Dynamic optimization method and system for air transit route

Abstract

The invention is suitable for the field of air transition transportation, and provides an air transition route dynamic optimization method and system, and the method comprises the steps: obtaining real-time hydrological monitoring data and historical tide model data, generating a dynamic golden navigation window, compressing an aqueduct shift interval in a window period, and adjusting the total amount; image recognition and reservation data are combined to generate a cross-shift vehicle ship-pooling combination, and an overwater vehicle-pooling network is constructed; and outputting a transportation scheme including temporary stop point optimization and connection path adjustment, triggering a sprint mode through a coupling mechanism of a tidal window and a ship-pooling demand, and constructing a vehicle path relay network during ebb tide to realize resource reuse. The system comprises a data processing module, a shift adjusting module, a ship-pooling combination generating module and the like. According to the scheme, the problems of low scheduling efficiency, resource waste and the like of a traditional fixed shift are solved, the dynamic adaptability and the resource utilization rate of steam transition transportation are improved, and the method is suitable for a steam transition route optimization scene influenced by tides.

Description

Technical Field

[0001] The present invention belongs to the field of ferry transportation, and in particular relates to a method and system for dynamically optimizing ferry routes. Background Art

[0002] As an important component of water transportation, ferry transport is responsible for transporting vehicles and personnel across waterways in areas such as estuaries and straits. Significantly affected by natural conditions such as tides and hydrology, traditional ferry operations rely on historical tidal data and fixed schedules, making it difficult to adapt to real-time water level changes and dynamic transportation needs. With the development of intelligent sensing technology and optimization algorithms, the digital and dynamic management of ferry routes has become a key direction for improving transportation efficiency. However, existing technologies still have significant shortcomings in terms of accurate tidal window prediction, flexible schedule adjustment, and optimized vehicle-to-vessel combinations. There is an urgent need to build a dynamic optimization system that integrates real-time data and intelligent algorithms.

[0003] Existing ferry route scheduling primarily relies on historical tidal models to set fixed navigation periods and intervals, with the total number of trips adjusted through manual experience or simple rules. Vehicle loading often utilizes a first-come, first-served fixed-schedule allocation model, lacking integrated analysis of vehicle dimensional characteristics (such as size, destination, and time tolerance). A lack of real-time response mechanisms makes it difficult to dynamically adjust navigation strategies and berth allocations in the face of rapidly changing tides. Return trips are typically empty, preventing cross-time resource reuse. While these solutions can meet basic transportation needs, they struggle to achieve efficient scheduling under complex hydrological conditions and fluctuating transport flows.

[0004] Existing technologies are unable to integrate hydrological monitoring data and tidal models in real time, resulting in insufficient accuracy in navigation window predictions, frequent suspension of navigation due to insufficient water levels, or waste of resources during high water periods; the fixed-shift model makes it difficult to dynamically compress shift intervals based on window duration, resulting in low berth utilization and long vehicle waiting times; there is a lack of a cross-shift vehicle ship combination mechanism, resulting in low deck space utilization and high empty mileage; in the face of emergencies such as rapidly rising tides, the sprint mode cannot be triggered to adjust the speed and berth allocation, which can easily lead to flight delays; there is a lack of a vehicle path relay network when returning at low tide, empty ship returns are common, and the ability to reuse resources across time periods is weak. Summary of the Invention

[0005] The purpose of the present invention is to provide a method and system for dynamically optimizing ferry routes, aiming to solve the technical problems existing in the prior art identified in the background technology.

[0006] The present invention is achieved by a method for dynamically optimizing ferry routes, the method comprising:

[0007] Obtain real-time hydrological monitoring data and historical tidal model data, calculate the time period that meets the water level conditions for continuous navigation on the same day, and generate a dynamically changing golden navigation window;

[0008] During the golden navigation window period, the ferry frequency will be shortened according to the duration of the golden navigation window, and the total number of flights will be adjusted dynamically;

[0009] Through image recognition and reservation data, the system scans vehicle dimensional features in real time, including the physical dimensions of the vehicle to be ferried, the GPS coordinates of the destination, the reservation time tolerance threshold, and the cargo type label. Based on these vehicle dimensional features, it generates vehicle pooling combinations across ferry schedules, building a flexible and scalable water carpooling network.

[0010] For vehicle-ship combinations, a transportation plan is output that includes optimization of temporary docking point coordinates, dynamic adjustment of connecting routes, and compensation factors for inter-shift connection time.

[0011] Through the coupling coordination mechanism of tidal windows and ship-sharing demand, the window period sprint mode is triggered when the tide level rise rate exceeds the preset threshold;

[0012] By utilizing the return window when the tide recedes, a vehicle path relay network is constructed between adjacent ferry crossings to reuse reverse-flow ferry resources across time periods.

[0013] As a further solution of the present invention, the generation of a dynamically changing golden navigation window specifically includes:

[0014] Collect real-time water level elevation and flow velocity vectors, integrate astronomical tide forecasts with meteorological typhoon storm surge correction parameters, establish a water level prediction model, and predict the water level change trajectory in the next 4 hours:

[0015] ;

[0016] in, For the future The water level at the moment, is the weight coefficient, For the current moment Real-time water level elevation, For the current moment The water velocity vector, is the prediction step length, Time for astronomical tide forecast Water level correction value, is the correction value for the impact of storm surge, is the error term;

[0017] Set the minimum safe draft for ferries and filter out periods where the water level is continuously higher than the minimum safe draft + 0.5m;

[0018] Set the time window and safety area, and calculate the water level fluctuation variance in the filtered period. When the water level fluctuation variance is less than the safety area for three consecutive windows, it is determined to be a stable navigation window.

[0019] By superimposing ship traffic density constraints, a golden navigation window sequence with time margin is generated.

[0020] As a further solution of the present invention, the interval compression of ferry flights and the dynamic adjustment of the total number of flights specifically include:

[0021] According to the duration of the golden navigation window Interval with regular shifts , calculate the compressed ferry interval:

[0022] ;

[0023] in, is the compressed ferry interval time, is the compression coefficient, is the window duration threshold;

[0024] Based on the predicted arrival time of the ship's AIS trajectory data, temporary flights are dynamically added until the berth utilization rate reaches 90%;

[0025] When the remaining time in the window is less than the full voyage time, the short-distance transfer mode is activated and only routes between adjacent ferry ports are operated.

[0026] As a further solution of the present invention, generating a vehicle-ship combination across ferry schedules specifically includes:

[0027] Construct a spatiotemporal network graph with ferries as mobile nodes, with node attributes including planned arrival and departure times and remaining capacity;

[0028] Taking the GPS coordinates of the destination as the center, the spatiotemporal density function is constructed by integrating the vehicle reservation time tolerance threshold;

[0029] The optimization objectives are defined, including minimizing weighted waiting time, minimizing empty mileage, ensuring deck area utilization ≤ 85%, and ensuring that the interval between shifts at the same berth is ≥ the safe berthing time. The cross-shift ship pooling scheme is solved based on the spatiotemporal density function, and a ship pooling combination matrix containing a list of ship IDs and license plate numbers is generated.

[0030] As a further solution of the present invention, the output transportation solution specifically includes:

[0031] Based on the GPS coordinate distribution of vehicle destinations in the ship-pooling matrix, the optimal docking center is solved and the temporary docking area is divided:

[0032] ;

[0033] in, is the GPS coordinate of the optimal stop, For the The weight of the vehicle model, To point to The Euclidean distance between the vehicle and the destination, is the distance weight index, For the The GPS coordinates of the vehicle's destination;

[0034] For vehicles that have been assigned to a ship group but have not yet boarded the ship, the delay probability of each vehicle is calculated and a delay threshold is set. If the delay probability exceeds the delay threshold, a backup slot is allocated to the vehicle:

[0035] ;

[0036] in, is the delay probability, To predict the time of embarkation, The original scheduled boarding time, is the standard deviation of the delay time, is the normal cumulative distribution function.

[0037] As a further solution of the present invention, the trigger window period sprint mode specifically includes:

[0038] Set a rising threshold. When the real-time tide level rise rate exceeds the rising threshold and lasts for 2 minutes, a sprint warning is triggered, including:

[0039] Increase the ferry's cruising speed to the maximum within safety thresholds;

[0040] Reorganize the unshipped boat combinations and prioritize boarding for vehicles whose destinations are the ferry with the largest unloading volume and whose reservation time tolerance is lower than the reservation time tolerance threshold.

[0041] Enable dynamic berth allocation strategy to shorten the time interval between ships’ berthing and unberthing to the safe minimum.

[0042] As a further solution of the present invention, the construction of the vehicle path relay network specifically includes:

[0043] 30 minutes before the low tide window opens, scan the database of waiting vehicles for ferrying, select vehicles whose destination coordinates are adjacent to the ferry and whose reservation time tolerance threshold is greater than the low tide window duration, and mark them as the set of relay objects;

[0044] A complete graph is constructed with all ferry crossings in the region as nodes, where nodes are the locations of each ferry crossing and edge weights are the estimated travel times between ferry crossings. The optimal route sequence is solved and output to minimize the total travel time and ensure that the time windows of each flight segment overlap.

[0045] Obtain the current data of each return ferry, including remaining available space, estimated arrival time at each ferry port, and basic transportation price, and establish an optimization allocation model to assign the corresponding ferry to each waiting vehicle, generating a vehicle-ferry matching matrix;

[0046] Get the available return ferry seats and corresponding prices, and calculate the discounted price based on the remaining available seats.

[0047] Another object of the present invention is to provide a ferry route dynamic optimization system, the system comprising:

[0048] The data processing module is used to obtain real-time hydrological monitoring data and historical tidal model data, calculate the time period that meets the water level conditions for continuous navigation on the same day, and generate a dynamically changing golden navigation window;

[0049] The ferry schedule adjustment module is used to compress ferry schedules during the golden navigation window period according to the duration of the golden navigation window and dynamically adjust the total number of schedules;

[0050] The carpooling combination generation module uses image recognition and reservation data to scan vehicle dimensional characteristics in real time, including the physical dimensions of the vehicle to be ferried, the GPS coordinates of the destination, the reservation time tolerance threshold, and the cargo type label. Based on the vehicle dimensional characteristics, it generates carpooling combinations that span ferry schedules, building a flexible and scalable water carpooling network.

[0051] The transport plan optimization module is used to output a transport plan for vehicle-ship combinations, including optimization of temporary docking point coordinates, dynamic adjustment of connecting routes, and compensation factors for inter-shift connection time;

[0052] The ship-sharing demand coordination module is used to trigger the window period sprint mode when the tide rise rate exceeds the preset threshold through the coupling coordination mechanism of the tide window and the ship-sharing demand;

[0053] The reverse boat-pooling module is used to utilize the return window when the tide recedes to build a vehicle path relay network between adjacent ferry crossings and reuse reverse boat-pooling resources across time periods.

[0054] The beneficial effects of the present invention are:

[0055] The present invention constructs a dynamic optimization system for ferry routes through multi-source data fusion and intelligent optimization algorithms. The generation of golden navigation windows driven by real-time hydrological and tidal data can accurately capture stable periods of high water levels and improve the utilization rate of navigation time in water areas; the compression and dynamic adjustment of flight intervals during the golden window period, combined with the temporary addition of flights based on AIS trajectory data, significantly improves berth utilization and flight turnover efficiency; cross-ferry combinations and flexible water carpooling networks based on vehicle dimensional characteristics achieve efficient use of deck space and reduce empty mileage; temporary stopover optimization and dynamic adjustment of connecting routes, combined with the allocation of spare space based on delay probability calculation, ensure the real-time and stability of the transportation plan; the coupling coordination mechanism of tidal windows and carpooling needs improves transportation efficiency during rapid tide changes through sprint mode; the vehicle path relay network and reverse carpooling resource reuse during the low tide return phase reduce the empty ship return rate and balance the transportation flow within the tidal cycle. The overall solution realizes intelligent and dynamic management of ferry transportation, with significant advantages in shipping safety, resource utilization and system responsiveness. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 A flow chart of a method for dynamic optimization of ferry routes provided by an embodiment of the present invention;

[0057] Figure 2 A flowchart of generating a dynamically changing golden navigation window provided by an embodiment of the present invention;

[0058] Figure 3 A flowchart of an embodiment of the present invention for compressing ferry schedules and dynamically adjusting the total number of schedules;

[0059] Figure 4 A flowchart of generating a vehicle-on-vehicle combination across ferry schedules provided by an embodiment of the present invention;

[0060] Figure 5 A flow chart of an output transportation solution provided by an embodiment of the present invention;

[0061] Figure 6 A flowchart of the trigger window period sprint mode provided by an embodiment of the present invention;

[0062] Figure 7 A flowchart of constructing a vehicle path relay network provided by an embodiment of the present invention;

[0063] Figure 8 This is a structural block diagram of a ferry route dynamic optimization system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0064] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0065] Figure 1 The flowchart of the method for dynamic optimization of ferry routes provided by the embodiment of the present invention is as follows: Figure 1 As shown, the method includes:

[0066] S100: Acquire real-time hydrological monitoring data and historical tidal model data, calculate the time period that satisfies the water level continuous navigation conditions on the day, and generate a dynamically changing golden navigation window;

[0067] Real-time hydrological monitoring data is achieved through a multi-type sensor network deployed in the water area, which continuously collects real-time information such as water level elevation and water flow velocity vector. Historical tidal model data integrates long-term astronomical tidal observation data, regional hydrological historical records, etc., and uses data mining technology to extract regular patterns of tidal changes.

[0068] When calculating the time period when the water level meets the conditions for continuous navigation, in addition to the water level prediction model mentioned in the document, it is also necessary to combine the ship draft characteristics database, which contains the minimum safe draft depth parameters of different types of ferries, and sets differentiated safety thresholds for different ship types such as passenger and cargo mixed ferries and large cargo ships.

[0069] In the process of generating the golden navigation window, a meteorological warning data interface will also be introduced to access real-time forecast information on extreme weather such as typhoons and storm surges, and dynamically correct the water level prediction model. When the meteorological department predicts a strong storm surge, the system will automatically increase the buffer value of the safe draft depth, temporarily adjusting it from the conventional 0.5 meters to 0.8 meters to ensure the safety of ferry navigation.

[0070] This step, through multi-source data fusion and dynamic modeling, constructs a navigation time prediction system that can adapt to changes in the water environment. On the one hand, the accurate water level prediction model can capture short periods of high water level stability during tidal fluctuations in advance. For example, in semi-diurnal waters, there may be multiple stable water level periods lasting several hours during the two daily high tides. The system can accurately identify these periods and generate golden navigation windows, allowing ferries to safely navigate during periods that traditional shipping models consider unnavigable, significantly improving the utilization of water navigation time.

[0071] On the other hand, the window sequence generated by combining the constraints on ship traffic density can effectively avoid the concentrated entry and exit of ferries during peak navigation periods. For example, in busy estuary ports, the system will dynamically adjust the time interval and duration of the golden navigation window based on real-time ship traffic data, making the order of ferry entry and exit more orderly and reducing the waiting time at anchorage due to congestion.

[0072] like Figure 2 As shown, the generation of a dynamically changing golden navigation window specifically includes:

[0073] S110 collects real-time water level elevation and flow velocity vectors, integrates astronomical tide forecasts with meteorological typhoon storm surge correction parameters, establishes a water level prediction model, and predicts the water level change trajectory in the next 4 hours:

[0074] ;

[0075] in, For the future The water level at the moment, is the weight coefficient, For the current moment Real-time water level elevation, For the current moment The water velocity vector, is the prediction step length, Time for astronomical tide forecast Water level correction value, is the correction value for the impact of storm surge, is the error term;

[0076] S120, setting the minimum safe draft of the ferry, and screening the period when the water level is continuously higher than the minimum safe draft + 0.5m;

[0077] S130, setting a time window and a safety area, and calculating the water level fluctuation variance in the filtered time period. When the water level fluctuation variance is less than the safety area for three consecutive windows, it is determined to be a stable navigation window;

[0078] S140, superimposing the ship traffic density constraint to generate a golden navigation window sequence with time margin.

[0079] S200: During the golden navigation window period, the ferry frequency will be shortened according to the duration of the golden navigation window, and the total number of flights will be dynamically adjusted;

[0080] When compressing ferry schedules and dynamically adjusting total volumes based on the golden navigation window duration, it is necessary to first obtain the precise duration of the window through real-time water level prediction models and ship traffic density monitoring. This duration data will be linked to the regular schedule intervals in the port operations database for calculation.

[0081] The value of the compression coefficient needs to be combined with historical navigation efficiency data and the type of ships in the current waters. For example, in waterways where cargo ships account for a relatively high proportion, the compression coefficient will automatically match a more aggressive compression strategy to improve heavy-load transportation efficiency.

[0082] During the process of dynamically adding temporary shifts, the system will synchronously analyze real-time parameters such as ship speed and course deviation in the AIS trajectory data, and predict the error range of arrival time through the Kalman filter algorithm. When the prediction error exceeds 15 minutes, the temporary shift scheduling process will be automatically triggered until the berth utilization reaches a dynamic equilibrium state.

[0083] The activation logic of the short-distance shuttle mode requires a comprehensive assessment of the remaining window time and the shortest voyage time between adjacent ferry terminals. When the remaining time can only meet a single short-distance transportation, the system will automatically lock in high-frequency round-trip ferry pairs, such as the route combination between the upstream industrial port and the downstream logistics terminal, and maintain basic transportation services through a circular shuttle mode.

[0084] This step establishes an adaptive matching mechanism between tidal patterns and shipping demand, and achieves granular segmentation of navigation resources by dynamically compressing the intervals between flights. For example, in estuary areas with significant semi-diurnal tides, the system can compress the regular 1-hour interval to 40 minutes during the high tide window, thereby increasing the ferry turnover per unit time by approximately 30%.

[0085] The temporary shift addition strategy based on AIS data can respond to sudden transportation needs in real time. When the number of vehicles waiting to be ferried increases sharply during a certain period of time, the system can complete the scheduling and deployment of temporary shifts in a short time to avoid large-scale queues.

[0086] The short-distance shuttle mode effectively solves the problem of resource waste at the end of the tidal window period. In the last hour before low tide, by focusing on short-distance transportation between adjacent ferry crossings, the capacity utilization rate during this period can be improved.

[0087] Taking coastal ferry ports with frequent tidal changes as an example, after applying this dynamic adjustment mechanism, the average daily effective operating frequency of ferries has increased significantly compared with the traditional fixed frequency model, and the idle time of berths has been significantly shortened. At the same time, by intelligently adjusting the frequency density, the average waiting time of vehicles has been reduced, fully demonstrating the core value of this step in improving the flexibility of the shipping system and resource utilization efficiency.

[0088] like Figure 3 As shown, the interval compresses the ferry schedule and dynamically adjusts the total number of schedules, specifically including:

[0089] S210, based on the duration of the golden navigation window Interval with regular shifts , calculate the compressed ferry interval:

[0090] ;

[0091] in, is the compressed ferry interval time, is the compression coefficient, is the window duration threshold;

[0092] S220: Based on the predicted arrival time of the ship's AIS trajectory data, temporary sailings are dynamically added until the berth utilization rate reaches 90%;

[0093] S230: When the remaining time in the window is less than the full voyage time, the short-distance connection mode is activated, and only routes between adjacent ferry ports are operated.

[0094] S300 uses image recognition and reservation data to scan vehicle dimensional characteristics in real time, including the physical dimensions of the vehicle to be ferried, the GPS coordinates of the destination, the reservation time tolerance threshold, and the cargo type label. Based on these vehicle dimensional characteristics, it generates vehicle pooling combinations across ferry schedules, building a flexible and scalable water carpooling network.

[0095] When generating vehicle combinations across ferry schedules through image recognition and reservation data, it is necessary to rely on multispectral image acquisition equipment deployed at the ferry entrance, use deep learning target detection algorithms to analyze the physical dimensions of the vehicles to be ferried in real time, and simultaneously connect to the port reservation system to obtain data such as the vehicle destination GPS coordinates, reservation time tolerance threshold, and cargo type labels.

[0096] When constructing the space-time network diagram, the system will treat each ferry as a dynamic node, update its planned arrival and departure times and remaining capacity data in real time, and combine historical navigation trajectory data to establish a capacity change prediction model to predict the available space for each route segment in advance.

[0097] The construction of the spatiotemporal density function requires the integration of the spatial distribution of the destination and the time tolerance parameters. For example, the destination coordinates are converted into a latitude and longitude grid matrix, and the reservation time window is used as the time axis to form a three-dimensional density field. The kernel density estimation method is used to identify high-frequency transportation demand areas and time periods.

[0098] When solving the inter-shift ship pooling plan, the deck area utilization constraint in the optimization objective needs to be combined with the vehicle size weight coefficient, and differentiated space occupancy coefficients should be set for different models such as large trucks and small cars to ensure that the actual loading does not exceed the safety threshold of 85%. At the same time, conflicts between ships berthing and leaving berths can be avoided by constraining the interval between shifts at the same berth.

[0099] This step transcends the time and space constraints of traditional fixed-schedule transport, establishing a water-based carpooling network with flexible scalability. By integrating image recognition with reservation data, the system can capture dynamic changes in vehicle dimensional characteristics in real time. During peak freight season, it automatically identifies the concentrated arrival of heavy-loaded trucks at the port and dynamically adjusts the carpooling strategy, prioritizing trucks destined for the same industrial zone onto the same ferry, reducing subsequent connecting transportation costs.

[0100] The cross-ferry sharing mechanism can realize the spatiotemporal reorganization of vehicles booked in different time periods. For example, vehicles with scattered destinations booked between 10 a.m. and 11 a.m. can be integrated into the 10:30 ferry after analyzing the spatiotemporal density function, thereby improving deck utilization.

[0101] The elastically scalable network feature enables the system to adapt to traffic fluctuations. When tourist vehicles arrive at the port in large numbers during the peak tourist season, it can quickly generate ship combinations across multiple shifts to avoid long waits due to insufficient cabin capacity on a single shift.

[0102] like Figure 4 As shown, generating a vehicle-ship combination across ferry schedules specifically includes:

[0103] S310, constructing a spatiotemporal network graph with ferries as mobile nodes, where node attributes include planned arrival and departure times and remaining capacity;

[0104] S320, taking the GPS coordinates of the destination as the center, integrating the vehicle reservation time tolerance threshold to construct a spatiotemporal density function;

[0105] S330 defines optimization objectives, including minimizing weighted waiting time, minimizing empty mileage, ensuring deck area utilization is ≤85%, and ensuring that the interval between shifts at the same berth is ≥ the safe berthing time. A cross-shift ship-pooling scheme is solved based on the spatiotemporal density function, generating a ship-pooling combination matrix containing a list of ship IDs and license plate numbers.

[0106] S400 outputs a transportation plan for vehicle-ship combinations, including optimization of temporary docking point coordinates, dynamic adjustment of connecting routes, and compensation factors for inter-shift connection time.

[0107] When developing a transportation plan for a vehicle-vessel combination, temporary docking point coordinates are optimized based on the spatial distribution of vehicle destinations within the vehicle-vessel combination matrix. A weighted Euclidean distance model is used to determine the optimal docking center. Vehicle weight coefficients are determined based on the vehicle's physical size and cargo type. Heavy trucks are given a higher weight than small passenger cars, and vehicles transporting hazardous goods receive an additional weight to shorten their docking distances.

[0108] Dynamic adjustment of the connecting route requires real-time access to real-time data such as water flow velocity and ship traffic density. When it is detected that the navigation time of a certain section of the waterway is extended due to turbulent water flow, the system will automatically switch to an alternative route and simultaneously update the inter-shift connection time compensation factor.

[0109] The calculation of this compensation factor requires a comprehensive consideration of the delay probability of the previous flight and the cabin capacity redundancy of the next flight. When the ferry is expected to be late, the system will allocate additional time buffers for affected vehicles in subsequent flights and adjust their connecting route priorities.

[0110] This step establishes a dynamic optimization system covering the entire "ship transport-land connection" chain. Optimizing temporary docking point coordinates allows vehicles with different destinations to evacuate via the shortest possible route after docking, avoiding land traffic congestion caused by illogical docking point layouts. This optimization, particularly in scenarios with multiple ferry terminals, improves vehicle departure efficiency.

[0111] Dynamic adjustment of connecting routes can avoid sudden water situations in real time. For example, when the traffic efficiency of a waterway decreases due to temporary construction, the system automatically switches routes to ensure that vehicles arrive at the temporary stop on time, reducing the risk of delays compared to fixed route solutions.

[0112] The inter-shift connection time compensation factor effectively solves the time coordination problem in multi-shift intermodal transport. When multiple shifts operate intensively during the tidal window period, the shift connection efficiency can be improved by dynamically compensating for the time difference, avoiding the subsequent transportation chain being broken due to delays in the previous shift.

[0113] like Figure 5 As shown, the output transportation plan specifically includes:

[0114] S410, based on the GPS coordinate distribution of the vehicle destinations in the ship-sharing combination matrix, the optimal docking center is solved and the temporary docking area is divided:

[0115] ;

[0116] in, is the GPS coordinate of the optimal stop, For the The weight of the vehicle model, To point to The Euclidean distance between the vehicle and the destination, is the distance weight index, For the The GPS coordinates of the vehicle's destination;

[0117] S420: For vehicles that have been assigned to a ship group but have not yet boarded the ship, the delay probability of each vehicle is calculated and a delay threshold is set. If the delay probability exceeds the delay threshold, a spare space is allocated to the vehicle:

[0118] ;

[0119] in, is the delay probability, To predict the time of embarkation, The original scheduled boarding time, is the standard deviation of the delay time, is the normal cumulative distribution function.

[0120] S500, through the coupling coordination mechanism of tidal windows and ship-sharing requirements, triggers the window period sprint mode when the tide level rise rate exceeds the preset threshold;

[0121] When the window period sprint mode is triggered through the coupling coordination mechanism of the tidal window and the demand for ship sharing, the system must first calculate the tide rise rate in real time through high-frequency water level monitoring sensors deployed in the water area (such as pressure water level gauges that sample once per second). When the rate exceeds the dynamic threshold preset based on historical tidal data and the safe draft of the ship (for example, the critical rise rate set in combination with the astronomical tide forecast for the day) and does not fall back for 2 minutes, the sprint warning will be officially triggered.

[0122] At this time, the system will synchronously retrieve the ship's power performance database, dynamically calculate the upper limit of safe speed based on the ferry type (such as mixed passenger and cargo ships, high-speed passenger ships), increase the cruising speed to the upper limit, and monitor the main engine load and navigation posture in real time through the ship automation system (IAS) to ensure the stability of the ship after the speed increase.

[0123] When reorganizing the unshipped combined fleet, the system will connect to the port's real-time unloading volume monitoring platform to obtain the current density data of vehicles to be unloaded at each ferry port, screen out vehicles whose destinations are the top three ferries in terms of unloading volume, and combine the reservation time tolerance threshold (for example, vehicles with a remaining waiting time of less than 30% of the original reservation time will be marked as high priority) to re-sort the boarding priority through a greedy algorithm.

[0124] The dynamic berth allocation strategy requires the use of laser scanning and AIS data fusion technology to generate berth occupancy heat maps in real time, optimize the order of ship berthing and unberthing through reinforcement learning models, shorten the traditional fixed berth interval to a safe minimum (for example, from the conventional 15 minutes to 8 minutes), and at the same time assist ships in precise berthing through shore-based guidance systems.

[0125] This step establishes a real-time response mechanism for tidal dynamics and transport demand, maximizing navigation efficiency during critical periods of rapidly changing tides. When the tide rise rate exceeds a threshold, a speed-up strategy shortens single-trip journey times, ensuring that ferries can complete more trips before the water level reaches its peak. A ship-pooling and reorganization mechanism prioritizes transport to high-demand ferry ports, avoiding the overall decline in route efficiency caused by localized congestion. This improves vehicle turnover at key ferry ports, especially during peak morning and evening hours. Dynamic berth allocation optimizes berthing and unberthing processes in real time, reducing idle berth time and enabling an additional 1-2 trips of operational capacity during the tidal window.

[0126] like Figure 6 As shown, the trigger window period sprint mode specifically includes:

[0127] S510: Set a rising threshold. When the real-time tide level rising rate exceeds the rising threshold and lasts for 2 minutes, a sprint warning is triggered, including:

[0128] S520, increasing the ferry cruising speed to a maximum value within a safety threshold range;

[0129] S530: Reorganize the unboarded ferry combinations and select vehicles whose destinations are the ferry with the largest unloading volume and whose reservation time tolerance is lower than the reservation time tolerance threshold to give priority to boarding;

[0130] S540: Activate the dynamic berth allocation strategy to shorten the time interval between ships berthing and leaving berth to the minimum safe value.

[0131] S600 utilizes the return window when the tide recedes to build a vehicle path relay network between adjacent ferry crossings, and reuses reverse-flow boat resources across time periods.

[0132] When using the return window during low tide to construct a vehicle route relay network, the system will activate a multi-dimensional data screening mechanism 30 minutes before the low tide window opens. Through a distributed database, it will scan the database of vehicles waiting to be crossed in real time. Combined with the spatial analysis function of the geographic information system (GIS), it will screen out vehicles whose destination coordinates are within the radiation range of adjacent ferry ports and whose reservation time tolerance threshold exceeds the expected duration of the low tide window. At the same time, it will automatically exclude vehicles carrying dangerous goods or with special timeliness requirements, forming a set of relayable objects.

[0133] When constructing the complete graph, in addition to using the locations of each ferry as nodes and the estimated voyage time as edge weights, hydrological data such as the water flow speed and wind direction at the current low tide stage will also be accessed in real time. The estimated sailing time of each section will be dynamically corrected through the fluid mechanics model to ensure the real-time and accuracy of the edge weight calculation.

[0134] The establishment of the optimization allocation model requires comprehensive consideration of the remaining available space on the return ferry, the estimated arrival time at each ferry port, the basic transportation price, and the vehicle priority label (such as ordinary vehicles and heavy-duty trucks). A mixed integer programming algorithm is used to solve the optimal matching solution between vehicles and ferries, generating a matching matrix containing ship ID, vehicle information, and connection time.

[0135] The discount price calculation mechanism needs to combine the number of remaining vacant cabins with the remaining time of the low tide window to establish a tiered pricing model. For example, when the remaining cabins of a ferry exceed 40% of the total capacity, incremental discounts will be provided to subsequent reserved vehicles to improve the utilization rate of empty cabins.

[0136] This step utilizes the low tide window to construct a reverse transport network, enabling the cross-time recycling of shared ferry resources. By pre-screening available relay vehicles, the system can complete transport plan before the tide begins to recede, ensuring that return ferries are fully or nearly fully loaded upon leaving the original port. This significantly reduces the occurrence of empty return trips, effectively reducing unit freight transportation costs, especially on waterways primarily used for freight.

[0137] Dynamically modified edge weights on the complete graph ensure that the optimal route sequence can adapt to changing water flows during low tide, avoiding delays caused by reduced speeds and improving return trip efficiency compared to traditional fixed-route solutions. A tiered discount pricing strategy uses economic leverage to regulate transport demand, guiding users to choose relay transport during low tide. This balances flow distribution within the tidal cycle and alleviates transport pressure during high tide periods.

[0138] like Figure 7 As shown, the construction of the vehicle path relay network specifically includes:

[0139] S610: 30 minutes before the low tide window opens, scan the database of vehicles waiting for ferries, select vehicles whose destination coordinates are adjacent to the ferry and whose reservation time tolerance threshold is greater than the low tide window duration, and mark them as a set of relay objects;

[0140] S620 , constructing a complete graph with all ferry crossings in the region as nodes, where the nodes are the locations of the ferry crossings and the edge weights are the estimated travel times between the ferry crossings, and solving and outputting an optimal route sequence that minimizes the total travel time and has overlapping time windows for each flight segment;

[0141] S630, obtaining current data of each return ferry, including remaining available space, estimated arrival time at each ferry port, and basic transportation price, and establishing an optimization allocation model to allocate a corresponding ferry to each waiting vehicle, thereby generating a vehicle-ferry matching matrix;

[0142] S640: Obtain the available berths and corresponding prices of the return ferry, and calculate a discounted price based on the number of remaining available berths.

[0143] Figure 8 The structural block diagram of the ferry route dynamic optimization system provided by the embodiment of the present invention is as follows: Figure 8 As shown, the system includes:

[0144] The data processing module 100 is used to obtain real-time hydrological monitoring data and historical tidal model data, calculate the time period that meets the water level continuous navigation conditions on the day, and generate a dynamically changing golden navigation window;

[0145] The ferry schedule adjustment module 200 is used to compress the ferry schedule according to the duration of the golden navigation window during the golden navigation window, and dynamically adjust the total number of flights;

[0146] The carpooling combination generation module 300 is used to scan vehicle dimensional characteristics in real time through image recognition and reservation data, including the physical dimensions of the vehicle to be ferried, the GPS coordinates of the destination, the reservation time tolerance threshold, and the cargo type label. Based on the vehicle dimensional characteristics, it generates carpooling combinations that span ferry schedules, thereby building a flexible and scalable water carpooling network.

[0147] The transport plan optimization module 400 is used to output a transport plan for the vehicle-vehicle combination, including optimization of temporary docking point coordinates, dynamic adjustment of connecting routes, and compensation factors for inter-shift connection time;

[0148] The ship-sharing demand coordination module 500 is used to trigger the window period sprint mode when it detects that the tide level rise rate exceeds a preset threshold through a coupling coordination mechanism between the tide window and the ship-sharing demand;

[0149] The reverse boat-pooling module 600 is used to utilize the return window when the tide recedes to build a vehicle path relay network between adjacent ferry crossings and perform cross-time reuse of reverse boat-pooling resources.

[0150] The technical features of the above-mentioned embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above-mentioned embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0151] The above-described embodiments merely illustrate several implementations of the present invention, and while their descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art would be able to make numerous variations and improvements without departing from the spirit of the present invention, all of which fall within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be determined by the appended claims.

[0152] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for dynamic optimization of ferry routes, characterized in that: The method comprises: Obtain real-time hydrological monitoring data and historical tidal model data, calculate the time period that meets the water level conditions for continuous navigation on the same day, and generate a dynamically changing golden navigation window; During the golden navigation window period, the ferry frequency will be shortened according to the duration of the golden navigation window, and the total number of flights will be adjusted dynamically; Through image recognition and reservation data, the system scans vehicle dimensional features in real time, including the physical dimensions of the vehicle to be ferried, the GPS coordinates of the destination, the reservation time tolerance threshold, and the cargo type label. Based on these vehicle dimensional features, it generates vehicle pooling combinations across ferry schedules, building a flexible and scalable water carpooling network. For vehicle-ship combinations, a transportation plan is output that includes optimization of temporary docking point coordinates, dynamic adjustment of connecting routes, and compensation factors for inter-shift connection time. Through the coupling coordination mechanism of tidal windows and ship-sharing demand, the window period sprint mode is triggered when the tide level rise rate exceeds the preset threshold; By utilizing the return window when the tide recedes, a vehicle path relay network is constructed between adjacent ferry crossings to reuse reverse-flow ferry resources across time periods.

2. The method according to claim 1, characterized in that The generating of the dynamically changing golden navigation window specifically includes: Collect real-time water level elevation and flow velocity vectors, integrate astronomical tide forecasts with meteorological typhoon storm surge correction parameters, establish a water level prediction model, and predict the water level change trajectory in the next 4 hours: ； in, For the future The water level at the moment, is the weight coefficient, For the current moment Real-time water level elevation, For the current moment The water velocity vector, is the prediction step length, Time for astronomical tide forecast Water level correction value, is the correction value for the impact of storm surge, is the error term; Set the minimum safe draft for ferries and filter out periods where the water level is continuously higher than the minimum safe draft + 0.5m; Set the time window and safety area, and calculate the water level fluctuation variance in the filtered period. When the water level fluctuation variance is less than the safety area for three consecutive windows, it is determined to be a stable navigation window. By superimposing ship traffic density constraints, a golden navigation window sequence with time margin is generated.

3. The method according to claim 2, characterized in that The intervals mentioned above compress the ferry schedules and dynamically adjust the total number of schedules, specifically including: According to the duration of the golden navigation window Interval with regular shifts , calculate the compressed ferry interval: ； in, is the compressed ferry interval time, is the compression coefficient, is the window duration threshold; Based on the predicted arrival time of the ship's AIS trajectory data, temporary flights are dynamically added until the berth utilization rate reaches 90%; When the remaining time in the window is less than the full voyage time, the short-distance transfer mode is activated and only routes between adjacent ferry ports are operated.

4. The method according to claim 3, characterized in that The generation of a vehicle-ship combination across ferry schedules specifically includes: Construct a spatiotemporal network graph with ferries as mobile nodes, with node attributes including planned arrival and departure times and remaining capacity; Taking the GPS coordinates of the destination as the center, the spatiotemporal density function is constructed by integrating the vehicle reservation time tolerance threshold; The optimization objectives are defined, including minimizing weighted waiting time, minimizing empty mileage, ensuring deck area utilization ≤ 85%, and ensuring that the interval between shifts at the same berth is ≥ the safe berthing time. The cross-shift ship pooling scheme is solved based on the spatiotemporal density function, and a ship pooling combination matrix containing a list of ship IDs and license plate numbers is generated.

5. The method according to claim 4, characterized in that The export transportation plan specifically includes: Based on the GPS coordinate distribution of vehicle destinations in the ship-pooling matrix, the optimal docking center is solved and the temporary docking area is divided: ； in, is the GPS coordinate of the optimal stop, For the The weight of the vehicle model, To point to The Euclidean distance between the vehicle and the destination, is the distance weight index, For the The GPS coordinates of the vehicle's destination; For vehicles that have been assigned to a ship group but have not yet boarded the ship, the delay probability of each vehicle is calculated and a delay threshold is set. If the delay probability exceeds the delay threshold, a backup slot is allocated to the vehicle: ； in, is the delay probability, To predict the time of embarkation, The original scheduled boarding time, is the standard deviation of the delay time, is the normal cumulative distribution function.

6. The method according to claim 4, characterized in that The trigger window period sprint mode specifically includes: Set a rising threshold. When the real-time tide level rise rate exceeds the rising threshold and lasts for 2 minutes, a sprint warning is triggered, including: Increase the ferry's cruising speed to the maximum within safety thresholds; Reorganize the unshipped boat combinations and prioritize boarding for vehicles whose destinations are the ferry with the largest unloading volume and whose reservation time tolerance is lower than the reservation time tolerance threshold. Enable dynamic berth allocation strategy to shorten the time interval between ships’ berthing and unberthing to the safe minimum.

7. The method according to claim 4, characterized in that The constructing of the vehicle path relay network specifically includes: 30 minutes before the low tide window opens, scan the database of waiting vehicles for ferrying, select vehicles whose destination coordinates are adjacent to the ferry and whose reservation time tolerance threshold is greater than the low tide window duration, and mark them as the set of relay objects; A complete graph is constructed with all ferry crossings in the region as nodes, where nodes are the locations of each ferry crossing and edge weights are the estimated travel times between ferry crossings. The optimal route sequence is solved and output to minimize the total travel time and ensure that the time windows of each flight segment overlap. Obtain the current data of each return ferry, including remaining available space, estimated arrival time at each ferry port, and basic transportation price, and establish an optimization allocation model to assign the corresponding ferry to each waiting vehicle, generating a vehicle-ferry matching matrix; Get the available return ferry seats and corresponding prices, and calculate the discounted price based on the remaining available seats.

8. The dynamic optimization system of ferry routes is characterized by: The system comprises: The data processing module is used to obtain real-time hydrological monitoring data and historical tidal model data, calculate the time period that meets the water level conditions for continuous navigation on the same day, and generate a dynamically changing golden navigation window; The ferry schedule adjustment module is used to compress ferry schedules during the golden navigation window period according to the duration of the golden navigation window and dynamically adjust the total number of schedules; The carpooling combination generation module uses image recognition and reservation data to scan vehicle dimensional characteristics in real time, including the physical dimensions of the vehicle to be ferried, the GPS coordinates of the destination, the reservation time tolerance threshold, and the cargo type label. Based on the vehicle dimensional characteristics, it generates carpooling combinations that span ferry schedules, building a flexible and scalable water carpooling network. The transport plan optimization module is used to output a transport plan for vehicle-ship combinations, including optimization of temporary docking point coordinates, dynamic adjustment of connecting routes, and compensation factors for inter-shift connection time; The ship-sharing demand coordination module is used to trigger the window period sprint mode when the tide rise rate exceeds the preset threshold through the coupling coordination mechanism of the tide window and the ship-sharing demand; The reverse boat-pooling module is used to utilize the return window when the tide recedes to build a vehicle path relay network between adjacent ferry crossings and reuse reverse boat-pooling resources across time periods.

Images