Channel scheduling method and storage medium

Abstract

The invention discloses a channel scheduling method and a storage medium, and the method comprises the steps: building a network flow-discretization ship kinematics model based on channel parameters which comprise ship lock parameters and ship channel parameters; first real-time data of the ship arriving at the ship lock is obtained in real time, a queuing strategy is obtained, and the first real-time data comprises ship arrival time, departure port information and destination port information; generating a ship lock scheduling plan based on the network flow-discrete ship kinematics model and a queuing strategy; and if the ship leaves the ship lock, generating a ship lane scheduling plan based on the network flow-discretization ship kinematics model. According to the method, the channel topology is abstracted into the network flow model with capacity constraint, and discretization simulation is carried out on ship microcosmic motion in a superposition manner, so that multi-scale coupling of macroscopic scheduling and microcosmic behaviors can be realized, the channel congestion prediction error rate is reduced, and the scheduling efficiency is improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of waterway safety, and particularly relates to a waterway scheduling method and a storage medium thereof. BACKGROUND

[0002] Inland river shipping is an important way to realize the connection between coastal ports and inland areas, and has an important influence on the economic growth of the river basin area and the development of international trade. However, the traffic capacity of inland waterways and ship locks is limited, which has become the main bottleneck restricting the development of inland waterway transportation. In addition, the scheduling of ship locks and waterways of most water conservancy hubs is independently decided, which further limits the improvement of ship navigation efficiency.

[0003] In the related art, the traditional ship lock scheduling method can only be estimated by a model or calculated by relying on historical data, resulting in a high error rate of waterway congestion prediction and low scheduling efficiency. SUMMARY

[0004] The purpose of the application is to realize multi-scale coupling of macroscopic scheduling and microscopic behavior by combining network flow and discretization, which can effectively reduce the error rate of waterway congestion prediction and improve the scheduling efficiency.

[0005] In order to achieve the above purpose, the application provides a waterway scheduling method, which comprises the following steps: constructing a network flow-discretized ship kinematics model based on waterway parameters, wherein the waterway parameters comprise ship lock parameters and ship channel parameters; obtaining first real-time data of ship arrival at a ship lock in real time to obtain a queuing strategy, wherein the first real-time data comprises ship arrival time, departure port and destination port information; generating a ship lock scheduling plan based on the network flow-discretized ship kinematics model and the queuing strategy; and if the ship leaves the ship lock, generating a ship channel scheduling plan based on the network flow-discretized ship kinematics model.

[0006] In an optional embodiment, the network flow-discretized ship kinematics model is constructed based on waterway parameters, and specifically comprises the following steps: topologically expanding the waterway into a network flow waterway model based on the waterway parameters, wherein the nodes of the network flow waterway model represent ship locks or bridges, and the edges represent waterway segments, the network flow waterway model has a capacity constraint, the capacity constraint comprises a ship lock parameter constraint, and the ship channel parameters comprise ship static parameters and ship dynamic parameters; performing discretized simulation on ship movement to obtain a discretized simulation model; and constructing the network flow-discretized ship kinematics model based on the network flow waterway model and the discretized simulation model.

[0007] In one optional implementation, the discretized simulation model is a spatial discretized model. The discretized simulation of ship motion is performed to obtain the discretized simulation model, which specifically includes: discretizing the channel space into cells based on cellular automata; defining cell states and neighbor relationships based on the cells to obtain the discretized simulation model, wherein the cell states include idle, occupied, and prohibited.

[0008] In one optional implementation, the first real-time data of a ship arriving at the lock is acquired in real time to obtain a queuing strategy, specifically including: obtaining the channel congestion status based on the first real-time data; dynamically switching the queuing mode based on the channel congestion status to obtain the queuing strategy, wherein the queuing mode includes a FIFO mode and a priority queue mode, and the priority of the priority queue mode is determined according to the ship type or urgency level.

[0009] In one optional implementation, generating a lock scheduling plan based on the queuing strategy specifically includes: dynamically generating a lock sequence scheme, a vessel staging scheme, and a lock chamber allocation scheme based on the queuing strategy; generating the lock scheduling plan based on the dynamically generated lock sequence scheme, vessel staging scheme, and lock chamber allocation scheme, wherein the lock scheduling plan includes a lock objective function and lock constraints, the lock constraints satisfying draft compatibility constraints, and the lock objective function being: minimizing the average vessel passage time through the lock. In the formula, The total number of ships during the scheduling period. For the first The time it takes for a ship to pass through the lock; maximizing the utilization rate of the lock chamber. In the formula, For the first The area occupied by a ship in the lock chamber. This represents the total usable area of ​​the gate chamber.

[0010] In one optional implementation, the channel scheduling plan includes: a following rule: if the vessel spacing meets the following conditions... Then deceleration is triggered; where, For ship spacing, For safe length, The safe distance for a vessel's speed at time t; Overtaking limit: Overtaking is permitted within the designated overtaking zone, while maintaining a safe lateral distance. In the formula, Define the ship's width; set the objective function for the waterway: equalize the traffic flow at the waterway nodes: ; In the formula, The number of key waterway nodes. For nodes Traffic flow The average traffic flow of all nodes.

[0011] In an optional embodiment, the scheduling rule is optimized online based on reinforcement learning, and the reward function is set as: ; wherein, to minimize the average passing time of the ship, to maximize the utilization rate of the lock chamber, is a weight parameter automatically generated in the reinforcement learning process, which is dynamically adjusted according to the seasonal flow pattern.

[0012] In an optional embodiment, the waterway scheduling method further comprises: setting a parking area scheduling rule for a one-way navigation bridge area, specifically comprising: obtaining a first time that a downlink anchorage waiting ship travels to an upstream prohibited meeting area; obtaining a first average exit time of the ship at the upstream hub ship lock after exiting the lock; based on the first time, the first average exit time and the second time, a time after the upstream hub ship lock is opened is calculated, and the ship is controlled not to leave the anchorage and travel downstream; wherein, a third time that the uplink anchorage waiting ship travels to the downstream prohibited meeting area is obtained; obtaining a second average exit time of the ship at the downstream hub ship lock after exiting the lock; based on the third time, the second average exit time and the fourth time, a time after the downstream hub ship lock is opened is calculated, and the ship is controlled not to leave the anchorage and travel upstream; wherein, ;

[0013] In an optional embodiment, the waterway scheduling method further comprises: obtaining second real-time data of the ship leaving the ship lock, the second real-time data comprising ship position and queuing time; the ship position update frequency is Δt≤the first time; if the queuing time delay exceeds a first threshold, a yellow warning is triggered; if the queuing time delay exceeds a second threshold, a red warning is triggered and an emergency channel is started; wherein, the first threshold is greater than the second threshold.

[0014] The beneficial effects of the present application are: the present application can realize multi-scale coupling of macroscopic scheduling and microscopic behavior by abstracting the channel topology into a network flow model with capacity constraints and superimposing discrete simulation on the microscopic motion of ships, solve the contradiction between system-level resource allocation and individual-level collision avoidance rules under the traditional single model, reduce the channel congestion prediction error rate, and improve the scheduling efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0015] Figure 1 A flowchart of a channel scheduling method provided by an embodiment of the present application is shown. Figure 2 A flowchart of a channel scheduling method provided by another embodiment of the present application is shown. DETAILED DESCRIPTION

[0016] The traditional method of ship scheduling uses a single objective function, such as lock chamber utilization rate or waiting time, without considering the time waste of reverse lock operation, the difference in ship emergency level, etc. Complex constraints, and without joint modeling of channel navigation rules and ship lock scheduling. For example, the existing channel model does not include the berthing area in the dynamic scheduling range, resulting in too long waiting time of ships in the one-way navigation bridge area.

[0017] The present application can solve the multi-objective optimization problem in multi-ship lock cooperative scheduling, simulate the channel traffic flow by cellular automata, optimize the lock chamber arrangement by combining the two-dimensional packing algorithm, and realize the global optimization of the ship lock time, transit time and lock chamber utilization rate. Especially for the one-way navigation bridge area, design the berthing area scheduling rules to reduce the invalid waiting of ships in the channel.

[0018] The present application will be further described in detail through the drawings and specific embodiments.

[0019] As shown in Figure 1 and Figure 2 According to an embodiment of the present application, on the one hand, a channel scheduling method is provided, comprising the following steps: Step S101: constructing a network flow-discrete ship kinematics model based on channel parameters, the channel parameters including ship lock parameters and ship channel parameters.

[0020] Step S103: obtaining first real-time data of ships arriving at the ship lock in real time to obtain a queuing strategy, the first real-time data including ship arrival time, departure port and destination port information.

[0021] Step S105: generating a ship lock scheduling plan based on the network flow-discrete ship kinematics model and the queuing strategy.

[0022] Step S107: if the ship leaves the ship lock, generating a ship channel scheduling plan based on the network flow-discrete ship kinematics model.

[0023] In this embodiment, the lock parameters include lock chamber size, gate opening and closing time, traffic capacity, ship transit time, total available area of the lock chamber, etc., and the channel parameters include ship static parameters, ship dynamic parameters, channel width, water depth, curvature, flow rate, etc. The ship static parameters mainly include ship length, ship width, draft depth, and the ship dynamic parameters mainly include expected speed, acceleration limit, braking distance threshold. Based on these parameters, a network flow model is constructed, which discretizes the channel into multiple nodes and edges, with nodes representing key locations such as lock chamber entrances and channel junctions, and edges representing ship navigation paths. At the same time, a dynamic behavior model of the ship in the channel is established by combining the ship kinematics model, such as ship acceleration, deceleration, turning radius, etc.

[0024] By discretizing the channel and introducing the kinematic characteristics of the ship, the dynamic behavior of the ship in the channel can be more accurately simulated, avoiding the scheduling errors caused by traditional static models. The network flow model can convert the complex channel scheduling problem into a graph theory problem, which is convenient for solving by using efficient optimization algorithms. Combined with the physical limitations of the lock and channel, it ensures that the scheduling scheme is feasible in the actual environment, reduces congestion and conflicts.

[0025] Real-time information such as ship arrival time, departure port, destination port, ship size, and cargo type is obtained through the Automatic Identification System (AIS), radar, or the Arrival module of the port management system. Real-time data updates ensure that the scheduling system can adapt to the uncertainty of ship arrival and reduce scheduling failures caused by information lag. By using a reasonable queuing strategy, the waiting time of different ships can be balanced, which can improve the overall traffic efficiency. Automatic data collection and queuing strategy generation can reduce the dependence on manual scheduling and reduce human errors.

[0026] By combining the queuing strategy with the network flow-discretization model, an optimization algorithm such as integer programming or heuristic algorithm is used to allocate lock chambers and transit time windows for ships, taking into account factors such as ship size, draft depth, and lock chamber capacity, to generate the optimal order and schedule for ships passing through the lock. By optimizing ship grouping and lock chamber allocation, the idle time of the gate can be reduced, and the throughput of the lock can be improved. A reasonable scheduling plan can shorten the ship's detention time before the lock and improve shipping efficiency. By using the discretization model to accurately calculate the ship's motion trajectory, collisions or blockages in the lock area can be prevented. After the ship leaves the lock, the channel segment and navigation speed are allocated to the ship based on the real-time state of the channel and the ship kinematics model, and the optimal path is planned to ensure safe navigation. The real-time state of the channel can include the positions of other ships and the occupancy of the channel. By dynamically detecting the space-time conflicts between ships, rear-end collisions and other accidents can be avoided, and the allocation of ship speed and path can be optimized to reduce channel congestion, especially in narrow or busy sections. By planning a reasonable speed, fuel consumption can be reduced, meeting the demand for green shipping.

[0027] In addition, based on the channel node topology, such as branch channel intersection, anchorage position, the path feasibility can be checked.

[0028] Further, in step S101, a network flow-discrete ship kinematics model is constructed based on the channel parameters, specifically including the following steps: Step S1011: Based on the channel parameters, the channel topology is converted into a network flow channel model, the nodes of the network flow channel model represent ship locks or bridges, and the edges represent channel segments. The network flow channel model has capacity constraints, including ship lock parameter constraints. The channel parameters include static ship parameters and dynamic ship parameters.

[0029] Step S1013: Discretize the ship motion to obtain a discretized simulation model.

[0030] Step S1015: Based on the network flow channel model and the discretized simulation model, a network flow-discrete ship kinematics model is constructed.

[0031] The ship locks, bridges and intersections in the channel are abstracted as nodes of a directed graph, and each sub-node can carry a 3D attribute. After the network flow model generates an initial path for the ship, the discretized model performs fine-grained simulation in key areas, such as ship lock entrances. If micro-simulation finds conflicts, such as channel segment congestion, the macro layer is triggered to re-plan the path. Understandably, the network flow model handles global path planning and flow allocation, and the discretized model handles local collision avoidance and motion control.

[0032] The upper limit of the ship lock capacity can be calculated based on the maximum queuing theory, such as the M / M / 1 queuing model, and the flow conservation constraint in graph theory is used to ensure that the ship flow does not exceed the edge capacity, such as the maximum number of parallel ships in the channel segment. The physical channel is converted into a calculable network flow model, supporting path planning, congestion detection and other algorithms. Capacity constraints can prevent the model from generating scheduling schemes that exceed the actual navigation capacity.

[0033] This scheme takes into account the global scheduling efficiency and local motion authenticity, is suitable for large-scale channel simulation, supports "planning-simulation-feedback" closed-loop optimization, and can improve the feasibility of scheduling schemes. The network flow model and the discretized simulation achieve complementarity in macro and micro, solving the problem that traditional methods cannot handle flow optimization and motion details at the same time. By updating the network flow capacity constraints in real time, the robustness of the model is enhanced.

[0034] Based on step S1013, the discretized simulation model is a spatial discretization model, which discretizes the ship motion to obtain a discretized simulation model, specifically including the following steps: Step S10131: Discretize the channel space into cells based on cellular automata.

[0035] Step S10133: Based on the cell definition cell state and neighbor relationship, get the discrete simulation model, the cell state includes free, occupied and forbidden.

[0036] Basic environment construction: Based on cellular automata (Cellular Automata, CA), the channel space is discretized into cells, the typical size is 1.1-1.5 times the length of the ship, the cell state and the neighbor relationship are defined, and the neighbor relationship can adopt Moore type or Von Neumann type neighborhood.

[0037] Ship attribute loading: Import ship static parameters and dynamic parameters, ship static parameters include ship length, ship width, draft, ship dynamic parameters include expected speed, acceleration limit, braking distance threshold, establish ship kinematics model, ship speed update formula is as follows vn ( t +1)=min( v max, vn ( t )+ a Δ t , dn ( t ) / τ ); In the formula, is the front ship distance, is the maximum ship speed, is the real-time ship speed, τ is the safety time interval, is the acceleration, is the discrete time step.

[0038] Ship lock rule configuration: Set the size of the lock chamber: length x width x sill depth, gate opening and closing time, maximum capacity of single lock based on plane arrangement algorithm, etc. Constraint conditions.

[0039] By decomposing the channel topology into a series of parallel cell spaces, such as upstream and downstream lock groups, using a distributed task scheduling strategy, the single node computing load can be reduced, and the response time of multi-ship lock cooperative scheduling can be shortened by 30%-50%.

[0040] Of course, it can also be through the discrete event simulation (DES), which discretizes the channel into an event queue, such as ship arrival and lock departure, and simulates ship movement based on event-driven, instead of the space-time discretization modeling of cellular automata.

[0041] It can also be through Agent-Based modeling (ABM), which defines autonomous decision rules for each ship, such as path selection and speed adjustment, and realizes traffic flow simulation through multi-agent interaction, which can avoid the fixed limitations of cell space division.

[0042] Based on step S103, the first real-time data of the ship arriving at the ship lock is acquired in real time, and a queuing strategy is obtained, specifically including the following steps: Step S1031: Acquire the channel congestion state based on the first real-time data.

[0043] Step S1033: Dynamically switch the queuing mode based on the channel congestion state, and obtain the queuing strategy. The queuing mode includes FIFO mode and priority queue mode. The priority of the priority queue mode is determined according to the ship type or the emergency degree.

[0044] The channel congestion state is expressed by whether the node queuing length exceeds the threshold Lqueue-max. If the queuing mode is dynamically switched, for example, FIFO / priority queue, the priority is determined according to the ship type or the emergency degree, for example, the priority according to the ship type is: dangerous chemical ship>passenger ship> cargo ship, or the priority is determined according to the emergency degree. Based on these data, a queuing strategy is generated by using queuing theory or dynamic priority algorithm, such as first-come-first-served, emergency ship priority, large ship group passage, etc.

[0045] The ship position, speed and destination can be received in real time, and the update frequency is ≥1 time / minute. The lock chamber occupancy state and the ship queuing length can be acquired by a laser scanner or an RFID. Meteorological and hydrological information, such as wind speed>6, can also be integrated to trigger the navigation capacity degradation. The total length of the queuing ship or the length of the channel buffer area is triggered to alarm when the threshold is ≥80%. The actual passing time / theoretical passing time is determined as congestion if the threshold is ≥1.5. The LSTM neural network can be used to predict the congestion trend in the next 15 minutes. The congestion state can be classified, such as smooth / mild / severe, with an accuracy of ≥90%, and the edge weight of the dynamic network flow model can be supported, such as the cost coefficient of the congested channel segment is automatically increased.

[0046] If the congestion level is ≤mild (space congestion rate<60%), the FIFO mode is triggered.

[0047] If the congestion level is ≥severe or an emergency event such as emergency medical transport, the priority queue is triggered.

[0048] In the priority mode, part of the FIFO queue can also be reserved, for example, 30% of the lock times are allocated to ordinary ships. The regret value mechanism can also be used: for example, the ship that is delayed for 3 times in a row is automatically promoted in priority.

[0049] In this embodiment, the dynamic strategy improves the channel passing efficiency by 18%-32% compared with the fixed strategy, and also supports the compliance with regulations, such as the mandatory guarantee of priority ships in relevant laws and regulations.

[0050] Further, in step S105, the ship lock scheduling plan is generated based on the queuing strategy, specifically including the following steps: Step S1051: dynamically generating the lock time scheme, the ship arrangement scheme and the lock chamber allocation scheme based on the queuing strategy; Step S1053: generating the ship lock scheduling plan based on the state generation lock time scheme, ship arrangement scheme and lock chamber allocation scheme, the ship lock scheduling plan including a ship lock objective function and a ship lock constraint condition, the ship lock constraint condition satisfying the draft compatibility constraint, and the ship lock objective function being: minimizing the average ship transit time: ; In the formula, is the total number of ships in the scheduling period, is the transit time of the i-th ship; maximizing the lock chamber utilization rate: ; In the formula, is the area occupied by the i-th ship in the lock chamber, is the total available area of the lock chamber.

[0051] In this embodiment, overtaking is only allowed in designated overtaking areas such as straight wide channels.

[0052] Lock chamber arrangement optimization: based on a two-dimensional packing algorithm such as the BL algorithm, a lock time scheme can be dynamically generated, with the objective function being to maximize the lock chamber area utilization rate while satisfying the draft compatibility constraint, i.e., the draft difference of ships in the same lock time is Δh≤htolerance.

[0053] Improved BL (Bottom-Left) algorithm for dynamic packing algorithm with draft depth constraint, preferentially matching ships with similar drafts to reduce lock chamber water level adjustment energy consumption.

[0054] Introducing a two-dimensional packing algorithm combined with dynamic programming in lock chamber arrangement reduces the complexity of ship layout problem from O(n!) to O(n2), and the generation time of single lock time scheduling scheme is controlled within 5 seconds, meeting the real-time demand.

[0055] Genetic algorithm (GA) layout can also be used, encoding ship positions as chromosomes, and iteratively optimizing the layout scheme through a fitness function including area utilization rate and draft difference, instead of the BL heuristic algorithm.

[0056] Deep learning prediction can also be used to train a convolutional neural network (CNN) to predict the optimal arrangement pattern of ships, and to realize end-to-end decision-making using historical scheduling data, bypassing the traditional optimization calculation process. ​​

[0057] The super parameters of the NSGA-II algorithm, such as a cross rate and a mutation rate, are trained offline through historical data, and only a pre-optimized parameter library needs to be loaded during online inference, so that the number of multi-objective optimization iterations is reduced by 40%, and the consumption of computing resources is reduced by 25%.

[0058] In the embodiment, microscopic traffic flow modeling is performed: in the Channel module, an improved car-following model and a collision avoidance rule are used.

[0059] The channel scheduling plan comprises: The car-following rule is: if the distance between ships satisfies , then deceleration is triggered; in the formula, is the distance between ships, is a safety length, is a safety time interval of the ship speed at time t, and in the discretized simulation, the safety time interval is defined as The overtaking limit is: overtaking is performed in a designated overtaking area, and a transverse safety distance is satisfied; in the formula, is a ship width; The channel objective function is set as: The equilibrium channel node traffic flow is: ; ; In the formula, is the number of key channel nodes, is the traffic flow of the node , and is the average traffic flow of all nodes.

[0060] The NSGA-II algorithm with an elite reservation strategy is adopted, the utilization rate f2 of the lock chamber, the ship delay time f1 and the energy consumption index f3 are simultaneously optimized, a non-dominated solution set is output for decision selection, the lock chamber space utilization rate ≥ 95% can be realized, and the single lock scheduling scheme generation time is compressed to the order of 10 seconds, while the traditional method needs minutes.

[0061] The improved NSGA-II algorithm, that is, the non-dominated sorting genetic algorithm with an elite reservation strategy, can be used, and the specific implementation is as follows: The coding mode: the ship lock passing order and the lock distribution scheme are represented by real number coding.

[0062] Crossing and mutation: two-point crossing is combined with a dynamic mutation rate, that is, the mutation rate is adaptively adjusted with the iteration number.

[0063] Pareto solution screening: the diversity solution set is reserved by using a crowding degree comparison operator, and finally the decision maker selects an implementation scheme based on weights.

[0064] Computational mode separation: the offline training phase optimizes model parameters, such as the tau value of the car-following model, using historical data, and the online inference phase loads the parameter library to realize real-time scheduling.

[0065] The ship kinematics model based on the cellular automaton can accurately simulate the dynamic behaviors of ship acceleration and braking distance, so that the error rate of waterway congestion prediction is less than or equal to 8%, which is more than 60% higher than that of a traditional macro model.

[0066] The Pareto frontier screening mechanism is adopted to simultaneously optimize the lock chamber utilization rate f2 and the average lock passing time f1, experiments show that the lock chamber area utilization rate can reach 92%-95%, while the traditional method is 80%-85%, and the average waiting time of the ship for locking is shortened by 20%-30%.

[0067] The MOEA / D algorithm can also be used to decompose the multi-objective problem into multiple single-objective sub-problems, and the non-dominated sorting mechanism of NSGA-II is replaced by the weight vector allocation for collaborative optimization.

[0068] Particle swarm optimization (PSO) can also be used, which uses a particle swarm velocity update formula to replace the crossover and mutation operations of the genetic algorithm, reducing the complexity of the algorithm and being suitable for high-dimensional optimization scenarios.

[0069] Further, the scheduling rules are optimized online based on reinforcement learning, and the reward function is set as: ; In the formula, is to minimize the average lock passing time of the ship, is to maximize the lock chamber utilization rate, is a weight parameter automatically generated in the reinforcement learning process, which is dynamically adjusted according to the seasonal flow pattern.

[0070] The invention constructs a task decoupling mechanism between ship lock groups, and the upstream and downstream ship lock calculation units independently execute local scheduling, including lock chamber allocation and ship arrangement, and the central coordinator dynamically adjusts the global priority based on the flow balance principle, forming a hybrid control structure of “distributed decision + centralized correction”, which can break through the scalability bottleneck of the traditional centralized computing architecture, support real-time collaboration of hundreds of ship lock groups, and improve the system throughput by 3-5 times.

[0071] Further, the waterway scheduling method further includes the following steps: Step S201: obtaining second real-time data of the ship leaving the ship lock, the second real-time data including the ship position and the queuing time.

[0072] Step S202: the ship position update frequency is Δt≤the first time.

[0073] Step S203: If the delay time of the queuing duration exceeds the first threshold value, a yellow warning is triggered.

[0074] Step S204: If the delay time of the queuing duration exceeds the second threshold value, a red warning is triggered and an emergency channel is started.

[0075] The second threshold value is greater than the first threshold value.

[0076] In this embodiment, the hierarchical warning scheduling process is realized based on real-time data. When there is a delay in the ship passing through the lock, different levels of warning and emergency measures can be triggered in time through dynamic monitoring and threshold judgment, so as to relieve congestion and improve navigation efficiency. The first time can be 10 seconds, synchronized with the time step of the cellular automaton. The first threshold value can be 30 minutes, and the second threshold value can be one hour. The yellow warning is triggered when the queuing delay exceeds 30 minutes, and the red warning is triggered when the queuing delay exceeds one hour.

[0077] Further, the waterway scheduling method further comprises: setting a parking area scheduling rule for a one-way navigation bridge area, specifically comprising the following steps: Step S301: Obtain the first time for a downlink anchorage waiting ship to travel to an upstream prohibited meeting area .

[0078] Step S302: Obtain the first average exit time of the upstream hub ship lock for two-way passing .

[0079] Step S303: Obtain the second time for a ship to travel to an upstream prohibited meeting area after exiting the upstream hub ship lock .

[0080] Step S304: Calculate the time after the upstream hub ship lock upstream gate is opened based on the first time, the first average exit time and the second time , control the ship not to leave the anchorage and travel downstream.

[0081] The time after the upstream hub ship lock upstream gate is opened .

[0082] Step S305: Obtain the third time for an uplink anchorage waiting ship to travel to a downstream prohibited meeting area .

[0083] Step S306: Obtain the second average exit time of the downstream hub ship lock for two-way passing .

[0084] Step S307: Obtain the fourth time for a ship to travel to a downstream prohibited meeting area after exiting the downstream hub ship lock .

[0085] Step S308: calculating the time after the downstream hub ship lock downbound lock gate is opened based on the third time, the second average lock-out time and the fourth time , the ship is not allowed to leave the anchorage and sail upstream.

[0086] wherein, .

[0087] In this embodiment, the definition #1 anchorage is the upbound anchorage, which is about 3.7 km away from the upstream A hub. When all the upbound ships in the A hub ship lock parking area enter the lock chamber, the A hub lock scheduling team informs the anchorage waiting ships to be allowed to sail upbound through very high frequency. The definition #2 anchorage is the downbound anchorage, which is about 1.6 km away from the upstream prohibited meeting area of the downstream B hub, about 3.9 km away from the A hub, about 3.0 km away from the upstream prohibited meeting area of the B hub after the A hub lock, and about 6.4 min is needed for the #2 anchorage waiting ships to sail to the upstream prohibited meeting area of the B hub, about 4.1 min is the average lock-out time of the A hub lock for two-way passing, and about 16.4 min is needed for the ships to sail to the upstream prohibited meeting area of the B hub after the A hub lock is out. Therefore, in order to avoid the ships meeting in the B hub prohibited meeting area, the A hub lock scheduling team is not allowed to leave the anchorage and sail downstream within 15 min after the A hub upbound lock gate is opened for 4.1+16.4-6.1=14.4 min.

[0088] The definition #3 anchorage is the upbound anchorage, which is about 3.0 km away from the downstream prohibited meeting area of the C hub, and about 3.4 km away from the downstream prohibited meeting area of the C hub after the C hub lock is out. Since there is a navigation scheduling anchorage upstream of the C hub, the #3 anchorage only needs to avoid the ships meeting in the downstream prohibited meeting area of the bridge area. It takes about 16.4 min for the #3 anchorage waiting ships to sail to the downstream prohibited meeting area of the C hub, about 4.1 min is the average lock-out time of the C hub lock for two-way passing, and about 13.6 min is needed for the ships to sail to the downstream prohibited meeting area of the C hub after the C hub lock is out. Therefore, the C hub lock scheduling team is not allowed to leave the anchorage and sail upstream within 5 min after the C hub downbound lock gate is opened for 4.1+13.6-16.4-1.3 min.

[0089] Through this embodiment, the number of overall passing ships is increased from 18-25 per hour to 50-60.

[0090] In another aspect, the present application also provides an electronic device, comprising: at least one processor; a memory connected with the at least one processor in communication; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform any of the channel scheduling methods and the network-constructed converter small signal stability margin evaluation method of claim 14.

[0091] In another aspect, the present application also provides a computer storage storage medium, storing a computer program, and the computer program is executed by a processor to implement any of the channel scheduling methods.

[0092] The computer storage storage medium can be simply referred to as a storage medium. Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be completed by a computer program instructing related hardware. The program can be stored in a non-volatile computer readable storage medium, and when the program is executed, it can include the processes of the above-mentioned embodiments. Any reference to memory, storage, database or other storage medium used in the embodiments provided by the present application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM). Each embodiment in the specification is described in a progressive manner, and the same or similar parts of each embodiment can be referred to. Each embodiment focuses on the differences from other embodiments. In particular, for device, equipment, non-volatile computer storage storage medium embodiments, because they are basically similar to method embodiments, the description is relatively simple, and the relevant parts can be referred to the part of the method embodiment.

[0093] The above embodiments are merely exemplary and are not intended to limit the embodiments. Based on the above description, those skilled in the art can further make modifications and variations to the embodiments. The modifications and variations do not depart from the scope of the embodiments.

Claims

1. A waterway scheduling method, characterized in that, include: A network flow-discrete ship kinematics model is constructed based on channel parameters, including lock parameters and channel parameters. The first real-time data of ships arriving at the lock is obtained to develop a queuing strategy. The first real-time data includes information on ship arrival time, port of departure, and port of destination. A lock scheduling plan is generated based on the network flow-discrete ship kinematics model and the queuing strategy. If a vessel leaves the lock, a channel scheduling plan is generated based on the network flow-discrete ship kinematics model.

2. The waterway scheduling method according to claim 1, characterized in that, A network flow-discrete ship kinematics model is constructed based on channel parameters, specifically including: Based on the channel parameters, the channel topology is converted into a network flow channel model. The nodes of the network flow channel model represent locks or bridges, and the edges represent channel segments. The network flow channel model has capacity constraints, including lock parameter constraints. The channel parameters include ship static parameters and ship dynamic parameters. Discretize the ship's motion to obtain a discretized simulation model; The network flow-discrete ship kinematics model is constructed based on the network flow channel model and the discretized simulation model.

3. The waterway scheduling method according to claim 2, characterized in that, The discretized simulation model is a spatial discretized model. By discretizing and simulating the ship's motion, a discretized simulation model is obtained, specifically including: The channel space is discretized into cells based on cellular automata; Based on the cell definition, cell states and neighbor relationships are obtained, and the discretized simulation model is obtained. The cell states include idle, occupied, and prohibited.

4. The waterway scheduling method according to any one of claims 1 to 3, characterized in that, Real-time data on the arrival of ships at the lock is acquired to derive a queuing strategy, which includes: The channel congestion status is obtained based on the first real-time data; The queuing strategy is obtained by dynamically switching queuing modes based on the channel congestion status. The queuing modes include FIFO mode and priority queue mode. The priority of the priority queue mode is determined according to the ship type or urgency level.

5. The waterway scheduling method according to any one of claims 1 to 3, characterized in that, The lock scheduling plan is generated based on the queuing strategy, specifically including: Based on the queuing strategy, the gate sequence scheme, vessel arrangement scheme, and lock chamber allocation scheme are dynamically generated. The lock scheduling plan is generated based on the state generation lock sequence scheme, vessel arrangement scheme, and lock chamber allocation scheme. The lock scheduling plan includes a lock objective function and lock constraints. The lock constraints satisfy draft compatibility constraints. The lock objective function is: Minimize the average lock passage time for ships: ； In the formula, The total number of ships during the scheduling period. For the first The time it takes for a ship to pass through the lock; Maximize the utilization rate of the gate chamber; ； In the formula, For the first The area occupied by a ship in the lock chamber. This represents the total usable area of ​​the gate chamber.

6. The waterway scheduling method according to any one of claims 1 to 3, characterized in that, The channel scheduling plan includes: Following rule: If the distance between vessels meets the following conditions... Then deceleration is triggered; where, For ship spacing, For safe length, The safe time interval for the ship's speed at time t; Overtaking Restriction: Overtaking is permitted within the designated overtaking zone, provided that the lateral safety distance is maintained. In the formula, For the width of the boat; Set the objective function for the navigation channel: Equalizing traffic flow at waterway nodes: ； ； In the formula, The number of key waterway nodes. For nodes Traffic flow This represents the average throughput across all nodes.

7. The waterway scheduling method according to any one of claims 1 to 3, characterized in that, Based on reinforcement learning, the online optimization scheduling rules are set, and the reward function is: ； In the formula, To minimize the average lock passage time for ships, To maximize the utilization rate of the gate chamber, These are weight parameters automatically generated during reinforcement learning and dynamically adjusted according to seasonal traffic patterns.

8. The waterway scheduling method according to any one of claims 1 to 3, characterized in that, Also includes: The following are the specific rules for setting up berthing area scheduling for one-way navigation bridge areas: Obtain the first time that a vessel waiting at the downstream anchorage reaches the upstream prohibited passage zone. ; Obtain the first average exit time for bidirectional passage through the upstream hub lock. ; Obtain the second time after the vessel exits the upstream lock and proceeds to the upstream prohibited passage zone. ; The time after the upstream gate of the upstream hub ship lock was opened was calculated based on the first time, the first average exit time, and the second time. The vessel must not leave the anchorage and sail downstream; Among them, the time after the upstream lock gate of the upstream hub ship lock opens ; Obtain the third time when the upstream anchorage vessel reaches the downstream prohibited passage zone. ; Obtain the second average exit time for bidirectional passage through the downstream hub lock. ; Obtain the fourth time after the vessel exits the downstream hub lock and proceeds to the downstream prohibited passage zone. ; The time after the downstream lock gate of the downstream hub ship lock opens is calculated based on the third time, the second average exit time, and the fourth time. Vessels are prohibited from leaving the anchorage and sailing upstream. in, .

9. The waterway scheduling method according to any one of claims 1 to 3, characterized in that, Also includes: Acquire second real-time data on the departure of the vessel from the lock, including the vessel's position and queuing time. The ship's position update frequency is Δt ≤ first time; If the delay in queuing exceeds the first threshold, a yellow alert will be triggered; If the delay in queuing exceeds the second threshold, a red alert is triggered and the emergency channel is activated. The first threshold is greater than the second threshold.

10. A storage medium, characterized in that, The system contains a computer program that, when executed by a processor, implements the waterway scheduling method as described in any one of claims 1 to 9.

Images