherein the comprehensive efficiency index of the time-space weighted container shipping dynamic network is The calculation formula is:

[0217]

[0218] in, are weights respectively. For example, They are 0.18, 0.22, 0.25, 0.15, 0.12, and 0.08 respectively.

[0219] Step S3: quantifying the resilience index of the spatiotemporal weighted container shipping dynamic network under extreme weather events based on the resilience triangle theory;

[0220] Preferably, the quantification of the resilience index of the spatiotemporal weighted container shipping dynamic network under extreme weather events based on the resilience triangle theory further includes:

[0221] The calculation formula of the resilience index Re of the spatiotemporal weighted container shipping dynamic network under extreme weather events is as follows:

[0222]

[0223] Where, between 0 and 1, The higher the value, the smaller the resilience loss of the shipping network and the greater the structural capriciousness; represents the starting time when the spatiotemporal weighted container shipping dynamic network is affected by extreme weather events, represents the time when the spatiotemporal weighted container shipping dynamic network returns to its original state, Indicates time Global network efficiency; represents the initial network efficiency;

[0224] Among them, if , indicating that the spatiotemporal weighted container shipping dynamic network is not affected by extreme weather events, and the performance of the spatiotemporal weighted container shipping dynamic network remains original;

[0225] Among them, if , indicating that the extreme weather event has a negative impact on the spatiotemporal weighted container shipping dynamic network, and a recovery strategy needs to be adopted to restore the performance of the spatiotemporal weighted container shipping dynamic network;

[0226] Among them, if , indicating that the spatiotemporal weighted container shipping dynamic network is severely damaged by the destructive event and will take a long time to recover.

[0227] Step S4: Evaluate the cost-effectiveness of the climate suitability improvement strategy for the spatiotemporal weighted dynamic container shipping network.

[0228] Preferably, the evaluation of the cost-effectiveness of the climate suitability improvement strategy of the spatiotemporal weighted container shipping dynamic network further includes:

[0229] Construct an improved greedy algorithm based on a multi-constrained marginal cost optimization model;

[0230] Specifically, the improved greedy algorithm based on the multi-constrained marginal cost optimization model further includes:

[0231] First, define a four-dimensional cost vector: C = [economic cost, carbon emissions, recovery time, climate vulnerability];

[0232] Secondly, construct the multi-constrained marginal cost optimization model :

[0233]

[0234] in, For the the implementation intensity of various climate suitability improvement strategies; For the port a climate vulnerability index (based on historical disaster frequency and sea level rise projections); is the climate risk sensitivity coefficient (calibrated through Monte Carlo simulation); The economic costs of implementing climate suitability improvement strategies (e.g. port facility upgrades, route adjustments, etc.); Carbon emissions resulting from the implementation of climate-suitability enhancement strategies (taking into account ship fuel consumption, port operation equipment, etc.); The recovery time of ports or shipping routes after climate disasters (reflecting network resilience); is the climate vulnerability threshold, used to distinguish the port climate risk level. It indicates that ports with a climate vulnerability index exceeding 0.5 require special attention; 、 、 is a weight coefficient used to balance the importance of economic costs, carbon emissions, and recovery time. + + =1; To strengthen the implementation of the strategy Find the partial derivative and calculate when The rate of change of the weighted total cost (a linear combination of economic cost, carbon emissions, and recovery time) when the change occurs;

[0235] The multi-constrained marginal cost optimization model is dynamically updated using the NSGA-II multi-objective optimization framework. The weight coefficient in 、 、 .

[0236] The innovation of the improved greedy algorithm of the present invention compared with the traditional greedy algorithm is shown in Table 1 below:

[0237] Table 1

[0238] Dimensions Traditional methods This method Target Dimension Single economic cost Four-dimensional goals + climate vulnerability adjustment Parameter staticity Fixed weight Dynamic Weight Decision granularity Port-level strategies Port-Route-Vessel Collaboration Strategy

[0239] Evaluating the cost-effectiveness of different climate suitability improvement strategies for the spatiotemporal weighted container shipping dynamic network based on the improved greedy algorithm to determine the climate suitability improvement strategy with the best cost-effectiveness;

[0240] A climate-adaptive recovery strategy for the spatiotemporal weighted container shipping dynamic network is formulated based on the Dynamic Adaptive Policy Path (DAPP) enhanced framework to ensure the rapid recovery of the spatiotemporal weighted container shipping dynamic network under extreme weather events.

[0241] It should be noted that the dynamic adaptive policy path enhancement framework sets a system failure signal (ATP) and formulates adjustment countermeasures for each system failure signal in advance to ensure the rapid recovery of the spatiotemporal weighted container shipping dynamic network under extreme weather events.

[0242] In the embodiment of the present invention, Figure 2 As shown, the dynamic adaptive policy path enhancement framework specifically includes:

[0243] (1) Constructing a system for quantifying system failure signals:

[0244] For example, a three-level warning signal is constructed as shown in Table 2 below:

[0245] Table 2

[0246] ATP level Trigger Conditions Response Mechanism ATP1 The wind speed at the port is ≥ level 10 within 3 days Activate the ship diversion plan ATP2 Weekly cumulative rainfall > historical 95% percentile Activate backup dock + loading and unloading robot reinforcements ATP3 Real-time monitoring of sea level exceeding the warning line by 0.5m Perform network-wide cascade shutdown and diversion

[0247] ATP calculation model:

[0248]

[0249] in, For the port Real-time wind speed, unit: m / s or knots; For the port The highest wind speed in history; For the port The cumulative rainfall, unit: mm; port The maximum cumulative rainfall in history; For the port hinterland economic equivalent, unit: GDP / year; is the sum of the hinterland economic equivalents of all ports (used for weight normalization); is the total number of ports;

[0250] (2) Dynamic generation of policy paths:

[0251] Strategy tree construction based on deep reinforcement learning (DRL):

[0252] State space: ATP level, network connectivity, and ship availability;

[0253] Action space: {reroute, temporary call, activation of alternate port, initiation of land transport connection};

[0254] Reward function: ;

[0255] in, and is the weight coefficient, the sum of which is 1, which is used to balance the importance of cost and time (determined by expert scoring); The cost change after the implementation of climate adaptation restoration strategies (such as diversion costs, land transportation costs, unit: 10,000 yuan); The time change after the implementation of climate adaptation recovery strategies (e.g., ship delay time, unit: hours); is a reward factor used to amplify the impact of recovery benefits (needs to be calibrated according to network resilience goals); The restoration benefits after the implementation of climate-adaptive restoration strategies (such as reduced carbon emissions and improved network connectivity) are obtained.

[0256] Compared with traditional emergency response mechanisms (strategies), as shown in Table 3 below:

[0257] Table 3

[0258]

[0259] In the embodiment of the present invention, a response strategy is also proposed based on the cost-effectiveness and climate suitability of the solution: a response strategy for the container shipping network to adapt to climate change that integrates the resilience theory from the three perspectives of improving the structural resilience, timeliness and sustainability of the container shipping network. Specifically:

[0260] First, we need to improve the transport and carrying capacity of the container transport network to alleviate congestion and enhance overall resilience. Particular attention should be paid to facilities located in climate-sensitive areas to ensure they can withstand the impact of extreme weather events.

[0261] Secondly, we develop detailed transportation plans based on the nature of the goods and shipping routes, and establish an early warning mechanism based on real-time shipping data. This provides prompt alerts for any delays or anomalies, allowing for timely processing and ensuring timely delivery of goods.

[0262] The third approach is to replace traditional fuels with low-carbon fuels such as ammonia and liquefied natural gas (LNG) to reduce carbon emissions from ships. At the same time, efforts are underway to install auxiliary power systems such as sails and solar panels on ships to reduce fuel consumption and carbon dioxide emissions.

[0263] In summary, the embodiments of the present invention provide a method for climate-adaptive recovery of a container shipping network that integrates resilience theory, including: (1) a method for constructing a topological structure of a container shipping network: by collecting liner route data of shipping companies, the topological structure of a container shipping network is constructed to reveal the cascading effect of the network; (2) a method for evaluating the efficiency of a container shipping network: based on complex network theory, a method for evaluating the efficiency of a container shipping network is proposed to quantitatively analyze the impact of extreme weather events on network efficiency; (3) a method for quantitatively evaluating the resilience of a container shipping network: based on the resilience triangle theory, the resilience of a container shipping network under extreme weather events is quantified to evaluate the resilience and adaptability of the network; (4) a greedy strategy based on marginal cost minimization: a greedy strategy based on marginal cost minimization is proposed to evaluate the cost-effectiveness of different recovery strategies; and (5) a dynamic adaptive policy path (DAPP): through the dynamic adaptive policy path method, a climate-adaptive recovery strategy for a container shipping network is formulated to ensure rapid recovery of the network under extreme weather events.

[0264] The above is only a detailed description of the preferred embodiments of the present invention, which is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

[0265] It should be understood that, although the various steps in the flow chart of each embodiment of the present invention are shown in sequence according to the indication of the arrows, these steps are not necessarily performed in sequence according to the order indicated by the arrows. Unless otherwise specified herein, the execution of these steps is not strictly limited in order, and these steps can be performed in other orders. Moreover, at least a portion of the steps in each embodiment may include a plurality of sub-steps or a plurality of stages, and these sub-steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these sub-steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with at least a portion of other steps or sub-steps or stages of other steps.

[0266] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing the relevant hardware through a computer program. The program can be stored in a non-volatile computer-readable storage medium. When executed, the program can include the processes of the above-described method embodiments. Any reference to memory, storage, database, or other media used in the various embodiments provided herein may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct RAMbus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM).

[0267] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the disclosure in the specification and examples. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art not disclosed herein. The description and examples are to be considered merely as exemplary, and the true scope and spirit of the present disclosure are indicated by the claims.

Claims

1. A climate adaptive recovery method for container shipping networks that integrates resilience theory, characterized by: The climate-adaptive recovery approach for container shipping networks that integrates resilience theory includes the following steps: Obtain liner schedule data, vessel dynamic data, route adjustment data, port geographic coordinate data, and multi-scale geo-fence data, and introduce time dimension parameters to construct a spatiotemporal weighted container shipping dynamic network; quantifying the strength of cascading failures in the spatiotemporal weighted container shipping dynamic network based on a cascading failure propagation model, and then calculating a comprehensive efficiency index of the spatiotemporal weighted container shipping dynamic network based on transportation timeliness, operational economy, and network robustness; Based on the resilience triangle theory, the resilience index of the spatiotemporal weighted container shipping dynamic network under extreme weather events is quantified; Evaluate the cost-effectiveness of climate suitability improvement strategies for the proposed spatiotemporally weighted container shipping dynamic network; The quantification of the strength of cascading failures in the spatiotemporal weighted container shipping dynamic network based on the cascading failure propagation model further includes: (1) First, after the construction of the spatiotemporal weighted container shipping dynamic network is completed, the logical layer network is further constructed. , where the logical node It is a set of cargo flow paths, representing the transportation path of cargo from the starting point to the end point; logical edge is a set of transport capacity scheduling relationships, representing the transport capacity allocation or scheduling relationship between paths; (2) Secondly, define the bidirectional impact factor between nodes : ； When the physical node When the logical node fails, The load will be Proportional interruption; (3) Again, construct a cascading failure propagation model: First define the initial load: the load of the physical node is the port weekly throughput , the load of the logical node is the capacity demand of the path ; When the node When a node fails, its load is distributed according to the capacity ratio of adjacent nodes: ； in, is a node Newly added loads; is the allocation preference coefficient, when σ>1, the nodes with high degree are preferred; when σ<1, the nodes with low degree are preferred; The capacity transfer efficiency is the ratio of the load of the failed node transferred to the adjacent nodes; Is a failed node The original load; is a node The degree of Power, representing a node influence; is the degree of all adjacent nodes the sum of powers; Wherein, in constructing the cascading failure propagation model, it also includes: Construct the propagation dynamics equation: ； in, is the elasticity coefficient of the physical facility; For physical node p The failure state at the moment; Logical Node In time load; Logical Node In time load; Logical Node In time load; For physical nodes In time load; For physical nodes In time load; (4) Finally, based on the cascading failure propagation model, the strength of the cascading failure of the spatiotemporal weighted container shipping dynamic network is quantified, revealing the cascading effect of the spatiotemporal weighted container shipping dynamic network.

2. The container shipping network climate adaptability recovery method integrating resilience theory according to claim 1 is characterized by: The liner plan data includes the route code, starting and ending ports and the name of the shipping company, and the ship dynamic data includes the latitude and longitude coordinates, speed, heading, current status and next port of call.

3. The container shipping network climate adaptability recovery method integrating resilience theory according to claim 1 is characterized by: The step of constructing a spatiotemporal weighted container shipping dynamic network further includes: (1) First, the traditional topology definition is performed: Node collection: ,in, For the indivual ; Edge collection: ,in, for The starting point, for The end point; Static weight: ; (2) Secondly, expand the time dimension: The first step is to discretize the time layer and divide the continuous time axis into equally spaced time periods: ； in, is the number of days; The second step is to build a three-dimensional space-time network , where each spatiotemporal node is ; The third step is to define the spatiotemporal edge weight as a multidimensional vector : ； in, ; for During the period Actual sailing time; for During the period liner frequency; for During the period Available capacity; (3) Again, perform dynamic parameter quantization: The first step is to calculate During the period Actual sailing time : ； in, For distance, It is the undisturbed speed; for Wind and wave level coefficient during the period; for Actual route cost index during the period; 、 is the calibration parameter, for The maximum route cost index within the time period; The second step is to calculate During the period Frequencies of liner services : ； in, for Annual benchmark flights; is the seasonal fluctuation amplitude; for During the period Temporary increase or decrease of shifts; The third step is to calculate During the period Available capacity : ； in, for Number of ships deployed during the time period; For ships Actual utilization rate of the voyage; For ships Rated load capacity; for Ship schedule reliability coefficient within the time period; Based on the above steps, the time-space weighted container shipping dynamic network is constructed, wherein the time-space weighted container shipping dynamic network is a physical layer network.

4. The method for climate adaptability recovery of container shipping networks integrating resilience theory according to claim 1 is characterized in that: After constructing the spatiotemporal weighted container shipping dynamic network, the method further includes: Construct linear programming models; Performing data cleaning and entity alignment on the spatiotemporal weighted container shipping dynamic network through the linear programming model; Wherein, the construction of the linear programming model further includes: Define the decision variables: ; Objective function: ; Constraints: ; in, Assemble for routes; for gather; For ships Rated load capacity; For routes Total capacity demand; The loading and unloading time for a single vessel is , For quay crane efficiency, is the number of quay cranes; For ships The maximum total available laytime, For ships The maximum number of available shifts.

5. The method for climate adaptability recovery of container shipping network integrating resilience theory according to claim 1 is characterized in that: The comprehensive efficiency index of the time-space weighted container shipping dynamic network calculated based on transportation timeliness, operational economy, and network robustness also includes: The first step is to determine the boundaries of the spatiotemporal weighted container shipping dynamic network, and determine the number of nodes, routes, and cargo flows covered in the evaluation area; The second step is to establish a three-tier indicator system including transportation timeliness, operational economy and network robustness; (1) Transport timeliness indicators include average speed ratio and on-time arrival rate Two sub-indicators: Average speed ratio The calculation formula is: ； in, For routes design speed; For routes Actual average speed; For routes any vessel on board; For routes The collection of ships on board; Assemble for the ship China Shipbuilding the total number of On-time arrival rate The calculation formula is: ； in, For ships Scheduled arrival time; For ships The actual arrival time of mustering for ships; is the allowed time deviation threshold; is an indicator function, which takes 1 when the condition is met and 0 otherwise; (2) Operational economic indicators include unit TEU cost and capacity utilization Two sub-indicators: Unit TEU cost The calculation formula is: ； in, For fuel costs, For ships Total distance traveled, For ships The average speed of is the main engine fuel consumption rate, is the unit price of fuel; in, For port fees, , For the port Basic berthing fee, It is a port The total number of containers, For the port Single box loading and unloading fee, mustering for the port; in, For canal tolls, is the overall length of the ship, For the width of the ship, is the deadweight tonnage; in, For other costs; Capacity Utilization The calculation formula is: ； in, For ships The actual loading box volume, For ships Maximum loading capacity, mustering for ships; (3) Network robustness indicators include node betweenness centrality and edge redundancy index Two sub-indicators: Node betweenness centrality The calculation formula is: ； in, is the total number of shortest paths from node s to node t; is the number of shortest paths passing through node v; where the summation range is for all node pairs sum; Node betweenness centrality Perform normalization: ; Wherein, V is the total number of nodes in the spatiotemporal weighted container shipping dynamic network; is the normalized node betweenness centrality, ranging from [0, 1]; Edge redundancy index The calculation formula is: ； in, For the path collection , the number of paths for which there are alternative paths; is the total number of edges in the spatiotemporal weighted container shipping dynamic network; For the side , the proportion of paths with alternative paths; The definition of alternative paths is: first, for each edge (u→v), find the set of all paths that use this edge ; For each path Then check whether there is an alternative path p' that satisfies: p' does not pass through edge e, and the weight of p' is ≤ 1.2 × the weight of the original path; finally, count the proportion of paths with alternative paths; The third step is to calculate the comprehensive efficiency index of the time-space weighted container shipping dynamic network based on the three-layer index system; wherein the comprehensive efficiency index of the time-space weighted container shipping dynamic network is The calculation formula is: ； in, are weights respectively.

6. The method for climate adaptability recovery of container shipping network integrating resilience theory according to claim 1 is characterized in that: The resilience triangle theory-based quantification of the spatiotemporal weighted container shipping dynamic network resilience index under extreme weather events also includes: The calculation formula of the resilience index Re of the spatiotemporal weighted container shipping dynamic network under extreme weather events is as follows: ； Where, between 0 and 1; represents the starting time when the spatiotemporal weighted container shipping dynamic network is affected by extreme weather events, represents the time when the spatiotemporal weighted container shipping dynamic network returns to its original state, Indicates time Global network efficiency; represents the initial network efficiency; Among them, if , indicating that the spatiotemporal weighted container shipping dynamic network is not affected by extreme weather events, and the performance of the spatiotemporal weighted container shipping dynamic network remains original; Among them, if , indicating that the extreme weather event has a negative impact on the spatiotemporal weighted container shipping dynamic network, and a recovery strategy needs to be adopted to restore the performance of the spatiotemporal weighted container shipping dynamic network; Among them, if , indicating that the spatiotemporal weighted container shipping dynamic network is severely damaged by the destructive event and will take a long time to recover.

7. The container shipping network climate adaptability recovery method integrating resilience theory according to claim 1 is characterized by: The evaluation of the cost-effectiveness of the climate suitability improvement strategy for the spatiotemporal weighted dynamic container shipping network also includes: Construct an improved greedy algorithm based on a multi-constrained marginal cost optimization model; Evaluating the cost-effectiveness of different climate suitability improvement strategies for the spatiotemporal weighted container shipping dynamic network based on the improved greedy algorithm to determine the climate suitability improvement strategy with the best cost-effectiveness; A climate adaptive recovery strategy for the spatiotemporal weighted container shipping dynamic network is formulated based on a dynamic adaptive policy path enhancement framework to ensure the rapid recovery of the spatiotemporal weighted container shipping dynamic network under extreme weather events.

8. The method for climate adaptability recovery of container shipping network integrating resilience theory according to claim 7 is characterized in that: The improved greedy algorithm based on the multi-constrained marginal cost optimization model further includes: First, define a four-dimensional cost vector: C = [economic cost, carbon emissions, recovery time, climate vulnerability]; Secondly, construct the multi-constrained marginal cost optimization model : ； in, For the the implementation intensity of various climate suitability improvement strategies; For the port Climate Vulnerability Index; is the climate risk sensitivity coefficient; the economic costs of implementing climate suitability improvement strategies; Carbon emissions resulting from the implementation of climate suitability improvement strategies; The recovery time of ports or shipping routes after climate disasters; is the climate vulnerability threshold; 、 、 is the weight coefficient; To strengthen the implementation of the strategy Find partial derivatives; The multi-constrained marginal cost optimization model is dynamically updated using the NSGA-II multi-objective optimization framework. The weight coefficient in 、 、 .

9. The method for climate adaptability recovery of container shipping network integrating resilience theory according to claim 7 is characterized in that: The dynamic adaptive policy path enhancement framework specifically includes: (1) Constructing a system for quantifying system failure signals: ATP calculation model: ； in, For the port Real-time wind speed; For the port The highest wind speed in history; For the port cumulative rainfall; port The maximum cumulative rainfall in history; For the port hinterland economic equivalent; is the sum of the economic equivalents of all port hinterlands; is the total number of ports; (2) Strategy tree construction based on deep reinforcement learning: State space: ATP level, network connectivity, and ship availability; Action space: {reroute, temporary call, activation of alternate port, initiation of land transport connection}; Reward function: ; in, and is the weight coefficient; Cost changes after implementation of climate-resilient restoration strategies; Temporal changes after implementation of climate-adaptive restoration strategies; is the reward coefficient; The restoration benefits after the implementation of climate-adaptive restoration strategies.

Images