A port ship traffic organization optimization and regulation method

Abstract

The application discloses a port ship traffic organization optimization and regulation method, and belongs to the technical field of waterway traffic optimization and intelligent management and control. The method comprises the following steps: collecting port data, constructing an information physical system framework, using the information physical system framework to realize dynamic management of the port, uniformly fusing features after extraction and quantization, using an event triggering mechanism to obtain a judgment result, updating a system state, using a model prediction control algorithm to predict the system step length, using an adaptive horizon mechanism to obtain a real-time step length, using a multiplier alternating direction method to solve according to the judgment result and the real-time step length, and obtaining a port ship traffic scheduling control strategy. The application can realize efficient ship traffic scheduling and resource optimization in a complex and changeable port environment, significantly improve port transportation efficiency, reduce traffic congestion, and enhance the adaptive ability of the system.

Description

Technical Field

[0001] This invention relates to the field of waterway traffic optimization and intelligent control technology, and in particular to a method for optimizing and controlling port vessel traffic organization based on dynamic optimization theory and intelligent scheduling. Background Technology

[0002] In recent years, with the acceleration of globalization and the surge in international trade, ports have become crucial hubs in the global transportation network. The efficient operation of their transportation systems plays a vital role in the stability of international logistics chains and the development of the port economy. According to data from the United Nations Conference on Trade and Development, over 80% of global international trade is conducted via maritime transport. As the core node of maritime logistics, the operational efficiency of ports directly determines the smoothness and economic benefits of the global supply chain. The rapid development of cutting-edge technologies such as artificial intelligence, the Internet of Things, and big data has greatly promoted the intelligent and optimized integration of data-driven port facilities. Through this deep integration, data has transformed from its previous static storage form into a real-time, dynamic basis for decision-making, thus laying a more reliable technological foundation for the optimization of intelligent transportation systems and the efficient operation of smart ports.

[0003] Against the backdrop of intelligent port development, port waterway traffic operations exhibit significant dynamic characteristics. Intelligent development has significantly enhanced the perception capabilities of port waterway traffic operations, greatly strengthening the collaborative perception capabilities of physical equipment and information systems. The significantly improved real-time data acquisition and processing capabilities based on big data and artificial intelligence technologies have also significantly enhanced the modeling and prediction capabilities of port waterway traffic operations. However, intelligent transformation also brings many challenges. First, the complexity of port waterway traffic operations has increased significantly, involving dynamic interactions across multiple dimensions such as ship motion, channel conditions, meteorological environment, and port operation arrangements. Existing modeling methods mostly focus on the independent analysis of ship behavior, environmental factors, and traffic flow. However, when faced with multi-factor interactions and time-varying characteristics, these methods often fail to capture the comprehensive dynamic characteristics of the system. Furthermore, port waterway traffic control involves not only ship navigation behavior but also channel congestion, inter-ship interactions, and port operation scheduling. Traditional control methods, such as fixed-time control and proportional-integral-derivative (PI-DE) control, struggle to cope with the changing port environment, especially under conditions of high ship traffic or strong external disturbances, often failing to provide sufficient flexibility and robustness. In recent years, novel control methods such as model predictive control and sliding mode control have been gradually applied. These methods can accurately adjust control by sensing changes in the system's state in real time and predicting future traffic dynamics. However, in systems like port waterways that require consideration of multi-agent interactions, the high computational resource consumption of traditional model predictive control becomes even more severe.

[0004] Therefore, how to construct a comprehensive model that can reflect the dynamic changes in the state of complex port waters, ship operations, and environmental characteristics, and maintain efficient real-time control and scheduling of port ships in a highly uncertain dynamic environment, remains one of the key issues that urgently need to be addressed in the field of waterway traffic optimization and intelligent management technology. This invention is proposed to solve this problem. Summary of the Invention

[0005] The purpose of this invention is to address the problems of insufficient research on the interaction and coupling characteristics of models in dynamic environments of port waters and the lag in control methods. By introducing a comprehensive model based on cyber-physical architecture and an adaptive predictive control algorithm, an optimization method is proposed that can achieve dynamic feature capture, real-time feedback control and efficient ship traffic scheduling.

[0006] To achieve the above objectives, the present invention provides a method for optimizing and controlling port vessel traffic organization, characterized in that the method includes the following steps:

[0007] S1: Constructing a port vessel traffic organization optimization model under the cyber-physical system architecture;

[0008] Sa1: Construct a cyber-physical system framework, collect port data, input the port data into the cyber-physical system framework to realize dynamic management and rolling control of port vessels, and obtain the overall state vector and scheduling instructions. The scheduling instructions are used to obtain the port vessel traffic scheduling and control strategy. The cyber-physical system framework includes a physical layer and an information layer. The physical layer consists of sensing devices and berth management and scheduling control systems, and the information layer consists of a computing center and a control center.

[0009] Preferably, in step Sa1, the specific process is as follows:

[0010] After constructing a cyber-physical system framework, port data is input into the physical layer to obtain real-time environmental data. Then, the real-time environmental data is input into the computing center of the information layer. Through distributed computing and data storage, the overall state vector is obtained. Finally, the overall state vector is input into the control center to obtain scheduling instructions, which are used to derive port vessel traffic scheduling and control strategies.

[0011] The process of inputting real-time environmental data into the computing center in the information layer and processing it through distributed computing and data storage to obtain the overall state vector includes the following steps:

[0012] Sa11: The information layer uses time-series statistical indicators to calculate indicator variables for real-time environmental data, including the mean. Standard deviation and periodic fluctuation amplitude The formula for calculating the indicator variables of any real-time environmental data is uniformly expressed as:

[0013]

[0014] in the formula For the first Observations at each sampling point The number of sampling points. This represents the mean of real-time environmental data variables. The standard deviation of real-time environmental data variables. This represents the periodic fluctuation range of real-time environmental data variables. This represents the maximum value among the observed values ​​of real-time environmental data variables. This represents the minimum value among the observed values ​​of real-time environmental data variables;

[0015] use and ， Obtain the environmental disturbance feature state vector ， The environmental disturbance feature state vector is represented as follows:

[0016]

[0017] in, This represents the mean of the real-time wind speed variable. The standard deviation of the real-time wind speed variable is represented by... This represents the periodic fluctuation range of the real-time wind speed variable. This represents the mean of the real-time wind direction variable. The standard deviation of the real-time wind direction variable is represented. This represents the periodic fluctuation range of the real-time wind direction variable. This represents the mean of the real-time tide level variable. The standard deviation of the real-time tide level variable. This represents the periodic fluctuation amplitude of the real-time tide level variable. This represents the mean of the real-time wave height variable. The standard deviation of the real-time wave height variable. This represents the periodic fluctuation amplitude of the real-time wave height variable. This represents the mean of the visibility variable. The standard deviation of the visibility variable is represented by... This indicates the periodic fluctuation range of the real-time visibility variable;

[0018] According to the ship's position 、 Bow angle Longitudinal velocity lateral velocity and angular velocity Obtain the ship behavior feature state vector The ship behavior characteristic state vector is represented as follows:

[0019]

[0020] in, and These represent the ship's acceleration in the longitudinal and lateral directions, respectively. and This indicates the characteristics of longitudinal and lateral acceleration as a function of time; It indicates the rate of change of the ship's heading angle, that is, how quickly the ship turns; This indicates a further increase or decrease in the rate of change of heading due to changes in torque during a ship's turning process;

[0021] Traffic flow characteristics are quantified using the flow-density relationship method, as shown in the following formula:

[0022]

[0023] in the formula Indicates the number of ships. Indicates the length of the ship;

[0024] The traffic flow characteristics include ship traffic volume. Ship density Average speed queue length berth utilization rate and congestion duration ,

[0025] Utilizing ship traffic Ship density Average speed queue length berth utilization rate and congestion duration Obtain the traffic flow feature state vector The traffic flow characteristic state vector It is expressed as follows:

[0026]

[0027] Sa12: Characteristic state vector for environmental disturbances Ship behavior characteristic state vector Traffic flow characteristic state vector Organizing them structurally according to their functional attributes and integrating them into a single vectorized state description framework, we obtain the overall state vector of the port waterway multi-source information fusion algorithm. The formula is as follows:

[0028]

[0029] The real-time environmental data includes: wind speed. ,wind direction , tide level High waves ,visibility 、 Ship's position 、 Bow angle Longitudinal velocity lateral velocity and angular velocity and traffic flow characteristics; the traffic flow characteristics include ship traffic volume. Ship density Average speed queue length berth utilization rate and congestion duration ;

[0030] Sa2: Construct a port vessel traffic organization optimization model. Input the overall state vector into the port vessel traffic organization optimization model to obtain discrete decisions and the real-time state of the vessel. The port vessel traffic organization optimization model includes: FSM module and MLD module.

[0031] Preferably, in step Sa2, a port vessel traffic organization optimization model is constructed. The overall state vector is input into the port vessel traffic organization optimization model to obtain discrete decisions and the real-time state of the vessels. The specific process is as follows:

[0032] After constructing the port vessel traffic organization optimization model, the overall state vector is input into the FSM module to obtain the vessel's behavior, state changes, and discrete decisions; the vessel's behavior, state changes, and discrete decisions are input into the MLD module to obtain the vessel's real-time state.

[0033] S2: Based on an event-triggered model predictive control algorithm under adaptive horizon, a port vessel traffic scheduling control strategy is obtained. The model predictive control algorithm includes an event-triggered mechanism and an adaptive horizon mechanism.

[0034] Sb1: Based on an event-triggered mechanism and discrete decision-making and the real-time state of the ship, the judgment results of the input conditions of the model predictive control algorithm are obtained; an intelligent scheduling control system is constructed, the judgment results are used to update the control, and the system state is obtained through iterative updates by inputting the intelligent scheduling control system. The system step size is then obtained by predicting the system state using the model predictive control algorithm. The system status includes scheduling logic and real-time dynamic information; the intelligent scheduling and control system uses the discrete decision output by the FSM module to provide the system with high-level, rule-based scheduling logic, and uses the real-time changes in the ship's status output by the MLD module to provide the system with accurate, low-level real-time dynamic information.

[0035] Preferably, the judgment result of the input conditions of the model predictive control algorithm is obtained in step Sb1, and the specific process is as follows:

[0036] A preset parameter is set to trigger an event triggering mechanism when a data change exceeds a certain threshold. Based on the real-time status of the ship, discrete decisions, and preset parameters, the triggering condition is set as real-time error or difference exceeding the threshold, as shown below:

[0037] Formula 1:

[0038]

[0039] Formula 2:

[0040]

[0041] or

[0042] Formula 3:

[0043]

[0044] In the formula, This represents a fixed threshold for time error during berthing. A fixed threshold representing changes in external environmental conditions. Indicates the current moment. This indicates the error between the current berthing time and the expected berthing time. This represents the difference between the current quantity and the expected quantity of the external state. express Control input signal at any time, Indicates the next departure time. Indicates the first Control input within a time interval, This represents the minimum time at which the condition is satisfied. Indicates the current trigger time. Represents a continuous time variable;

[0045] Based on the judgment result obtained from the input conditions, an event-triggered mechanism is adopted during the ship's navigation process, which is divided into two types of control actions: ① If the trigger condition is not met, that is, according to Formula 1, the control input signal is less than the threshold in Formula 2. The last control input signal at which the maximum continuous non-trigger time is reached will be applied to the intelligent scheduling control system. ① The berth allocation order will be reset; ② If the triggering condition is met, that is, according to Formula 1, the control input signal is less than the threshold in Formula 3. Update control input signals for intelligent dispatching and control systems ;

[0046] Sb2: Based on an adaptive horizon mechanism, with system step size Based on this, obtain the real-time step size. ;

[0047] Preferably, in step Sb2, the specific process is as follows:

[0048] Adjusting the system step size based on the changing trend of the current state to achieve a balance between response speed and computational efficiency can be described by the following formula:

[0049]

[0050] In the formula, For real-time step size, and These are the minimum and maximum view lengths, respectively. Represents the actual deviation function;

[0051] Sb3: The judgment result of the input conditions of the model predictive control algorithm and the real-time step size. As the setting of predictive control for model predictive control algorithm, the multiplier alternation direction method is used to optimize the model predictive control algorithm, and then the port vessel traffic scheduling control strategy is finally obtained according to the scheduling instructions.

[0052] Preferably, in step Sb3, the model predictive control algorithm is optimized using the alternating direction method of multipliers. The specific process is as follows:

[0053] Sb31: Optimizing the Model Predictive Control Algorithm using the alternating direction method of multipliers, solving the convex optimization problem, and defining the set of index functions. and the introduction of slack variables The optimization objective can be transformed into the following problem, as shown below:

[0054]

[0055]

[0056] Indicates constraints. Represents a set Indicator functions, Indicates in variable Within the range of values, minimize the objective function;

[0057] Using Lagrange multipliers And add a secondary penalty item. Construct the augmented Lagrangian function, expressed as follows:

[0058]

[0059] in This represents the augmented Lagrange function. This represents the original objective function for optimization. Represents the Lagrange multiplier terms;

[0060] Sb32: Using the alternating direction method of multipliers for vectors and Perform alternating optimization, then optimize the Lagrange multipliers. Gradient ascent updates are performed, resulting in the following three-step iteration until a desired convergence tolerance is reached. The specific process is as follows:

[0061] First, fix the current and Minimize about Part To minimize the augmented Lagrange function ， variables The update can be written as a closed-form solution or a solution for a simple linear system, as follows:

[0062]

[0063] This represents the minimum value of the function. express The updated value;

[0064] Then, update To make it as close as possible to the result obtained after alternating optimization. Update value And still satisfy the constraints ， In this second iteration Update value Linear projection can be performed using simple matrix operations, as shown below:

[0065]

[0066] express The updated value;

[0067] Finally, the Lagrange multipliers are updated using the gradient ascent method to approximate the optimal constrained multiplier values, as shown below:

[0068]

[0069] express The updated value.

[0070] In this invention, traditional control methods often struggle to effectively address the challenges of high-density vessel traffic, complex channel structures, dynamic environmental changes, and multi-agent interactions in port waters. While Multi-Process Control (MPC) is widely used in port waterway traffic control due to its efficiency in constrained optimization problems, its reliance on a fixed-frequency decision update strategy can lead to decision lags in rapidly changing dynamic environments, thus affecting system response speed and safety. To address this issue, existing research has introduced event-triggered mechanisms and adaptive horizon methods into MPC, aiming to dynamically adjust the timing of decision updates, reduce computational burden, and improve response flexibility. However, the application of these methods in the specific environment of port waterways remains insufficient, requiring further research to ensure their effectiveness in practical applications. Attached Figure Description

[0071] Figure 1 This invention provides a flowchart of a method for optimizing and controlling port vessel traffic organization.

[0072] Figure 2 This is the port scheduling optimization control architecture based on CPS in this invention.

[0073] Figure 3 This refers to the delays in ship spatiotemporal trajectory and berth allocation in this invention.

[0074] Figure 4 For the comparison of the triggering time distribution of this invention (free navigation).

[0075] Figure 5 This is a comparison of the triggering time distribution of the present invention (collision avoidance navigation). Detailed Implementation

[0076] The purpose of this invention is to address the problems of insufficient research on the interaction and coupling characteristics of models in dynamic environments of port waters and the lag in control methods. By introducing a comprehensive model based on cyber-physical architecture and an adaptive predictive control algorithm, an optimization method is proposed that can achieve dynamic feature capture, real-time feedback control and efficient ship traffic scheduling.

[0077] This invention proposes a method for optimizing and controlling port vessel traffic organization, such as... Figure 1 As shown, it includes the following steps:

[0078] S1: Constructing a port vessel traffic organization optimization model under the cyber-physical system architecture;

[0079] Sa1: Construct a cyber-physical system framework, collect port data, input the port data into the cyber-physical system framework to obtain the overall state vector and scheduling instructions, realize dynamic management and rolling control of port vessels, and use the scheduling instructions to obtain port vessel traffic scheduling and control strategies; the cyber-physical system framework includes a physical layer and an information layer. The physical layer consists of sensing devices and berth management and scheduling control systems, and the information layer consists of a computing center and a control center.

[0080] Under the port CPS rolling optimization control mode, the port waterway cyber-physical system optimizes ship status prediction and control behaviors in each sampling period through real-time information interaction between the information layer and the physical layer, thereby achieving dynamic management and rolling control of the port. A specific CPS-based port optimization control system is as follows: Figure 2 As shown, the specific implementation steps of Sa1 are as follows:

[0081] After constructing a cyber-physical system framework, port data is input into the physical layer to obtain real-time environmental data. Then, the real-time environmental data is input into the computing center of the information layer. Through distributed computing and data storage, the overall state vector is obtained. Finally, the overall state vector is input into the control center to obtain scheduling instructions, which are used to derive port vessel traffic scheduling and control strategies.

[0082] The aforementioned process involves inputting real-time environmental data into the computing center within the information layer, and then processing it through distributed computing and data storage to obtain the overall state vector. The specific implementation steps are as follows:

[0083] Sa11: The information layer uses time-series statistical indicators to calculate indicator variables for real-time environmental data, including the mean. Standard deviation and periodic fluctuation amplitude The formula for calculating the indicator variables of any real-time environmental data is uniformly expressed as:

[0084]

[0085] in the formula For the first Observations at each sampling point The number of sampling points. This represents the mean of real-time environmental data variables. The standard deviation of real-time environmental data variables. This represents the periodic fluctuation range of real-time environmental data variables. This represents the maximum value among the observed values ​​of real-time environmental data variables. This represents the minimum value among the observed values ​​of real-time environmental data variables;

[0086] use and ， Obtain the environmental disturbance feature state vector ， The environmental disturbance feature state vector is represented as follows:

[0087]

[0088] in, This represents the mean of the real-time wind speed variable. The standard deviation of the real-time wind speed variable is represented by... This represents the periodic fluctuation range of the real-time wind speed variable. This represents the mean of the real-time wind direction variable. The standard deviation of the real-time wind direction variable is represented. This represents the periodic fluctuation range of the real-time wind direction variable. This represents the mean of the real-time tide level variable. The standard deviation of the real-time tide level variable. This represents the periodic fluctuation amplitude of the real-time tide level variable. This represents the mean of the real-time wave height variable. The standard deviation of the real-time wave height variable. This represents the periodic fluctuation amplitude of the real-time wave height variable. This represents the mean of the visibility variable. The standard deviation of the visibility variable is represented by... This indicates the periodic fluctuation range of the real-time visibility variable;

[0089] According to the ship's position 、 Bow angle Longitudinal velocity lateral velocity and angular velocity Obtain the ship behavior feature state vector The ship behavior characteristic state vector is represented as follows:

[0090]

[0091] in, and These represent the ship's acceleration in the longitudinal and lateral directions, respectively. and This indicates the characteristics of longitudinal and lateral acceleration as a function of time; It indicates the rate of change of the ship's heading angle, that is, how quickly the ship turns; This indicates a further increase or decrease in the rate of change of heading due to changes in torque during a ship's turning process;

[0092] Traffic flow characteristics are quantified using the flow-density relationship method, as shown in the following formula:

[0093]

[0094] in the formula Indicates the number of ships. Indicates the length of the ship;

[0095] The traffic flow characteristics include ship traffic volume. Ship density Average speed queue length berth utilization rate and congestion duration ,

[0096] Utilizing ship traffic Ship density Average speed queue length berth utilization rate and congestion duration Obtain the traffic flow feature state vector The traffic flow characteristic state vector It is expressed as follows:

[0097]

[0098] Sa12: Characteristic state vector for environmental disturbances Ship behavior characteristic state vector Traffic flow characteristic state vector Organizing them structurally according to their functional attributes and integrating them into a single vectorized state description framework, we obtain the overall state vector of the port waterway multi-source information fusion algorithm. The formula is as follows:

[0099]

[0100] The real-time environmental data includes: wind speed. ,wind direction , tide level High waves ,visibility 、 Ship's position 、 Bow angle Longitudinal velocity lateral velocity and angular velocity and traffic flow characteristics; the traffic flow characteristics include ship traffic volume. Ship density Average speed queue length berth utilization rate and congestion duration ;

[0101] Sa2: Construct a port vessel traffic organization optimization model. Input the overall state vector into the port vessel traffic organization optimization model to obtain discrete decisions and the real-time state of the vessel. The port vessel traffic organization optimization model includes: FSM module and MLD module.

[0102] The specific implementation steps of Sa2 are as follows:

[0103] After constructing the port vessel traffic organization optimization model, the overall state vector is input into the FSM module to obtain the vessel's behavior, state changes, and discrete decisions; the vessel's behavior, state changes, and discrete decisions are input into the MLD module to obtain the vessel's real-time state.

[0104] The operating mechanism of the FSM module and MLD module

[0105] The FSM module, by explicitly defining finite states and state transition conditions, can clearly and intuitively describe the evolution process of such discrete behaviors, giving the model higher expressiveness and maintainability at the logical level;

[0106] Sa21: The 5-tuple model of FSM is a fundamental way to describe the structure and behavior of FSM. It defines the operating mechanism of FSM through five key elements:

[0107]

[0108] In the formula, Represents a finite set of states; Represents a set of input events; Represents the state transition function. Indicates the initial state of the system; Indicates the target state. This indicates the operating mechanism of the FSM;

[0109] Sa22: Building upon this, the subsequent introduction of the MLD module aims to further integrate continuous dynamic behavior and logical constraints within the discrete-state framework. While the FSM module can describe discrete events and state transitions, it struggles to directly handle the continuous motion of ships and the dynamic effects of environmental disturbances. The MLD module, by combining logical conditions with dynamic equations, can simultaneously consider continuous state variables such as ship speed and acceleration, along with their constraints. This allows for the construction of a comprehensive model that conforms to logical rules and reflects actual physical dynamics. This integration enables the model to more accurately describe the hybrid dynamic characteristics of port waterway traffic systems. The MLD module contains three equations, describing the system's dynamic evolution, output calculation, and logical constraints, respectively. Its basic form can be described as the following discrete-time system:

[0110]

[0111] In the formula, Indicates time, Indicates the next moment. Represents the state matrix, Represents the control input matrix. Represents a matrix of logical variables. Represents the auxiliary variable matrix. Represents continuous state variables, the current dynamic state of the system. Represents the state output matrix. Represents the control input / output matrix. Represents the output matrix of logical variables. This represents the auxiliary variable output matrix. Represents a constraint matrix vector. This represents the control input variable, which affects the state and output of the system. Output variables represent the overall behavior of the system. Represents discrete logic variables, used to describe discrete events or logical conditions; This represents a mixed variable used to connect continuous and discrete behaviors;

[0112] Sa23: The state update equation describes continuous state variables. How to input Discrete logic variables and mixed variables Under the influence of [the process], it evolves into the state of the next moment; the output equation defines the output variable. How is determined by the state, input, logical variables, and mixed variables? The logic and constraints, combined with the discrete and continuous characteristics of the system, define the constraint relationships between logical variables and mixed variables, thus fully describing the mixed dynamic characteristics of the system.

[0113] The MLD model for ship navigation integrates state transitions, system dynamic behavior, input constraints, and state switching from the FSM model of the channel segment, by considering... Given the ship's speed, position, external environmental factors, and navigation target at any given time, solve and optimize. The control sequence at each moment is used to make control decisions during navigation.

[0114] The dynamic and kinematic characteristics of the ship under different operating conditions are captured in a segmented manner. In the dynamic model, the ship's propulsion power... propulsion by propeller and total resistance of navigation The difference determines:

[0115]

[0116] Among them, propeller thrust Represented as:

[0117]

[0118] in the formula Indicates the thrust coefficient. Indicates the propeller speed. ; Indicates the propeller diameter;

[0119] Total resistance Including hydrostatic resistance and wind resistance Hydrostatic resistance is composed of frictional resistance 、 Residual resistance and attachment resistance The composition and wind resistance of the ship are derived from empirical formulas and calculated as follows:

[0120]

[0121] in the formula Indicates the coefficient of frictional resistance. Indicates the drag coefficient of the appendage. Indicates the air drag coefficient; and These represent the densities of seawater and air, respectively, in kg / m³. 3 ; The wetted area of ​​the hull surface is expressed in meters (m). 2 Unit; The projected area of ​​the mid-section is expressed in meters. 2 Unit; and These represent ship speed and relative wind speed, respectively, in m / s; and These represent the amplitude of the fluctuation and the length between the ship's perpendiculars, respectively, in meters (m).

[0122] Sa24: based on sampling period The velocity state of a ship over a time interval is described by the following state equation:

[0123]

[0124] Indicates the first The speed status of the ship in each sampling period Indicates the first The speed status of the ship in each sampling period express The rate of change of ship speed at any given moment;

[0125] Changes in ship position and course can be predicted using the following formula:

[0126]

[0127] in, Indicates the first State variables at discrete time points, Indicates input variables, Represents state variables, Indicates the first Discrete time;

[0128] In the formula, It is the position vector in the inertial coordinate system, including position coordinates. and heading angle ; It is the velocity vector in the ship's fixed coordinate system, including the sway velocity. sway speed and bow roll rate ; It is the rotation matrix from the hull coordinate system to the inertial coordinate system, expressed as:

[0129]

[0130] Sa25: The MLD module uses logical variables to describe the state transitions of a ship in different navigation segments. The Boolean variables are defined as follows:

[0131] This indicates a free navigation segment, in which there is no risk of collision. ,otherwise , This indicates the collision avoidance phase. If a collision avoidance maneuver is performed, then... ,otherwise , This indicates a channel segment change. If the vessel enters a deceleration state and undergoes a sharp turn, then... ;otherwise ;

[0132] When the relative distance between a ship and other vessels approaches the danger value D, the logical variable switches from free navigation state to collision avoidance state. When the ship enters a deceleration state and its steering angle changes, the logical variable triggers a channel switch.

[0133] According to the conversion rules between the two in MLD:

[0134] True is equivalent to ,in, , , Indicates the allowable error amount;

[0135] Logical variables , With ship speed The relationship can be expressed by the following inequality:

[0136]

[0137] This indicates the maximum value of the control input. This indicates the minimum value of the control input;

[0138] When a ship enters a deceleration state accompanied by a sharp change in steering angle, logical variables The rate of change of the ship's speed will change from 0 to 1. It will be subject to the following constraints:

[0139]

[0140] Indicates the maximum acceleration. Indicates the minimum acceleration. Indicates the allowable error amount;

[0141] Sa26: Through the transformation between logical variables and auxiliary variables, the MLD model can be represented in the following form:

[0142]

[0143] in, , , Indicates the system at time... The state vector, Indicates the system at time... The state vector, Represents a binary logical variable. Indicates time Control input;

[0144] Finally, the correctness of the discrete state switching of the constructed model, the degree of satisfaction of logical constraints, the accuracy of continuous dynamic behavior, and the coordination of discrete and continuous interaction processes are verified through historical data and simulation environment tests, thereby verifying the model's descriptive ability and practicality in complex hybrid dynamic systems.

[0145] S2: Based on an event-triggered model predictive control algorithm under adaptive horizon, a port vessel traffic scheduling control strategy is obtained. The model predictive control algorithm includes an event-triggered mechanism and an adaptive horizon mechanism.

[0146] Sb1: Based on an event-triggered mechanism, and using discrete decision-making and the real-time state of the ship as a foundation, the judgment results of the input conditions of the model predictive control algorithm are obtained; an intelligent scheduling control system is constructed, the judgment results are used to update the control, and the system state is obtained through iterative updates by inputting the intelligent scheduling control system; the system state is predicted by the model predictive control algorithm to obtain the system step size Np; the system state includes scheduling logic and real-time dynamic information; the intelligent scheduling control system uses the discrete decision-making output of the FSM module to provide the system with high-level, rule-based scheduling logic, and uses the real-time changes in the ship's state output of the MLD module to provide the system with accurate, low-level real-time dynamic information;

[0147] The specific implementation steps of Sb1 are as follows:

[0148] A preset parameter is set to trigger an event triggering mechanism when a data change exceeds a certain threshold. Based on the real-time status of the ship, discrete decisions, and preset parameters, the triggering condition is set as real-time error or difference exceeding the threshold, as shown below:

[0149] Formula 1:

[0150]

[0151] Formula 2:

[0152]

[0153] or

[0154] Formula 3:

[0155]

[0156] In the formula, This represents a fixed threshold for time error during berthing. A fixed threshold representing changes in external environmental conditions. Indicates the current moment. This indicates the error between the current berthing time and the expected berthing time. This represents the difference between the current quantity and the expected quantity of the external state. express Control input signal at any time, Indicates the next departure time. Indicates the first Control input within a time interval, This represents the minimum time at which the condition is satisfied. Indicates the current trigger time. Represents a continuous time variable;

[0157] Based on the judgment result obtained from the input conditions, an event-triggered mechanism is adopted during the ship's navigation process, which is divided into two types of control actions: ① If the trigger condition is not met, that is, according to Formula 1, the control input signal is less than the threshold in Formula 2. The last control input signal at which the maximum continuous non-trigger time is reached will be applied to the intelligent scheduling control system. ① The berth allocation order will be reset; ② If the triggering condition is met, that is, according to Formula 1, the control input signal is less than the threshold in Formula 3. Update control input signals for intelligent dispatching and control systems .

[0158] Sb2: Based on an adaptive horizon mechanism, with system step size Based on this, obtain the real-time step size. ;

[0159] The specific implementation steps of Sb2 are as follows:

[0160] Adjusting the prediction step size based on the changing trend of the current state achieves a balance between response speed and computational efficiency, which can be described by the following formula:

[0161]

[0162] In the formula, For the predicted length that changes in real time, and These are the minimum and maximum view lengths, respectively. Represents the actual deviation function;

[0163] Sb3: The judgment result of the input conditions of the model predictive control algorithm and the real-time step size. As the setting of predictive control for model predictive control algorithm, the multiplier alternation direction method is used to optimize the model predictive control algorithm, and then the port vessel traffic scheduling control strategy is finally obtained according to the scheduling instructions.

[0164] The optimization of the model predictive control algorithm using the alternating direction method of multipliers in Sb3 is implemented through the following steps:

[0165] Sb31: Optimizing the Model Predictive Control Algorithm using the alternating direction method of multipliers, solving the convex optimization problem, and defining the set of index functions. and the introduction of slack variables The optimization objective can be transformed into the following problem, as shown below:

[0166]

[0167] Indicates constraints. Represents a set Indicator functions, Indicates in variable Within the range of values, minimize the objective function;

[0168] Using Lagrange multipliers And add a secondary penalty item. Construct the augmented Lagrangian function, expressed as follows:

[0169]

[0170] in This represents the augmented Lagrange function. This represents the original objective function for optimization. Represents the Lagrange multiplier terms;

[0171] Sb32: The alternating direction method of multipliers optimizes vectors through alternation. and Instead of minimizing both simultaneously, then applying Lagrange multipliers... Gradient ascent updates are performed, resulting in the following three-step iteration until a desired convergence tolerance is reached. The specific process is as follows:

[0172] First, fix the current and Minimize about Part To minimize the augmented Lagrange function ， In many optimization problems The update can be written as a closed-form solution or a solution for a simple linear system, compared to considering both simultaneously. and The computational complexity is low, as shown below:

[0173]

[0174] This represents the minimum value of the function. express The updated value;

[0175] Then, update To make it as close as possible to the result obtained after alternating optimization. Update value However, the constraints are still satisfied. ， In this second iteration Only linear projection is needed, which can be accomplished with simple matrix operations. It is a sparsity constraint, and the projection can be directly set to zero to improve computational efficiency, as shown below:

[0176]

[0177] Indicates update The value;

[0178] Finally, the Lagrange multipliers are updated using the gradient ascent method to approximate the optimal constrained multiplier values, as expressed by the formula: and If there is inconsistency, increase the dual variable. The value of is used to impose stronger constraints in the next iteration, as shown below:

[0179]

[0180] Indicates update The value;

[0181] Sb33: The alternating direction method of multipliers ensures that the dual variable gradually converges during the iteration process, making... x and z They are gradually converging.

[0182] In response to the port vessel traffic organization optimization and control method proposed above, this paper takes the entry and exit of vessels in Wuhan port as the research object, establishes a vessel traffic optimization organization model, generates control schemes, and verifies the accuracy of the method of this invention.

[0183] The area measures 2000m × 2000m. There are three available channels within the waterway, one for small vessels and two for large vessels. The port has four berths: berth 1 is a large berth, and berths 2, 3, and 4 are small berths. Tidal windows are defined as 0-5 hours, 10-15 hours, 20-25 hours, 30-35 hours, and 40-45 hours. Large vessels must navigate within a specific tidal window. The port has opened one large berth and three small berths to meet vessel berthing needs. The state description, state transition function, and output function tables for the port vessel entry and exit FSM model are shown in Tables 1 and 2. The simulation scheduling includes the entry and exit operations of 10 vessels; vessel information, including arrival time, load capacity, and berth requirements, is shown in Table 3.

[0184] Simulation analysis was performed using Matlab modeling, which monitored the real-time operational status of ships in the waiting area, channel area, and berth area. Berth allocation and navigation scheduling were accomplished using model predictive control algorithms. The system dynamically optimized the allocation scheme based on ship characteristics and berth status. Considering the minimum prediction time domain and computational cost, an initial prediction time domain NP(0) of 40-50 is recommended. In the two simulation scenarios of this experiment, the initial prediction time domain was set to 40, and the controller parameters are shown in Table 4.

[0185] Table 1 Current information on vessels entering and leaving the port

[0186]

[0187] Table 2 State transition function and output function table

[0188]

[0189] Table 3 Current Port Entry and Exit Information

[0190]

[0191] Table 4 Controller Parameters

[0192]

[0193] (2) Experimental results

[0194] The simulation analysis examined the berth allocation strategy and vessel scheduling optimization effects under different vessel traffic conditions. Due to the large number of vessels and limited berth resources, vessels often need to wait in the port waiting area for a period of time before entering a berth, leading to increased waiting time. However, by optimizing the berth allocation strategy and sailing time arrangement, the model can effectively reduce vessel waiting time and improve berth utilization. Specific berth allocation results and sailing time statistics are shown in Table 5, and the spatiotemporal trajectory and sailing time distribution of vessels are shown in [Table 5]. Figure 3 .

[0195] Table 5. Statistics on ship berth allocation and sailing time

[0196]

[0197] By incorporating berth allocation strategies, the model optimizes vessel berth allocation and navigation time scheduling. Simulations show that vessels can efficiently complete port entry, berthing, and departure operations with multi-vehicle cooperation. Within the waterway, when multiple vessels are navigating simultaneously and the safe distance is insufficient, the system automatically adjusts speed to avoid collisions; if the safe distance condition is met, vessels navigate along a free path.

[0198] To verify the performance of the proposed control method, a comparative analysis was conducted with traditional model predictive control methods in two scenarios: free navigation and obstacle avoidance navigation. The computational performance evaluation indicators for different schemes are shown in Table 6. (Trigger time distribution diagram) Figure 4 and Figure 5 The results show that traditional model predictive control maintains a high triggering frequency in both navigation scenarios, and its iteration time fluctuates significantly in the early and middle stages. In contrast, event-triggered adaptive horizon model predictive control triggers more frequently in the early stages of navigation, while the triggering frequency decreases significantly in the middle and later stages, demonstrating good adaptive adjustment capabilities. By introducing an optimized triggering strategy, this method significantly reduces the total number of triggers and the overall iteration time. Specifically, in the free navigation scenario, the number of triggers is reduced by approximately 45.0%, and the total iteration time is shortened by approximately 62.4%; in the obstacle avoidance navigation scenario, the number of triggers is reduced by approximately 28.8%, and the total iteration time is reduced by approximately 34.2%.

[0199] Simulation results show that the method provided by this invention significantly improves scheduling efficiency in complex environments while reducing computational costs, providing an effective solution for intelligent scheduling optimization.

[0200] Table 6 Evaluation indicators of computational performance under different schemes

[0201] .

Claims

1. A port vessel traffic organization optimization and regulation method, characterized in that, The application relates to a port ship traffic organization optimization method based on a cyber-physical system framework. S1: constructing a port ship traffic organization optimization model under a cyber-physical system framework; Sa1: constructing a cyber-physical system framework, collecting port data, inputting the port data into the cyber-physical system framework to obtain an overall state vector and a scheduling instruction, realizing dynamic management and rolling control of port ships, and using the scheduling instruction to obtain a port ship traffic scheduling control strategy; the cyber-physical system framework comprises a physical layer and an information layer, the physical layer is composed of sensing equipment, berth management and scheduling control systems, the information layer is composed of a computing center and a control center, and the overall state vector comprises an environmental disturbance characteristic state vector, a ship behavior characteristic state vector and a traffic flow characteristic state vector; Sa2: constructing a port ship traffic organization optimization model, inputting the overall state vector into the port ship traffic organization optimization model to obtain discrete decisions and real-time states of ships, and the port ship traffic organization optimization model comprising an FSM module and an MLD module; After the port ship traffic organization optimization model is constructed, the overall state vector is input into the FSM module to obtain behaviors, state changes and discrete decisions of the ships; and the behaviors, state changes and discrete decisions of the ships are input into the MLD module to obtain real-time states of the ships; S2: obtaining a port ship traffic scheduling control strategy based on an event-triggered model predictive control algorithm under an adaptive horizon, and the model predictive control algorithm comprising an event-triggering mechanism and an adaptive horizon mechanism; Sb1: based on the event triggering mechanism, the judgment result of the input condition of the model prediction control algorithm is obtained based on the discrete decision and the real-time state of the ship; an intelligent scheduling control system is constructed, the judgment result is updated and controlled, input into the intelligent scheduling control system for iterative updating to obtain the system state, and the system step is obtained by predicting the system state by using the model prediction control algorithm ; the system state includes scheduling logic and real-time dynamic information; the intelligent scheduling control system uses the discrete decision output by the FSM module to provide the system with high-level, rule-based scheduling logic, and uses the real-time state change of the ship output by the MLD module to provide the system with accurate, bottom-level real-time dynamic information; Sb2: Based on an adaptive field of view mechanism, taking system steps as a basis, acquiring real-time steps ; According to the change trend of the current state, the system step is adjusted to balance the response speed and the calculation efficiency, and the following formula is used to describe the same: ； where, is the real-time step size, and are the minimum and maximum horizon lengths, respectively, denotes the actual deviation function; Sb3: judge the result of model predictive control algorithm input condition and real-time step As the setting of the predictive control of the model predictive control algorithm, the model predictive control algorithm is optimized by the multiplier alternating direction method, the convex problem is solved, and finally the port ship traffic scheduling control strategy is obtained according to the scheduling instruction.

2. The port vessel traffic optimization and regulation method of claim 1, wherein, In the Sa1, the overall state vector and the scheduling instruction are obtained, and the specific process is as follows: After the cyber-physical system framework is constructed, the port data is input into the physical layer to obtain real-time environmental data, the real-time environmental data is input into the computing center of the information layer, the overall state vector is obtained through distributed calculation and data storage, and the overall state vector is input into the control center to obtain the scheduling instruction, which is used to obtain the port ship traffic scheduling control strategy.

3. The port vessel traffic organization optimization and regulation method according to claim 2, characterized in that, In the Sa1, the overall state vector and the scheduling instruction are obtained, and the specific process is as follows: Sa11: The information layer adopts time series statistical indicators for index variable calculation of real-time environmental data, the indicators including mean value , standard deviation and periodic fluctuation amplitude , the index variable calculation formula of any real-time environmental data is uniformly expressed as: ； in the formula For the first Observations at each sampling point The number of sampling points. This represents the mean of real-time environmental data variables. The standard deviation of real-time environmental data variables. This indicates the periodic fluctuation range of real-time environmental data variables. This represents the maximum value among the observed values ​​of real-time environmental data variables. This represents the minimum value among the observed values ​​of real-time environmental data variables; Utilizing , and ， obtaining an environmental disturbance feature state vector ， The environmental disturbance feature state vector is represented as follows: ； wherein, represents a mean of the real-time wind speed variable, represents a standard deviation of the real-time wind speed variable, represents an amplitude of periodic fluctuations of the real-time wind speed variable, represents a mean of the real-time wind direction variable, represents a standard deviation of the real-time wind direction variable, represents an amplitude of periodic fluctuations of the real-time wind direction variable, represents a mean of the real-time tide level variable, represents a standard deviation of the real-time tide level variable, represents an amplitude of periodic fluctuations of the real-time tide level variable, represents a mean of the real-time wave height variable, represents a standard deviation of the real-time wave height variable, represents an amplitude of periodic fluctuations of the real-time wave height variable, represents a mean of the visibility variable, represents a standard deviation of the visibility variable, represents an amplitude of periodic fluctuations of the real-time visibility variable; According to the position of the ship 、 The heading angle The longitudinal velocity The lateral velocity And the angular velocity A ship behavior feature state vector Is obtained, which represents as follows: ; wherein, and respectively represent the acceleration of the ship in the longitudinal and transverse directions; and represent the characteristics of the longitudinal and transverse acceleration over time; represents the rate of change of the heading angle of the ship, i.e. how quickly the ship is turning; represents a further increase or decrease in the rate of change of the heading of the ship due to changes in the moment during the turning of the ship; The traffic flow characteristic is quantified by using a flow-density relationship method, and the formula is as follows: ; In the formula denotes the number of ships, denotes the length of the ship; The traffic flow characteristics include vessel flow rate , vessel density , average speed , queue length , berth utilization and congestion duration , Utilizing vessel flow , vessel density , average speed , queue length , berth utilization and congestion duration Obtaining a traffic flow feature state vector , said traffic flow feature state vector is represented as follows: ; Sa12: environmental disturbance feature state vector , ship behavior feature state vector , and traffic flow feature state vector According to the functional attributes, it is structurally organized and integrated into a single vectorized state description framework to obtain the overall state vector of the port water area multi-source information fusion algorithm , the formula is as follows: ； The real-time environmental data includes: wind speed , wind direction , tide level , wave height , visibility 、 The position of the ship 、 The bow angle , longitudinal speed , lateral speed , and angular velocity , and traffic flow characteristics; the traffic flow characteristics include ship flow , ship density , average speed , queue length , berth utilization rate and congestion duration .

4. The port vessel traffic optimization and regulation method of claim 1, wherein, In the Sb1, the judgment result of the input condition of the model predictive control algorithm is obtained, and the specific process is as follows: A preset parameter of triggering the event-triggering mechanism is set when the data change exceeds a certain threshold value, based on the event-triggering mechanism, the real-time state of the ship, the discrete decision and the preset parameter are used as the basis, the triggering condition is that the real-time error or difference exceeds the threshold value, and the triggering condition is represented as follows: Formula 1: ； Formula 2: ; Or Formula 3: ； in the formula, denotes a fixed threshold value for the time error during berthing, denotes a fixed threshold value for the change in external environmental conditions, denotes the current time, denotes the error between the current berthing time and the expected berthing time, denotes the difference between the current quantity and the predicted quantity of the external state, denotes the control input signal at the time, denotes the next departure time, denotes the control input in the time interval, denotes the minimum time that meets the condition, denotes the current triggering time, denotes the continuous time variable; According to the judgment result obtained from the input condition, an event triggering mechanism is adopted during the ship navigation process, which is divided into two kinds of control actions: ① If the triggering condition is not met, that is, the control input signal is less than the threshold value in formula 2 according to formula 1 , the last control input signal when the maximum continuous non-triggering time is reached is applied to the intelligent scheduling control system , the berth allocation order is reset; ② If the triggering condition is met, that is, the control input signal is less than the threshold value in formula 3 according to formula 1 , the control input signal is updated and applied to the intelligent scheduling control system .

5. The port vessel traffic organization optimization and regulation method according to claim 1, characterized in that, In the Sb3, the model predictive control algorithm is solved by using the multiplier alternating direction method to optimize the convex problem, and the specific process is as follows: Sb31: Utilize the Multiplicative Algebraic Direction Method to optimize the model predictive control algorithm, solve the convex optimization problem, and define the index function set and introduce a relaxation variable Convert the optimization objective into the following problem, which is represented as follows: ; denotes a constraint, denotes a set of indicator functions, denotes a minimization of an objective function over a range of values of a variable ; Using Lagrange multipliers and adding a quadratic penalty term The augmented Lagrangian function is constructed and is represented as follows: ； wherein represents an augmented Lagrangian function, represents the original optimization objective function, represents a Lagrange multiplier term; Sb32: Alternating Direction Method of Multipliers (ADMM) is used to optimize the vector and Alternating optimization is performed, and then the Lagrange multipliers are updated by gradient ascent, forming a three-step iteration as follows, until a desired convergence tolerance is reached, as follows: First, fix the current and Minimize about Part To minimize the augmented Lagrange function ， variables The update can be written as a closed-form solution or a solution for a simple linear system, as follows: ; denotes minimizing a function, denotes an updated value of Then, update To make it as close as possible to the result obtained after alternating optimization. Update value And still satisfy the constraints ， In this second iteration Update value Linear projection can be performed using simple matrix operations, as shown below: ; represents an updated value of Finally, the Lagrange multiplier is updated by using the gradient ascent method, so that the constraint multiplier value approaches the optimal value, and the updating is represented as follows: ; represents an updated value of the variable.

Images