A container route analysis method based on artificial intelligence analysis

Abstract

This invention discloses a container shipping route analysis method based on artificial intelligence analysis, belonging to the field of shipping intelligence technology. The method includes: collecting route operation data during the operation and management of a target container shipping route, performing time alignment and normalization processing to generate an actual route event sequence; embedding route governance rules into corresponding route segments based on the actual route event sequence to construct a route governance semantic representation; outputting the arrival time series of each port of call and the reliability evolution characteristics as the route progresses through an artificial intelligence analysis process based on the route governance semantic representation; and generating a counterfactual route trajectory sequence using the arrival time series and reliability evolution characteristics. This invention reconstructs the theoretical evolution process of a shipping route under given operational constraints by constructing a counterfactual route trajectory with route governance semantics, providing a predictable and alignable expressive basis for the operational state of the route in both time and space dimensions.

Description

Technical Field

[0001] This invention relates to the field of shipping intelligence technology, and in particular to a container shipping route analysis method based on artificial intelligence analysis. Background Technology

[0002] With the continuous expansion of the global shipping network and the increasing complexity of shipping operations, container shipping route management is gradually evolving from traditional experience-based scheduling to data-driven and intelligent analysis. In existing shipping management systems, the dimensions of route operation data collection are constantly being enriched, covering multi-source information such as speed, location, voyage plans, port operation status, weather conditions, and costs and carbon emissions. Artificial intelligence methods are also being gradually introduced to model and evaluate route operation status. Based on this, an analytical framework centered on route event sequences, trajectory analysis, and operational indicator calculation has been formed, focusing on status identification, operational feature extraction, and route performance evaluation during route operations. This framework provides data support for route scheduling optimization and operational decision-making.

[0003] Existing technologies for route operation analysis typically focus on statistical analysis of single indicators or localized behaviors, lacking a systematic characterization of multidimensional constraints, temporal evolution relationships, and potential deviations throughout the entire route operation process. In particular, they struggle to establish a unified semantic correlation mechanism between route governance rules, operational behavior evolution, and outcome impacts. This leads to route anomaly identification and adjustment decisions often relying on experience-based judgments, lacking interpretability and quantitative evidence. When deviations occur in route operations, it is difficult to accurately characterize the underlying causes of these deviations and their impact on overall route reliability, thus limiting the precision and reliability of route optimization and replanning decisions. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a container route analysis method based on artificial intelligence analysis to solve the problem of difficulty in unified modeling and quantitative analysis of route governance constraints, operational status evolution and abnormal behavior impacts during route operation.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: This invention provides a container shipping route analysis method based on artificial intelligence analysis. The method includes: collecting route operation data during the operation and management of a target container shipping route, performing time alignment and normalization processing to generate an actual route event sequence; embedding route governance rules into corresponding route segments based on the actual route event sequence to construct a route governance semantic representation; outputting the arrival time sequence of each port of call and the reliability evolution characteristics as the route progresses through artificial intelligence analysis based on the route governance semantic representation, and generating a counterfactual route trajectory sequence using the arrival time sequence and reliability evolution characteristics; performing spatiotemporal alignment and difference analysis on the route segments to identify abnormal route behavior by reconstructing the actual route trajectory sequence and the counterfactual route trajectory sequence from the actual route event sequence; performing differential calculations based on the abnormal route behavior to form impact quantification information; and generating route adjustment and replanning suggestions based on the arrival time sequence, reliability evolution characteristics, abnormal route behavior, and impact quantification information, constructing corresponding interpretable route scores and counterfactual difference evidence chains, and outputting the analysis content.

[0007] As a preferred embodiment of the container shipping route analysis method based on artificial intelligence analysis described in this invention, the route operation data includes trajectory data, voyage plan data, port berthing and departure and operation status data, waterway passage status data, meteorological and sea condition data, fuel consumption data, and cost and carbon emission data.

[0008] As a preferred embodiment of the container route analysis method based on artificial intelligence analysis described in this invention, the specific steps for generating the actual route event sequence are as follows: The route operation data is timestamped and normalized, and speed information, position coordinate information and operation status identification information are extracted from the route operation data to form a route trajectory point sequence. Based on the sequence of route trajectory points, the operation status switching nodes are identified, and the route operation process is divided with adjacent switching nodes as boundaries to generate the actual route event sequence.

[0009] As a preferred embodiment of the container route analysis method based on artificial intelligence analysis described in this invention, the route governance rules include on-time requirements, route compliance requirements, cost budget constraints, and carbon emission constraints.

[0010] As a preferred embodiment of the container route analysis method based on artificial intelligence analysis described in this invention, the specific steps for constructing the route governance semantic representation are as follows: For each event segment in the actual route event sequence, features such as arrival time deviation, flight path deviation, port of call change, unit voyage cost deviation, and unit voyage carbon emission deviation are extracted, and the embedding strength of the route governance rules on the corresponding route event segments is calculated. The embedding strength is jointly processed with the speed information, position coordinate information and operation status identification information of the corresponding route event fragment, and combined according to the time sequence of route operation to generate a semantic representation of route governance.

[0011] As a preferred embodiment of the container route analysis method based on artificial intelligence analysis described in this invention, the specific steps for generating the counterfactual route trajectory sequence are as follows: Based on the semantic representation of route governance, the governance semantic features of each route event segment are read in the time sequence of route operation and used as the state input in the route advancement process to generate a semantic state sequence. Based on the semantic state sequence, the arrival time of the route at each port of call is extrapolated port by port. At the same time, the timeliness stability and fluctuation of the route are evaluated during the route's progress, forming the arrival time sequence and reliability evolution characteristics. Using arrival time series and reliability evolution characteristics as constraint information, the route operation process is retrospectively reconstructed, and the temporal and spatial positions of each route event segment are adjusted to generate a counterfactual route trajectory sequence.

[0012] As a preferred embodiment of the container route analysis method based on artificial intelligence analysis described in this invention, the specific process for identifying abnormal route behavior is as follows: Based on the event segments of each actual flight route in the actual flight route event sequence, the actual flight route trajectory sequence is reconstructed. Trajectory data is extracted from the actual flight path sequence and the counterfactual flight path sequence to form actual flight path segments and counterfactual flight path segments; Using the counterfactual flight path segment as the time reference, the corresponding actual flight path segment is subjected to time alignment and spatial resampling to generate aligned trajectory segment pairs; Based on the aligned trajectory segment pairs, the differences between the actual route trajectory segment and the counterfactual route trajectory segment in terms of arrival time deviation, navigation path offset, operating speed change and operating status switching frequency are calculated to form difference description information. The difference description information is continuously analyzed according to the flight route sequence. When the difference of one flight route segment is continuously amplified and abruptly changes, it is judged as abnormal flight route behavior.

[0013] As a preferred embodiment of the container route analysis method based on artificial intelligence analysis described in this invention, the specific process of performing differential calculations based on abnormal route behavior to form impact quantification information is as follows. Within the time range corresponding to the abnormal route behavior, differential calculations are performed on the arrival time, sailing distance, energy consumption per unit distance, and duration of operation status of the actual route trajectory segment and the counterfactual route trajectory segment to form a multidimensional differential feature set. Based on a multidimensional differential feature set, the collaborative changes of each differential feature within the same abnormal flight path behavior are integrated and processed to generate quantitative descriptive information; The quantitative description information is associated with the corresponding abnormal flight path behavior and recorded to form the impact quantitative information.

[0014] As a preferred embodiment of the container route analysis method based on artificial intelligence analysis described in this invention, the specific process for generating route adjustment and replanning suggestions is as follows: Based on arrival time series and reliability evolution characteristics, a comprehensive assessment of the timeliness stability and risk change trends of the route in each segment is conducted to identify the route segments that need to be adjusted. For each route segment, by combining abnormal route behavior and corresponding quantitative information on its impact, we analyze the specific ways in which abnormal behavior affects the route's timeliness, operating costs and energy consumption, and form a set of adjustable factors. Based on the set of adjustable factors, the route operating parameters, call order and operating strategies are combined and simulated, and the feasibility and expected effects are evaluated to form recommendations for route adjustment and replanning.

[0015] As a preferred embodiment of the container route analysis method based on artificial intelligence analysis described in this invention, the specific process for outputting the analysis content is as follows: Based on arrival time series, reliability evolution characteristics, abnormal route behavior and impact quantification information, a set of scoring elements is generated by route segment, and the contribution relationship of each scoring element within the route segment is attributed and organized to generate an interpretable route score. Based on the actual route trajectory sequence and the counterfactual route trajectory sequence, the evidence of changes in arrival time deviation, navigation path deviation, operating speed change and operating status switching within the route segment is organized in a way to generate a counterfactual difference evidence chain. The explainable route scores, counterfactual discrepancy evidence chains, and route adjustment and replanning recommendations are linked and encapsulated to form the analysis content.

[0016] The beneficial effects of this invention are as follows: by constructing counterfactual route trajectories with semantics of route governance, the theoretical evolution process of routes under given operational constraints is reconstructed, providing a predictable and alignable basis for the operational status of routes in both time and space dimensions; by performing segment-by-segment alignment analysis between counterfactual route trajectories and actual route trajectories, the sources of deviations caused by timeliness fluctuations, path deviations, or changes in operational status during route operation can be accurately characterized, transforming route anomalies that were originally difficult to quantify into calculable and traceable difference features, thus completing the structured expression and quantitative description of the operational status of routes. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart of a container shipping route analysis method based on artificial intelligence.

[0019] Figure 2 A flowchart for constructing semantics for route governance.

[0020] Figure 3 A flowchart for generating counterfactual flight paths.

[0021] Figure 4 A flowchart for evaluating route adjustments. Detailed Implementation

[0022] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0023] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0024] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0025] Reference Figures 1-4 This is one embodiment of the present invention, which provides a container shipping route analysis method based on artificial intelligence analysis, including the following steps: S1: Collect route operation data of the target container route during the operation and management process, and perform time alignment and normalization processing to generate the actual route event sequence.

[0026] S1.1: Perform timestamp alignment and field normalization on the route operation data, and extract air speed information, position coordinate information and operation status identification information from the route operation data to form a route trajectory point sequence.

[0027] Specifically, the timestamp field in the route operation data is uniformly organized, and route operation data from different sources or different sampling frequencies are mapped to a unified time base; when the timestamps of adjacent data records are inconsistent, the speed information, position coordinate information and operation status identification information of the intermediate time point are supplemented by time interpolation or missing filling methods, so as to keep various types of route operation data aligned on the same time series.

[0028] The field names, units of measurement, and value formats in the aligned route operation data are standardized. Speed ​​information, position coordinate information, and operation status identifier information are extracted from the standardized route operation data in chronological order and combined with the corresponding timestamps to form a sequence of route trajectory points arranged in chronological order.

[0029] S1.2: Identify the operation status switching nodes based on the route trajectory point sequence, divide the route operation process with adjacent switching nodes as the boundary, and generate the actual route event sequence.

[0030] Specifically, in the sequence of route trajectory points, the operation status identification information corresponding to each trajectory point is read in chronological order, and the operation status identification information of adjacent trajectory points is compared; when the operation status identification of adjacent trajectory points changes, or the duration of the operation status exceeds the preset time condition, the corresponding time point is marked as the operation status switching node.

[0031] Based on the running status switching node, the timestamps corresponding to adjacent running status switching nodes are read as the time range of the event segment, and the start and end position coordinate range of continuous route trajectory points within the time range is used as the spatial range of the corresponding route event segment. Adjacent switching nodes are used as boundaries to divide the continuous route operation process into multiple route event segments.

[0032] The event segments of each route are organized according to the time sequence of the route operation, and the corresponding operation status type, start and end time, spatial range and operation attribute information are recorded for each route event segment, thereby generating the actual route event sequence.

[0033] It should be noted that the operational status switching node refers to the point in time in the flight path where the operational status of the flight path changes due to changes in external or internal factors.

[0034] The preset time conditions are obtained by analyzing the statistical distribution of time intervals between historical route trajectory points. The high-frequency normal sampling intervals are used as the benchmark, and a certain fluctuation range is allowed above and below the benchmark to cover the time variation range under normal navigation conditions.

[0035] A route trajectory point sequence represents a continuous set of trajectory points arranged in chronological order, consisting of speed, position coordinates, and operational status. A route event sequence, on the other hand, is a set of route segments with clear start and end times and operational attributes, formed by segmenting the trajectory based on operational status switching nodes according to the route trajectory point sequence.

[0036] S2: Based on the actual route event sequence, embed the route governance rules into the corresponding route segments to construct the semantic representation of route governance.

[0037] S2.1: For each route event segment in the actual route event sequence, extract the features of arrival time deviation, flight path deviation, port of call change, unit voyage cost deviation, and unit voyage carbon emission deviation, and calculate the embedding strength of the route governance rules on the corresponding route event segment.

[0038] Specifically, for each route event segment, the normalized offsets of arrival time deviation, flight path deviation, unit distance cost deviation, and unit distance carbon emission deviation are calculated in the corresponding route event segment. These offsets are obtained by proportionally converting the actual offset values ​​to the baseline offset range allowed by the corresponding governance rules. Based on the normalized offsets, segmented mapping relationships are set according to arrival time, route compliance, cost constraints, and carbon emission constraints, mapping different offset intervals to corresponding rule trigger intensity values, forming a rule trigger intensity vector corresponding to the route event segment. According to the priority of the governance rules in route operation management or the historical governance effect, weights are assigned to the governance rule trigger intensity, and a weighted summation calculation is performed on the governance rule trigger intensity vector to obtain the embedding strength of the route governance rules in the corresponding route event segment.

[0039] S2.2: The embedding strength is jointly processed with the speed information, position coordinate information and operation status identification information of the corresponding route event segment, and combined according to the time sequence of the route operation to generate a semantic representation of route governance.

[0040] Specifically, the embedding strength is combined with the speed information, position coordinate information and operation status identification information. The speed information, position coordinate information and operation status identification information are standardized respectively. Among them, the speed information is converted into units and normalized between intervals, the position coordinate information is expressed in a unified geographic coordinate, and the operation status identification information is mapped with status encoding.

[0041] After standardization, the embedding strength of the route governance rules is used as the weight benchmark. The speed information, position coordinate information, and operation status identification information are weighted according to their degree of influence in the operation of the route. A comprehensive representation is formed by weighted summation. According to the time sequence of the route operation, the comprehensive representations of each route event segment are continuously combined and adjusted in combination with the semantic consistency between adjacent segments to keep the temporal continuity and governance constraint expression of each route event segment consistent, thus generating a semantic representation of route governance.

[0042] Semantic consistency is measured by calculating the similarity between the comprehensive representation vectors corresponding to adjacent route event segments. When the similarity is lower than the reference consistency limit, the comprehensive representation of the current route event segment is corrected by a smooth adjustment method based on the comprehensive representation of the previous route event segment, thereby reducing semantic abrupt changes between adjacent route event segments and maintaining temporal continuity.

[0043] The reference consistency limit is obtained by statistical analysis of the similarity of adjacent comprehensive representations of historical normal flight event segments.

[0044] In contrast, existing methods typically analyze features such as speed, position coordinates, and operational status independently, lacking comprehensive processing and adjustment. By embedding governance rules into the strength of the data and combining it with information on speed, position coordinates, and operational status, and by weighting adjustments based on the relevance and importance of the route operation, the timeliness and stability consistency of the route status can be improved.

[0045] S3: Based on the semantic representation of route governance, the artificial intelligence analysis process outputs the arrival time series of each port of call and the reliability evolution characteristics as the route progresses. Then, based on the arrival time series and reliability evolution characteristics, a counterfactual route trajectory sequence is generated.

[0046] S3.1: Based on the semantic representation of route governance, the governance semantic features of each route event segment are read in the time sequence of route operation and used as the state input in the route advancement process to generate a semantic state sequence.

[0047] Specifically, based on the semantic representation of route governance, the governance semantic features of each route event segment are read in chronological order of route operation, and governance semantic features such as arrival time deviation, flight path deviation, and port of call change corresponding to each route event segment are extracted. According to the time stamp order of each route event segment in the route operation, the governance semantic features corresponding to each event segment are read sequentially, and the governance semantic features of the current event segment are continuously connected with the governance semantic features of the previous event segment as input information. Combined with the route operation status, a semantic state sequence is generated.

[0048] It should be noted that the semantic state sequence refers to the set of states formed by combining the governance semantic features of each route event fragment at the corresponding time point with the route operation status information and arranging them sequentially according to the route operation time.

[0049] S3.2: Based on the semantic state sequence, the arrival time of the route at each port of call is extrapolated port by port. At the same time, the timeliness stability and fluctuation of the route are evaluated during the route's progress, forming the arrival time sequence and reliability evolution characteristics.

[0050] Specifically, based on the temporal order of each route event segment in the semantic state sequence, the temporal characteristics, spatial location, speed information, operational status identifier, and governance semantic characteristics corresponding to each route event segment are read sequentially. Starting from the arrival time of the previous port of call, the expected arrival time of the next port of call is deduced by combining the operational characteristics of the current route event segment. Continuous arrival times are generated port by port in the order of port calls. During the port-by-port deduction process, the magnitude of the change in arrival time between adjacent ports of call and the trend of change in continuous segments are compared to assess whether the arrival time remains stable or fluctuates during the route advancement. The expected arrival times of each port of call are summarized in chronological order to form an arrival time sequence, and the diffusion or convergence trend of arrival time changes during the route advancement process is organized into reliability evolution characteristics.

[0051] S3.3: Using arrival time series and reliability evolution characteristics as constraint information, the route operation process is retrospectively reconstructed, and the temporal and spatial positions of each route event segment are adjusted to generate a counterfactual route trajectory sequence.

[0052] Specifically, using arrival time series and reliability evolution characteristics as time constraints, each event segment in the route operation process is retrospectively corrected according to the route operation sequence. Time and position adjustments are made to ensure that the time continuity between adjacent route event segments and the berthing order remain unchanged. After the time and position adjustments, the corresponding trajectory points in each route event segment are rearranged according to the corrected time sequence, and the navigation path and berthing position are simultaneously adjusted in the spatial dimension. The spatial position adjustments are carried out without changing the original berthing order, navigation direction, and basic waterway range of the route. By simultaneously satisfying the time continuity constraint, spatial coherence constraint, and basic compliance constraint of route operation, the route operation process is retrospectively reconstructed, and the adjusted route event segments are reconnected to generate a counterfactual route trajectory sequence.

[0053] S4: Reconstruct the actual flight path sequence from the actual flight path event sequence and the counterfactual flight path sequence, perform spatiotemporal alignment and difference analysis on the flight path segments to identify abnormal flight path behavior; perform differential calculation based on the abnormal flight path behavior to form impact quantification information.

[0054] S4.1: Based on each flight event segment in the actual flight event sequence, reconstruct the actual flight trajectory sequence; extract the corresponding trajectory data from the actual flight trajectory sequence and the counterfactual flight trajectory sequence to form the actual flight trajectory segment and the counterfactual flight trajectory segment.

[0055] Specifically, using the start and end times and spatial range of each event segment in the actual flight path event sequence as an index, the speed information, position coordinate information, and operational status identifier information within the corresponding time period are read sequentially according to the time sequence. The position information of consecutive time points is sequentially spliced ​​together, and the changes in navigation direction and speed between adjacent position points are continuously correlated. The consistency between the beginning and end of the segments is maintained on the time axis to form the actual flight path trajectory sequence. From the actual flight path trajectory sequence and the counterfactual flight path trajectory sequence, the corresponding trajectory data is extracted according to the time boundary and spatial range of each event segment to form the actual flight path trajectory segment and the counterfactual flight path trajectory segment.

[0056] It should be noted that the actual flight path sequence refers to the complete flight path reconstructed based on the time and space information of the flight path, using each flight path event segment in the actual flight path event sequence.

[0057] S4.2: Using the counterfactual flight path segment as the time reference, perform time alignment and spatial resampling on the corresponding actual flight path segment to generate aligned trajectory segment pairs.

[0058] Specifically, using counterfactual flight path segments as the time reference, the time nodes in the counterfactual flight path segments are identified and compared with the time nodes of the corresponding actual flight path segments. After aligning the time nodes, the time of the actual flight path segments is interpolated and adjusted to keep the time in the actual flight path segments consistent with that of the counterfactual flight path segments. The actual flight path segments are then spatially resampled, and the position coordinate information is adjusted through spatial interpolation to match the spatial position of the actual flight path segments with that of the counterfactual flight path segments, generating aligned trajectory segment pairs.

[0059] S4.3: Based on the aligned trajectory segment pairs, calculate the differences between the actual route trajectory segment and the counterfactual route trajectory segment in terms of arrival time deviation, navigation path offset, operating speed change and operating status switching frequency, and form difference description information.

[0060] Specifically, based on the aligned trajectory segment pairs, the deviation in arrival time between the actual and counterfactual route trajectory segments is calculated, and the difference in arrival time between the actual and counterfactual route trajectory segments is compared; the spatial offset between the actual and counterfactual route trajectory segments is measured, and the degree of offset is calculated; based on airspeed information, the difference in operating speed between the actual and counterfactual route trajectory segments is compared, and the magnitude of speed change is calculated; the frequency of operating state switching is analyzed, and the frequency difference of state switching between the actual and counterfactual route trajectory segments within the same time range is statistically analyzed; and through the differences in arrival time, operating speed, and frequency, a description of the degree of difference is formed.

[0061] S4.4: Perform continuous analysis of the difference description information according to the flight route operation sequence. When the difference of one flight route segment continues to increase and abruptly occurs, it is judged as abnormal flight route behavior.

[0062] Specifically, according to the flight route operation sequence, the difference description information corresponding to each flight route segment is arranged chronologically to form a difference time series. In the difference time series, a continuous flight route segment of a preset length is used as an analysis window to calculate the average level and dispersion of the difference within the analysis window, and to construct a reference threshold to characterize the normal fluctuation range. The difference changes between adjacent flight route segments are compared segment by segment. When the difference of multiple consecutive flight route segments within the same analysis window is continuously higher than the reference threshold, it is determined that the difference shows a continuous amplification trend. The difference increment between adjacent flight route segments is compared. When the difference increment between a flight route segment and the previous flight route segment is significantly higher than the normal change range within the window, it is determined that the difference has abruptly changed. When the continuous amplification trend and abrupt change characteristics of the difference appear simultaneously on the same flight route segment, the corresponding flight route segment is identified as abnormal flight route behavior.

[0063] It should be noted that the preset length is obtained based on the statistical results of the stable interval of the difference between the average duration of adjacent route event segments in the route event sequence and the difference under historical normal operation conditions.

[0064] The reference threshold for the normal fluctuation range refers to the upper limit boundary obtained based on statistical information describing the differences in historical normal flight route segments. The range is determined by using the historical average difference as the center and combining it with the upper and lower limits formed by historical fluctuation levels.

[0065] S4.5: Within the time range corresponding to abnormal route behavior, perform differential calculations on the arrival time, sailing distance, energy consumption per unit distance, and duration of operation status for the actual route trajectory segment and the counterfactual route trajectory segment to form a multidimensional differential feature set.

[0066] Specifically, within the time range corresponding to abnormal route behavior, data such as arrival time, sailing distance, energy consumption per unit distance, and duration of operation status are extracted from actual route trajectory segments and counterfactual route trajectory segments. Difference calculation is performed between actual and counterfactual route trajectory segments to calculate the difference of each indicator between the actual and counterfactual route trajectory segments, thereby obtaining the difference features of each indicator and forming a multidimensional difference feature set.

[0067] S4.6: Based on the multidimensional differential feature set, the collaborative changes of each differential feature within the same abnormal flight path behavior are integrated and processed to generate quantitative description information; the quantitative description information is associated with the corresponding abnormal flight path behavior to form influence quantitative information.

[0068] Specifically, the arrival time difference feature, flight distance difference feature, unit flight energy consumption difference feature, and operational status duration difference feature in the multidimensional difference feature set are all subjected to interval normalization processing to convert the difference features of different dimensions into comparable standardized values. Based on the relative importance feature weights of each difference feature to the abnormal route behavior, the feature weights can be set according to the correlation between each difference feature and changes in route timeliness, cost, or energy consumption in historical abnormal samples, or configured according to the priority preset by the governance rules. For each standardized difference feature within the same abnormal route behavior, a weighted summation or weighted average operation is performed according to the corresponding weight to obtain quantitative description information representing the overall impact of the abnormal route behavior. The quantitative description information is associated with and recorded one by one with the corresponding abnormal route behavior identifier and route event fragment to form impact quantification information.

[0069] Compared to analyzing abnormal flight path behavior based on a single difference feature, this method comprehensively considers the influence of each multidimensional difference feature in abnormal flight path behavior through normalization and weighted calculation of the multidimensional difference feature set. Through standardization and weighted summation or weighted averaging, the influence of different multidimensional difference features can be quantified, and the synergistic changes among multidimensional difference features can be analyzed, ensuring high-precision assessment of abnormal flight path behavior.

[0070] S5: Based on arrival time series, reliability evolution characteristics, abnormal route behavior and impact quantification information, generate route adjustment and replanning suggestions, construct corresponding interpretable route scores and counterfactual difference evidence chains, and output analysis content.

[0071] S5.1: Based on arrival time series and reliability evolution characteristics, conduct a comprehensive assessment of the timeliness stability and risk change trends of the route in each segment to identify the route segments that need to be adjusted.

[0072] Specifically, based on arrival time series and reliability evolution characteristics, statistical analysis is conducted on the average arrival time, fluctuation range, and trend of each segment to characterize the stability of the segment over time. Combined with reliability evolution characteristics, a comparative analysis is performed on the stability changes of the same segment within a continuous operating cycle, focusing on identifying whether reliability indicators show a continuous decline, frequent fluctuations, or phased deterioration. A joint analysis of arrival time fluctuation range and reliability evolution trend is performed; when a segment simultaneously exhibits increased arrival time fluctuation and decreased or intensified reliability, the operational risk of the segment is determined to be increasing. By comparing the variation ranges between different segments, segments with relatively high risk are identified and designated as route segments requiring focused attention and adjustment.

[0073] Among them, the relatively high-risk segments refer to segments within the same route that, after uniformly quantifying and ranking the arrival time fluctuations and reliability evolution indicators of each segment, have a higher risk characterization value than other segments and remain at a high level in the overall distribution for a long period of time.

[0074] For example, to identify whether key reliability indicators show a continuous decline, frequent fluctuations, or phased deterioration, the reliability indicators of a flight segment are statistically analyzed over multiple consecutive operating cycles. If the reliability indicator shows a monotonically declining trend over multiple consecutive cycles, and the decline exceeds the historical normal fluctuation range, it is judged as a continuous decline. If the reliability indicator frequently shows significant increases and decreases between adjacent cycles, and the fluctuation frequency is higher than that of other flight segments, it is judged as frequent fluctuations. If the reliability indicator suddenly deteriorates within a certain period of time and remains at a low level in subsequent cycles, it is judged as a phased deterioration.

[0075] For example, when a segment of a voyage simultaneously exhibits increased arrival time volatility and decreased or intensified reliability, it can be determined that the operational risk of the segment is on the rise. If the arrival time deviation of the segment has been stable within a small range in its historical operations, but in recent voyages, the dispersion of arrival time deviation has increased, and the corresponding reliability indicators have gradually decreased from the stable range or shown sharp fluctuations, then it can be concluded that the segment is simultaneously experiencing increased arrival time volatility and decreased or intensified reliability.

[0076] S5.2: For route segments, combine abnormal route behavior and corresponding quantitative information on impact to analyze the specific ways in which abnormal behavior affects route timeliness, operating costs and energy consumption, and form a set of adjustable factors.

[0077] Specifically, for each route segment, based on the quantitative impact information corresponding to abnormal route behavior, the actual changes in the impact indicators of arrival time deviation, flight path offset, speed fluctuation, and number of operational status switching are calculated within the current route segment. These impact indicators are then compared with the historical normal operating levels of the route segment to obtain the magnitude of change for each indicator under abnormal conditions. The magnitudes of change are categorized and analyzed according to their degree of impact on route operation results to determine the direction and extent of the impact of abnormal route behavior on route timeliness, operating costs, and energy consumption. By comparing the relative changes among the various impact indicators, the situations with the most significant impact on timeliness, cost, or energy consumption in abnormal route behavior—such as a significantly increased arrival time deviation, increased flight path offset, or abnormally high frequency of operational status switching—are identified as the primary adjustment targets. Finally, the influencing factors are summarized to form a set of adjustable factors.

[0078] It should be noted that abnormal behavior refers to situations in which the route deviates from the normal route plan or expectation during operation; including deviations in arrival time, deviations in flight path, fluctuations in speed, and abnormal frequency of switching of operating status.

[0079] The route operation results refer to the overall operational status and performance of the route under the current operating conditions, based on arrival time series, reliability evolution characteristics, and quantitative information on the impact of abnormal route behavior. This is achieved by comprehensively characterizing the changes in arrival time, operational stability, and resource consumption of each route segment during actual operation.

[0080] The formula for arrival time deviation is: ; In the formula, This indicates that the arrival time has deviated. Indicates the actual arrival time. Indicates the expected arrival time in Hong Kong. To indicate the actual situation, Indicates the expected situation.

[0081] The actual arrival time refers to the real arrival time of the vessel at the corresponding port obtained from the actual route trajectory sequence, while the expected arrival time refers to the theoretical arrival time derived from the voyage plan or counterfactual route trajectory.

[0082] The formula for deviation from the navigation path is: ; In the formula, Indicates deviation from the navigation path. Indicates the actual length of the navigation path. Indicates the length of the planned navigation route. This indicates the status of the plan.

[0083] The actual navigation path length refers to the cumulative navigation distance calculated based on the continuous position coordinates in the actual route trajectory sequence, while the planned navigation path length refers to the theoretical navigation distance calculated based on the preset route in the voyage plan or counterfactual route trajectory. Both the actual navigation path length and the planned navigation path length are measured in length and are obtained using the same coordinate system and distance calculation method.

[0084] Speed ​​fluctuations, the formula is: ; In the formula, Indicates fluctuations in speed. Indicates the actual speed. Indicates the planned speed.

[0085] The frequency of operation state switching is calculated using the following formula: ; In the formula, Indicates the frequency of operation state switching. Indicates the actual frequency of switching between operating states. This indicates the expected frequency of switching between operating states.

[0086] Among them, the time, distance, speed and frequency of the port time delay formula, the navigation path deviation formula, the speed fluctuation formula and the operation status switching frequency formula are calculated separately; there is no comparison across physical quantities, and the dimensions are consistent.

[0087] S5.3: Based on the set of adjustable factors, combine and extrapolate the route operation parameters, call order and operation strategy, evaluate the feasibility and expected effect, and select and form route adjustment and replanning recommendations.

[0088] Specifically, based on the set of adjustable factors, candidate adjustment schemes are generated by combining route operation parameters, call-in sequence, and operation strategies. For each candidate scheme, an executability determination is performed. The executability determination uses channel navigability, meteorological and sea state levels, port operation window availability, operating cost budget constraints, and carbon emission constraints as constraints. When a candidate scheme meets the following conditions within the corresponding route segment: the channel is navigable, the meteorological and sea state does not exceed the preset operating threshold, the target port has an operation window in the corresponding time period, and the cost increment and carbon emission increment do not exceed the allowable upper limit, the candidate scheme is deemed executable.

[0089] Provided that the feasibility assessment is passed, the degree of improvement in timeliness, cost, and energy consumption of the candidate solutions is quantitatively evaluated based on the changes in arrival time, operating costs, and energy consumption corresponding to the candidate solutions, and an expected effect evaluation is formed. The expected effect evaluations of each candidate solution are compared, and the solution with the best overall improvement effect under the condition of meeting the feasibility constraints is selected as the recommendation for route adjustment and replanning.

[0090] It should be noted that the operating threshold is obtained by statistically analyzing meteorological and hydrological records during the historical normal operation of the target route. The historical interval that meets the requirements for safe navigation of ships and does not cause abnormal route behavior is used as a reference range, and the value is limited to the typical quantile range of the historical interval.

[0091] For example, by comparing the expected effects of each candidate scheme, the scheme with the best overall improvement effect under the condition of meeting the executability constraints is selected; for each candidate scheme that passes the executability judgment, the changes in arrival time, operating costs and energy consumption are calculated respectively, and the three changes are uniformly converted into directly comparable relative improvement indicators; the multiple improvement indicators of each candidate scheme are comprehensively summarized to obtain the evaluation value reflecting the degree of overall improvement; the evaluation values ​​of each candidate scheme are compared horizontally, and the candidate scheme with the evaluation value showing a greater improvement and no significant deterioration of any single indicator is selected as the route adjustment and replanning suggestion with better overall improvement effect.

[0092] S5.4: Based on arrival time series, reliability evolution characteristics, abnormal route behavior, and impact quantification information, a set of scoring elements is generated by route segment, and the contribution relationship of each scoring element within the route segment is attributed and organized to generate an interpretable route score.

[0093] Specifically, based on arrival time series, reliability evolution characteristics, abnormal route behavior, and impact quantification information, a set of various scoring elements is extracted and generated according to route segments. The value changes of each scoring element within its corresponding segment are read one by one. By comparing and analyzing the magnitude, direction, and duration of changes in scoring elements within the same route segment, the degree of influence of each scoring element on the route's operational status is determined. During the analysis, the scoring elements are correlated with the actual operational performance of the route segment to identify scoring elements that affect the route's timeliness, stability, or operational risks. The degree of influence of each scoring element within the route segment is summarized and organized to form an interpretable route score.

[0094] The scoring element set consists of arrival time deviation statistics, reliability indicators, path deviation, speed fluctuation, state switching frequency, cost deviation, and carbon emission deviation.

[0095] S5.5: Based on the actual route trajectory sequence and the counterfactual route trajectory sequence, organize the evidence of changes in arrival time deviation, navigation path deviation, changes in operating speed and differences in operating status switching within the route segment, and generate a counterfactual difference evidence chain.

[0096] Specifically, the arrival time, flight path, operating speed, and operating status of each route segment are extracted from the actual route trajectory sequence and the counterfactual route trajectory sequence; the arrival time, flight path, operating speed, and operating status are compared item by item according to time and space, and the arrival time delay, flight path deviation, and speed change of each route segment are calculated to form difference data; the difference data are organized and integrated according to time order and spatial location to form a counterfactual difference evidence chain.

[0097] It should be noted that the difference data refers to the differences in arrival time, flight path, operating speed, and operating status calculated from the actual flight path sequence and the counterfactual flight path sequence.

[0098] S5.6: Link and encapsulate the explainable route score, counterfactual difference evidence chain, and route adjustment and replanning recommendations to form the analysis content.

[0099] Specifically, using route segments as a unified index, the explainable route scores for corresponding segments are aligned with the counterfactual discrepancy evidence chains. A one-to-one correspondence is established between the risk level and stability level reflected in the explainable route scores and the arrival time deviations, path offsets, and operational status changes reflected in the discrepancy evidence. Based on this correspondence, the adjustment content involved in route adjustment and replanning recommendations is matched with the explainable route scores and discrepancy evidence for the corresponding segments. Route adjustment and replanning recommendations are indexed by route segments, and each recommendation is aligned with the corresponding explainable route scores and counterfactual discrepancy evidence within the same segment. The route segment identifiers, explainable route scores, counterfactual discrepancy evidence, and corresponding adjustment recommendations are uniformly organized and packaged to form the analysis content.

[0100] In summary, this invention reconstructs the theoretical evolution of routes under given operational constraints by constructing counterfactual route trajectories with semantics for route governance, providing a predictable and alignable basis for the operational status of routes in both time and space dimensions. By performing segment-by-segment alignment analysis between the counterfactual route trajectory and the actual route trajectory, it can accurately characterize the sources of deviations caused by timeliness fluctuations, path deviations, or changes in operational status during route operation, transforming previously difficult-to-quantify route anomalies into calculable and traceable difference features, thus completing the structured expression and quantitative description of the operational status of routes.

[0101] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. 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 be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A container shipping route analysis method based on artificial intelligence analysis, characterized in that: include, Collect route operation data of the target container shipping route during the operation and management process, and perform time alignment and normalization processing to generate actual route event sequences; Based on actual flight route event sequences, flight route governance rules are embedded into corresponding flight route segments to construct a semantic representation of flight route governance; Based on the semantic representation of route governance, the arrival time series of each port of call and the reliability evolution characteristics as the route progresses are output through artificial intelligence analysis. Then, the counterfactual route trajectory sequence is generated using the arrival time series and reliability evolution characteristics. The actual flight trajectory sequence obtained by reconstructing the actual flight event sequence and the counterfactual flight trajectory sequence are subjected to spatiotemporal alignment and difference analysis by flight segment to identify abnormal flight behavior. Differential calculations are performed based on abnormal flight path behavior to generate impact quantification information; Based on arrival time series, reliability evolution characteristics, abnormal route behavior and impact quantification information, route adjustment and replanning suggestions are generated, and corresponding interpretable route scores and counterfactual difference evidence chains are constructed to output analysis content.

2. The container route analysis method based on artificial intelligence analysis as described in claim 1, characterized in that: The route operation data includes trajectory data, voyage plan data, port berthing and departure and operation status data, waterway passage status data, meteorological and sea condition data, fuel consumption data, and cost and carbon emission data.

3. The container route analysis method based on artificial intelligence analysis as described in claim 2, characterized in that: The specific steps for generating the actual flight path event sequence are as follows: The route operation data is timestamped and normalized, and speed information, position coordinate information and operation status identification information are extracted from the route operation data to form a route trajectory point sequence. Based on the sequence of route trajectory points, the operation status switching nodes are identified, and the route operation process is divided with adjacent switching nodes as boundaries to generate the actual route event sequence.

4. The container route analysis method based on artificial intelligence analysis as described in claim 3, characterized in that: The route governance rules include on-time requirements, route compliance requirements, cost budget constraints, and carbon emission constraints.

5. The container route analysis method based on artificial intelligence analysis as described in claim 4, characterized in that: The specific steps for constructing the semantic representation of route governance are as follows: For each event segment in the actual route event sequence, features such as arrival time deviation, flight path deviation, port of call change, unit voyage cost deviation, and unit voyage carbon emission deviation are extracted, and the embedding strength of the route governance rules on the corresponding route event segments is calculated. The embedding strength is jointly processed with the speed information, position coordinate information and operation status identification information of the corresponding route event fragment, and combined according to the time sequence of route operation to generate a semantic representation of route governance.

6. The container route analysis method based on artificial intelligence analysis as described in claim 5, characterized in that: The specific steps for generating the counterfactual flight path sequence are as follows. Based on the semantic representation of route governance, the governance semantic features of each route event segment are read in the time sequence of route operation and used as the state input in the route advancement process to generate a semantic state sequence. Based on the semantic state sequence, the arrival time of the route at each port of call is extrapolated port by port. At the same time, the timeliness stability and fluctuation of the route are evaluated during the route's progress, forming the arrival time sequence and reliability evolution characteristics. Using arrival time series and reliability evolution characteristics as constraint information, the route operation process is retrospectively reconstructed, and the temporal and spatial positions of each route event segment are adjusted to generate a counterfactual route trajectory sequence.

7. The container route analysis method based on artificial intelligence analysis as described in claim 6, characterized in that: The specific process for identifying abnormal flight path behavior is as follows. Based on the event segments of each actual flight route in the actual flight route event sequence, the actual flight route trajectory sequence is reconstructed. Trajectory data is extracted from the actual flight path sequence and the counterfactual flight path sequence to form actual flight path segments and counterfactual flight path segments; Using the counterfactual flight path segment as the time reference, the corresponding actual flight path segment is subjected to time alignment and spatial resampling to generate aligned trajectory segment pairs; Based on the aligned trajectory segment pairs, the differences between the actual route trajectory segment and the counterfactual route trajectory segment in terms of arrival time deviation, navigation path offset, operating speed change and operating status switching frequency are calculated to form difference description information. The difference description information is continuously analyzed according to the flight route sequence. When the difference of one flight route segment is continuously amplified and abruptly changes, it is judged as abnormal flight route behavior.

8. The container route analysis method based on artificial intelligence analysis as described in claim 7, characterized in that: The process of generating impact quantification information through differential calculation based on abnormal flight path behavior is as follows. Within the time range corresponding to the abnormal route behavior, differential calculations are performed on the arrival time, sailing distance, energy consumption per unit distance, and duration of operation status of the actual route trajectory segment and the counterfactual route trajectory segment to form a multidimensional differential feature set. Based on a multidimensional differential feature set, the collaborative changes of each differential feature within the same abnormal flight path behavior are integrated and processed to generate quantitative descriptive information; The quantitative description information is associated with the corresponding abnormal flight path behavior and recorded to form the impact quantitative information.

9. The container route analysis method based on artificial intelligence analysis as described in claim 8, characterized in that: The specific process for generating route adjustment and replanning suggestions is as follows. Based on arrival time series and reliability evolution characteristics, a comprehensive assessment of the timeliness stability and risk change trends of the route in each segment is conducted to identify the route segments that need to be adjusted. For each route segment, by combining abnormal route behavior and corresponding quantitative information on its impact, we analyze the specific ways in which abnormal behavior affects the route's timeliness, operating costs and energy consumption, and form a set of adjustable factors. Based on the set of adjustable factors, the route operating parameters, call order and operating strategies are combined and simulated, and the feasibility and expected effects are evaluated to form recommendations for route adjustment and replanning.

10. The container route analysis method based on artificial intelligence analysis as described in claim 9, characterized in that: The specific process for analyzing the output content is as follows. Based on arrival time series, reliability evolution characteristics, abnormal route behavior and impact quantification information, a set of scoring elements is generated by route segment, and the contribution relationship of each scoring element within the route segment is attributed and organized to generate an interpretable route score. Based on the actual route trajectory sequence and the counterfactual route trajectory sequence, the evidence of changes in arrival time deviation, navigation path deviation, operating speed change and operating status switching within the route segment is organized in a way to generate a counterfactual difference evidence chain. The explainable route scores, counterfactual discrepancy evidence chains, and route adjustment and replanning recommendations are linked and encapsulated to form the analysis content.

Images