Sea route planning method considering uncertainty factors

Abstract

The invention discloses a marine route planning method considering uncertain factors, and belongs to the field of marine traffic safety management, and the method comprises the steps: S1, carrying out the preprocessing and feature fusion of multi-source data; s2, carrying out the uncertainty modeling of the marine route, and quantifying the probability distribution of uncertainty parameters; s3, predicting a candidate route trajectory based on PCA hierarchical attention; s4, in combination with the probability distribution of the uncertainty parameters and the predicted candidate route trajectory, calculating the operability super probability of each segment of the route, and generating a probability operability index; and S5, by adopting an approximate MinSumA algorithm, outputting an optimal route by taking the integration of the minimum total navigation time and the maximum probability operability index as targets. By the adoption of the sea route planning method considering the uncertainty factors, a data-modeling-prediction-evaluation-optimization closed loop is constructed, the problems that a traditional method ignores uncertainty, is high in optimization complexity and poor in real-time performance are solved, and safety-efficiency-real-time performance collaborative optimization is achieved.

Description

Technical Field

[0001] This invention relates to the field of maritime traffic safety management technology, and in particular to a method for planning maritime routes that takes into account uncertainties. Background Technology

[0002] As a core carrier of global trade, maritime shipping handles over 80% of global cargo volume, and its efficiency and safety directly impact the stability of the global supply chain. With the acceleration of maritime intelligence and the widespread adoption of Automatic Identification System (AIS) data, satellite remote sensing sea state data, and ship motion modeling technology, maritime route planning has shifted from traditional "experience-driven" to "data-driven." The core objective is gradually focusing on a multi-objective balance of "safety, efficiency, and low carbon"—avoiding ship motion risks caused by complex sea conditions (such as waves and ocean currents) (such as excessive rolling and collisions), while optimizing sailing time and fuel consumption to comply with the latest International Maritime Organization (IMO) standards on ship energy efficiency (EEDI) and navigation safety.

[0003] Currently, a preliminary system of technologies related to maritime route planning has been established: trajectory prediction technology based on AIS data can predict vessel positions in advance, providing a basis for collision avoidance and path adjustment; vessel maneuverability analysis technology defines safe navigation boundaries by quantifying indicators such as vertical acceleration and pitch; and traffic allocation optimization algorithms are used for vessel flow scheduling between multiple routes to avoid local congestion. However, existing technologies still have significant shortcomings in adaptability to complex sea conditions, accuracy of multi-source data fusion, and real-time optimization capabilities, making it difficult to meet the route planning needs of complex sea areas such as near-shore wind farms and inland river estuaries. Summary of the Invention

[0004] The purpose of this invention is to provide a maritime route planning method that takes into account uncertainties, thereby solving the aforementioned technical problems.

[0005] To achieve the above objectives, the present invention provides a method for maritime route planning that considers uncertainties, comprising the following steps: S1. Multi-source data preprocessing and feature fusion: Integrate AIS data, sea state data and ship technical data, and generate key feature matrices by dimensionality reduction and redundancy removal through PCA-iPCA; S2. Uncertainty Modeling of Maritime Routes: Considering the errors in ship motion calculations and the ambiguity of operability limits, frequency-independent model errors and Gaussian distributions are used for modeling to quantify the probability distribution of uncertainty parameters. S3. Ship trajectory prediction based on PCA hierarchical attention: Based on the key feature matrix generated in step S1, a PCA hierarchical attention encoder-decoder model is adopted. The dynamic dependence of position, speed and heading is learned through three independent LSTM branches. The attention mechanism is combined to focus on key time steps to predict candidate route trajectories. S4. Probabilistic operability assessment: Combining the probability distribution of the uncertainty parameters obtained in step S2 and the candidate route trajectory predicted in step S3, the operability hyperprobability of each segment of the route is calculated based on the second-order reliability method, and a probabilistic operability index is generated. S5. Route optimization based on MinSumA: The approximate MinSumA algorithm is adopted to minimize the total flight time and maximize the probability operability index, and output the optimal route.

[0006] Therefore, the above-mentioned method for maritime route planning that considers uncertainties has the following beneficial effects: 1. Significantly improves route safety and avoids safety misjudgments: For the first time, two types of cognitive uncertainty are quantitatively modeled: the FIME error coefficient (normal distribution) is used to correct the calculation deviation of ship motion response, and the Gaussian distribution is used to characterize the ambiguity of operability limits, solving the "black and white" defect of traditional deterministic assessment; at the same time, based on SORM, the probability of exceeding the limit for each segment of the route is calculated and POI (Probability Operability Index) is generated to quantify the overall safety level (setting POI≥95%), which can identify high-risk sea state segments in advance and reduce the risk of accidents such as ship roll exceeding the limit and collisions; 2. Balancing navigation efficiency and real-time optimization to reduce operating costs: Based on PCA hierarchical attention trajectory prediction, accurate route length is obtained. Combined with the MinSumA optimization algorithm, the total navigation time is reduced by ≥10%, and the average navigation time difference between ships on each route is ≤20%, avoiding excessive congestion on a single route. At the same time, MinSumA uses approximate derivatives to replace traditional trial transfer calculations, which improves the speed by 2 to 170 times compared to the original MinSum algorithm. It is compatible with shipborne embedded systems and can meet the real-time scheduling needs of scenarios such as offshore wind farm operation and maintenance and inland waterway shipping.

[0007] 3. Multi-technology integration enhances technological advancement and breaks through the limitations of traditional methods: PCA-iPCA feature fusion reduces feature redundancy (redundancy reduction rate ≥40%) while retaining 90% of the information, avoiding noise interference from multi-source data (AIS + sea state + ship technical data), and providing high-quality input for subsequent modeling; at the same time, PCA hierarchical attention learns the dynamic dependence of position, velocity, and heading through three independent LSTM branches, and focuses key information by combining intra-branch / inter-branch attention, resulting in prediction accuracy superior to traditional single LSTM (average error reduction of more than 30%), providing a reliable trajectory basis for probability assessment; moreover, POI is used as a safety constraint for MinSumA optimization, forming a linkage of "prediction accuracy → risk quantification → efficiency optimization", avoiding the one-sidedness of single-objective optimization (such as only pursuing time while ignoring safety).

[0008] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0009] Figure 1 This is a flowchart of a maritime route planning method that takes into account uncertainties according to the present invention. Detailed Implementation

[0010] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely illustrative of the embodiments of the present invention and are not intended to limit the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of this application. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout.

[0011] It should be noted that the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion, such as a process, method, system, product, or server that includes a series of steps or units, not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such process, method, product, or device.

[0012] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0013] like Figure 1 As shown, a method for planning maritime routes that considers uncertainties includes the following steps: S1. Multi-source data preprocessing and feature fusion: Integrate AIS data, sea state data and ship technical data, and generate key feature matrices by dimensionality reduction and redundancy removal through PCA-iPCA; Step S1 specifically includes the following steps: S11. Synchronously acquire AIS data Sea state data and ship technical data The original multi-source dataset is obtained. ,in, These represent the latitude, longitude, ground speed, ground heading, and timestamp of the ship's location, respectively. These represent wave height, peak period, and wave direction, respectively. These represent the ship's length, beam, draft, block coefficient, and natural roll period, respectively. S12. For the original multi-source dataset Outlier removal, time synchronization, and standardization are performed sequentially to obtain a standardized multi-source dataset. ; S13, PCA-iPCA Feature Fusion: Utilizing Standardized Multi-Source Datasets Constructing a feature matrix from features Then, PCA-iPCA is used for feature fusion to obtain the key feature matrix. .

[0014] S2. Uncertainty Modeling of Maritime Routes: Considering the errors in ship motion calculations and the ambiguity of operability limits, frequency-independent model errors and Gaussian distributions are used for modeling to quantify the probability distribution of uncertainty parameters. Step S2 specifically includes the following steps: S21. Calculate the error coefficient of the frequency-independent model by correlating the ship's main parameters with its navigation status through linear regression. : ; In the formula, , , , These represent the regression intercept coefficient, wave direction angle regression coefficient, square coefficient regression coefficient, and Froude number regression coefficient, respectively. The heave error coefficient is calculated from these coefficients. hour, Calculate the pitch error coefficient hour, ; Indicates the angle between the ship and the wave direction, and ; Represent the Fruder number, and , Indicates the actual speed of the ship. Represents gravitational acceleration; S22. Based on model test data, the frequency-independent model error coefficients The model is based on a normal distribution, where the sag error coefficient is... , and These represent the mean and standard deviation of the heave error coefficient, respectively; and the pitch error coefficient. , and These represent the mean and standard deviation of the pitch error coefficient, respectively; the roll error coefficient... , and These represent the mean and standard deviation of the roll error coefficient, respectively. S23. Determine the mean of the operability limits: pitch limit Vertical acceleration limit Roll limit Collision probability limit Propeller water exit probability limit ; S24. Considering the differences in captains' experience and the ambiguity of standards, the limit values ​​follow a Gaussian distribution. , Represents the coefficient of variation. Indicates the limit value, and .

[0015] S3. Ship trajectory prediction based on PCA hierarchical attention: Based on the key feature matrix generated in step S1, a PCA hierarchical attention encoder-decoder model is adopted. The dynamic dependence of position, speed and heading is learned through three independent LSTM branches. The attention mechanism is combined to focus on key time steps to predict candidate route trajectories. Step S3 specifically includes the following steps: S31. Define the input branch: Spatial feature branch Velocity characteristic branch , heading characteristic branch , These represent the latitude and longitude of the ship's relative position, respectively. Indicates the standard ground velocity characteristics. Indicates the standard ground heading characteristics; S32, Utilizing Spatial Features in Branching Velocity characteristic branch and heading feature branch Independently train an LSTM encoder to convert the input temporal features into hidden state sequences. : ; In the formula, Indicates branch The corresponding LSTM encoder, Indicates the branch identifier. ; Indicates branch Input temporal characteristics; Indicates branch The initial hidden state sequence; S33. Use Luong dot product attention to calculate the weights of the hidden state sequence at each time step for each branch: ; in, ; In the formula, Indicates branch At time step Attention weights , Representing spatial feature branches respectively Velocity characteristic branch and heading feature branch Attention weights; Indicates branch At time step The hidden state; Indicates branch The final time step hidden state; Indicates the transpose operation; Indicates the total number of time steps; S34, Calculation Branch context vector ; and , These are context vectors representing position, velocity, and heading, respectively. Fusion attention weights and Obtain the comprehensive context vector : ; S35. Employ an LSTM decoder to synthesize the context vector. Given the initial state, predict the future position of the ship: , Linear ; In the formula, and These represent the LSTM decoder at the 1st... The and the first Hidden states for each prediction step; Indicates an LSTM encoder; Indicates the first The relative positions of each prediction step; Linear Indicates a fully connected layer; Indicates the first The relative positions of each prediction step, and , These represent the latitude and longitude of the predicted relative position, respectively. S36. Absolute position transformation: , ; In the formula, and These represent the latitude and longitude of the predicted absolute location, respectively. and These represent the latitude and longitude of the absolute location of the starting point of the flight route, respectively. Thus, the predicted candidate flight path trajectory is obtained. : ; In the formula, This indicates the prediction time step.

[0016] Step S36 is followed by S37, which calculates the distance error between the predicted location and the historical actual location using the Haversine formula to verify the accuracy.

[0017] S4. Probabilistic operability assessment: Combining the probability distribution of the uncertainty parameters obtained in step S2 and the candidate route trajectory predicted in step S3, the operability hyperprobability of each segment of the route is calculated based on the second-order reliability method, and a probabilistic operability index is generated. Step S4 specifically includes the following steps: S41. Calculate the ship's motion response by combining the probability distribution of the uncertainty parameters and the predicted candidate trajectories: ; ; ; in, ; In the formula, , and These represent the corrected heave, pitch, and roll responses, respectively. , and Let represent the transfer functions for heave, pitch, and roll, respectively. Indicates encountering a spectrum; Indicates the frequency of the spectral peak, and ; Indicates the frequency of encounters, and , Represents the circular frequency of ocean waves. Indicates wave number; Indicates the shape parameter of the encountered spectrum; S42. Define the following limit state function: Pitch safety function : ; Vertical acceleration safety function : ; Roll safety function : ; Collision probability safety function : ; Propeller water exit probability safety function : ; In the formula, Indicates the distance from the bow to the center of gravity; Indicates the vertical acceleration at the bow; Indicates effective draft; Indicates the critical vertical velocity; Represents the relative velocity variance; Represents the variance of relative motion; Indicates the depth of the propeller blade tip; S43. Standardize the uncertainty parameters to obtain the standardized uncertainty parameter variables. Then, the probability of exceeding the standard is calculated using the second-order reliability index. : ; In the formula, This represents the cumulative function of the standard normal distribution. This represents a second-order reliability index; Indicates the principal curvature of the limit state surface at the design point; The dimension representing the uncertainty parameter; S44. Calculate the probability operability index based on historical sea state data. : ; In the formula, Indicates the frequency of sea state types; Indicates sea state The probability of achieving the required operability.

[0018] S5. Route optimization based on MinSumA: The approximate MinSumA algorithm is adopted to minimize the total flight time and maximize the probability operability index, and output the optimal route.

[0019] Step S5 specifically includes the following steps: S51, Predicting candidate flight paths The coordinates of the prediction step are extracted as waypoints, and the route segment formed by two adjacent waypoints is verified to meet the no-navigation zone constraint, so as to obtain the predicted route under various sea conditions. S52. Calculate the total length of the predicted route generated in step S51 using the Haversine formula. : ; In the formula, Indicates the average radius of the Earth; Indicates the predicted flight path The Middle The waypoint and the first The latitude difference between each waypoint, and , and They represent the first The first waypoint and the first The latitude of each waypoint; Indicates the predicted flight path The Middle The waypoint and the first The longitude difference between each waypoint, and , and They represent the first The first waypoint and the first Longitude of each waypoint; S53, Calculate the initial sailing time : ; In the formula, Indicates the average speed of the ship; S54. Construct a multi-objective function that minimizes the total flight time and maximizes the probabilistic operability index. : ; In the formula, Indicates the sailing time, and , Indicates the congestion coefficient; Indicates the predicted flight path The number of ships; This indicates the total number of predicted flight routes; Indicates the penalty weight; Indicates the predicted flight path The probability operability index; S55. Calculate the derivative of a single route: ; S56, Calculate donor prediction trajectory Predicted route to receptor Change in the transfer objective function: ; In the formula, Indicates receptor-predicted flight path The number of ships; S57. Traverse all donor-recipient pairs and combine them with predicted candidate flight paths. Valid pairs are selected if the donor predicts the flight path. of Display the number of remaining ships. Or receptor predicts flight path of Display the number of ships after the transfer. If the trajectory deviates into a no-fly zone, then the donor-recipient pair is excluded, and a valid donor-recipient pair is obtained; among them, Indicates the number of ships transferred; S58, Filtering The smallest effective donor-recipient pair is considered the optimal route.

[0020] Step S5 is followed by S6, which verifies the prediction error, probabilistic operability index and time cost of the optimal route based on historical route data, and outputs a visualized route and evaluation report.

[0021] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for planning maritime routes that considers uncertainties, characterized in that: Includes the following steps: S1. Multi-source data preprocessing and feature fusion: Integrate AIS data, sea state data and ship technical data, and generate key feature matrices by dimensionality reduction and redundancy removal through PCA-iPCA; S2. Uncertainty Modeling of Maritime Routes: Considering the errors in ship motion calculations and the ambiguity of operability limits, frequency-independent model errors and Gaussian distributions are used for modeling to quantify the probability distribution of uncertainty parameters. S3. Ship trajectory prediction based on PCA hierarchical attention: Based on the key feature matrix generated in step S1, a PCA hierarchical attention encoder-decoder model is adopted. The dynamic dependence of position, speed and heading is learned through three independent LSTM branches. The attention mechanism is combined to focus on key time steps to predict candidate route trajectories. S4. Probabilistic operability assessment: Combining the probability distribution of the uncertainty parameters obtained in step S2 and the candidate route trajectory predicted in step S3, the operability hyperprobability of each segment of the route is calculated based on the second-order reliability method, and a probabilistic operability index is generated. S5. Route optimization based on MinSumA: The approximate MinSumA algorithm is adopted to minimize the total flight time and maximize the probability operability index, and output the optimal route.

2. The maritime route planning method considering uncertainties according to claim 1, characterized in that: Step S1 specifically includes the following steps: S11. Synchronously acquire AIS data Sea state data and ship technical data The original multi-source dataset is obtained. ,in, These represent the latitude, longitude, ground speed, ground heading, and timestamp of the ship's location, respectively. These represent wave height, peak period, and wave direction, respectively. These represent the ship's length, beam, draft, block coefficient, and natural roll period, respectively. S12. For the original multi-source dataset Outlier removal, time synchronization, and standardization are performed sequentially to obtain a standardized multi-source dataset. ; S13, PCA-iPCA Feature Fusion: Utilizing Standardized Multi-Source Datasets Constructing a feature matrix from the features in Then, PCA-iPCA is used for feature fusion to obtain the key feature matrix. .

3. The maritime route planning method considering uncertainties according to claim 2, characterized in that: Step S2 specifically includes the following steps: S21. Calculate the error coefficient of the frequency-independent model by correlating the ship's main parameters with its navigation status through linear regression. : ； In the formula, , , , These represent the regression intercept coefficient, wave direction angle regression coefficient, square coefficient regression coefficient, and Froude number regression coefficient, respectively. The heave error coefficient is calculated from these coefficients. hour, Calculate the pitch error coefficient hour, ; Indicates the angle between the ship and the wave direction, and ; Represent the Fruder number, and , Indicates the actual speed of the ship. Represents gravitational acceleration; S22. Based on model test data, the frequency-independent model error coefficients The model is based on a normal distribution, where the sag error coefficient is... , and These represent the mean and standard deviation of the heave error coefficient, respectively; and the pitch error coefficient. , and These represent the mean and standard deviation of the pitch error coefficient, respectively; the roll error coefficient... , and These represent the mean and standard deviation of the roll error coefficient, respectively. S23. Determine the mean of the operability limits: pitch limit Vertical acceleration limit Roll limit Collision probability limit Propeller water exit probability limit ; S24. Considering the differences in captains' experience and the ambiguity of the standard, the limit value follows a Gaussian distribution. , Represents the coefficient of variation. Indicates the limit value, and .

4. The maritime route planning method considering uncertainties according to claim 3, characterized in that: Step S3 specifically includes the following steps: S31. Define the input branch: Spatial feature branch Velocity characteristic branch , heading characteristic branch , These represent the latitude and longitude of the ship's relative position, respectively. Indicates the standard ground velocity characteristics. Indicates the standard ground heading characteristics; S32, Utilizing Spatial Features in Branching Velocity characteristic branch and heading feature branch Independently train an LSTM encoder to convert the input temporal features into hidden state sequences. : ； In the formula, Indicates branch The corresponding LSTM encoder, Indicates the branch identifier. ; Indicates branch Input temporal characteristics; Indicates branch The initial hidden state sequence; S33. Use Luong dot product attention to calculate the weights of the hidden state sequence at each time step for each branch: ； in, ； In the formula, Indicates branch At time step Attention weights , Representing spatial feature branches respectively Velocity characteristic branch and heading feature branch Attention weights; Indicates branch At time step The hidden state; Indicates branch The final time step hidden state; Indicates the transpose operation; Indicates the total number of time steps; S34, Calculation Branch context vector ； and , These represent context vectors for position, velocity, and heading, respectively. Fusion attention weights and Obtain the comprehensive context vector : ； S35. Employ an LSTM decoder to synthesize the context vector. Given the initial state, predict the future position of the ship: , Linear ； In the formula, and These represent the LSTM decoder at the 1st... The and the first Hidden states for each prediction step; Indicates an LSTM encoder; Indicates the first The relative positions of each prediction step; Linear Indicates a fully connected layer; Indicates the first The relative positions of each prediction step, and , These represent the latitude and longitude of the predicted relative position, respectively. S36. Absolute position transformation: ， ； In the formula, and These represent the latitude and longitude of the predicted absolute location, respectively. and These represent the latitude and longitude of the absolute location of the starting point of the flight route, respectively. Thus, the predicted candidate flight path trajectory is obtained. : ； In the formula, This indicates the prediction time step.

5. The maritime route planning method considering uncertainties according to claim 4, characterized in that: Step S36 is followed by S37, which calculates the distance error between the predicted location and the historical actual location using the Haversine formula to verify the accuracy.

6. The maritime route planning method considering uncertainties according to claim 5, characterized in that: Step S4 specifically includes the following steps: S41. Calculate the ship's motion response by combining the probability distribution of the uncertainty parameters and the predicted candidate trajectories: ； ； ； in, ； In the formula, , and These represent the corrected heave, pitch, and roll responses, respectively. , and Let represent the transfer functions for heave, pitch, and roll, respectively; Indicates encountering a spectrum; Indicates the frequency of the spectral peak, and ; Indicates the frequency of encounters, and , Represents the circular frequency of ocean waves. Indicates wave number; Indicates the shape parameter of the encountered spectrum; S42. Define the following limit state function: Pitch safety function : ; Vertical acceleration safety function : ; Roll safety function : ; Collision probability safety function : ; Propeller water exit probability safety function : ; In the formula, Indicates the distance from the bow to the center of gravity; Indicates the vertical acceleration at the bow; Indicates effective draft; Indicates the critical vertical velocity; Represents the relative velocity variance; Represents the variance of relative motion; Indicates the depth of the propeller blade tip; S43. Standardize the uncertainty parameters to obtain the standardized uncertainty parameter variables. Then, the probability of exceeding the standard is calculated using the second-order reliability index. : ； In the formula, This represents the cumulative function of the standard normal distribution. This represents a second-order reliability index; Indicates the principal curvature of the limit state surface at the design point; The dimension representing the uncertainty parameter; S44. Calculate the probability operability index based on historical sea state data. : ； In the formula, Indicates the frequency of sea state types; Indicates sea state The probability of achieving the required operability.

7. A maritime route planning method considering uncertainties according to claim 6, characterized in that: Step S5 specifically includes the following steps: S51, Predicting candidate flight paths The coordinates of the prediction step are extracted as waypoints, and the route segment formed by two adjacent waypoints is verified to meet the no-navigation zone constraint, so as to obtain the predicted route under various sea conditions. S52. Calculate the total length of the predicted route generated in step S51 using the Haversine formula. : ； In the formula, Indicates the average radius of the Earth; Indicates the predicted flight path The Middle The waypoint and the first The latitude difference between each waypoint, and , and They represent the first The first waypoint and the first The latitude of each waypoint; Indicates the predicted flight path The Middle The waypoint and the first The longitude difference between each waypoint, and , and They represent the first The first waypoint and the first Longitude of each waypoint; S53, Calculate the initial sailing time : ； In the formula, Indicates the average speed of the ship; S54. Construct a multi-objective function that minimizes the total flight time and maximizes the probabilistic operability index. : ； In the formula, Indicates the sailing time, and , Indicates the congestion coefficient; Indicates the predicted flight path The number of ships; This indicates the total number of predicted flight routes; Indicates the penalty weight; Indicates the predicted flight path The probability operability index; S55. Calculate the derivative of a single route: ； S56, Calculate donor prediction trajectory Predicted route to receptor Change in the transfer objective function: ； In the formula, Indicates receptor-predicted flight path The number of ships; S57. Traverse all donor-recipient pairs and combine them with predicted candidate flight paths. Valid pairs are selected if the donor predicts the flight path. of Display the number of remaining ships. Or receptor predicts flight path of Display the number of ships after the transfer. If the trajectory deviates into a no-fly zone, then the donor-recipient pair is excluded, and a valid donor-recipient pair is obtained; among them, Indicates the number of ships transferred; S58, Filtering The smallest effective donor-recipient pair is considered the optimal route.

8. A maritime route planning method considering uncertainties according to claim 7, characterized in that: Step S5 is followed by S6, which verifies the prediction error, probabilistic operability index and time cost of the optimal route based on historical route data, and outputs a visualized route and evaluation report.

Images