Ship offshore operation decision-making method, electronic equipment, medium and product

Abstract

The invention provides a ship offshore operation decision-making method, electronic equipment, a medium and a product. The ship offshore operation decision-making method comprises the steps of obtaining operation data of a ship; the operation data are input into a decision model, ship operation prediction information output by the decision model is obtained, a ship operation strategy is generated according to the ship operation prediction information, the decision model comprises a convolutional neural network, a recurrent neural network and an attention module which are connected in sequence, and hyper-parameters of the convolutional neural network are determined based on a preset optimization algorithm; the ship operation strategy comprises at least one of an operation time window, cable tension and posture adjustment. According to the method, the ship operation information in the future time period can be accurately predicted, so that targeted adaptive decision suggestions are quickly provided, the ship can flexibly deal with various complex and sudden offshore operation conditions, the decision efficiency is high, the decision is not influenced by personal experience and feeling, the occurrence of wrong decisions is effectively reduced, and the operation safety is improved.

Description

Technical Field

[0001] The present application relates to the field of ship operations. Specifically, the present application relates to a ship offshore operation decision-making method, an electronic device, a medium, and a product. Background Art

[0002] With the rapid development of offshore engineering, the scope of ship operations offshore has gradually increased and the operation time has also been greatly extended. However, the environment faced by ships during offshore operations is complex, and they need to face factors such as sea condition changes, wind and wave impacts, and the uncertainty of the state of operating equipment during operation. These factors have put forward higher requirements for the adaptability and decision-making ability of ships. Traditional decision-making methods rely on preset rules and personal experience to predict the next ship operation data, and then make ship operation decisions based on the prediction results. This method often has difficulty coping with the dynamically changing environment, may lead to low decision-making efficiency, and the decision-making results are prone to incorrect decisions due to the influence of personal experience and perception, thus posing potential safety hazards. Summary of the Invention

[0003] In view of the shortcomings of the existing methods, the present application provides a ship offshore operation decision-making method, an electronic device, a medium, and a product, which can solve the problems that the existing ship offshore operation decision-making depends on preset rules and personal experience judgment, is difficult to cope with the dynamically changing environment, has low decision-making efficiency and is prone to incorrect decisions, resulting in potential safety hazards in work.

[0004] According to one aspect of the embodiments of the present application, the embodiments of the present application provide a ship offshore operation decision-making method, the method including:

[0005] Obtain the operation data of the ship;

[0006] Input the operation data into a decision model, obtain the ship operation prediction information output by the decision model, and generate a ship operation strategy according to the ship operation prediction information. The decision model includes a convolutional neural network, a recurrent neural network, and an attention module connected in sequence. The hyperparameters of the convolutional neural network are determined based on a preset optimization algorithm. The ship operation strategy includes at least one of an operation time window, a cable tension, and an attitude adjustment.

[0007] In a possible implementation manner, the ship is a crane ship. The obtaining the operation data of the ship includes:

[0008] Determine that the ship is a crane ship, and collect the operation data of the crane ship in the working state. The operation data includes wave data, ship motion response data, and cable force data;

[0009] Preprocess the operation data and add position information to the operation data.

[0010] In a possible implementation manner, the preset optimization algorithm is the PLO algorithm. The determination of the hyperparameters includes:

[0011] Randomly generate an initial particle swarm in the solution space, call the objective function to evaluate the fitness of each particle in the initial particle swarm, and obtain the initial fitness value;

[0012] Iteratively update the velocity and position of the particles to generate a new particle swarm, obtain the fitness update value corresponding to the particles after each update, determine the optimal particle position according to the initial fitness value and the fitness update value, and determine the hyperparameters of the convolutional neural network based on the optimal particle position. The hyperparameters include the learning rate, the size of the convolutional kernel, the number of neurons, and the number of convolutional kernels.

[0013] In a possible implementation manner, the iterative update of the velocity and position of the particles includes:

[0014] Update the velocity and position of the particles according to the rotation motion and the aurora ellipse walking strategy;

[0015] Execute the particle collision strategy based on a preset probability to update the particle position.

[0016] In a possible implementation manner, the execution of the particle collision strategy based on a preset probability to update the particle position includes:

[0017] Update the particle position through the particle collision formula. The particle collision formula is:

[0018] X new (i, j) = X(i, j) + sin(r3 × π) × (X(i, j) - X(a, j))

[0019] In the formula, X new (i, j) is the parameter in the i-th row and j-th column of the updated particle swarm, X(i, j) is the parameter in the i-th row and j-th column of the particle swarm before update, r3 is a random number, and X(a, j) represents any parameter in the particle swarm.

[0020] In a possible implementation manner, the obtaining of the ship operation prediction information output by the decision model includes:

[0021] Input the operation data into the convolutional neural network to obtain the operation vector sequence output by the convolutional neural network;

[0022] Obtain the operation prediction data output by the recurrent neural network based on the operation vector sequence. The operation prediction data includes the hidden state sequence at different time steps;

[0023] Based on the weighted processing of the job prediction data by the attention module, ship operation prediction information is obtained.

[0024] In a possible implementation, the weighted processing includes:

[0025] Attention coefficients at different time steps are obtained according to the hidden state sequence, and the ship operation prediction data is weighted processed based on the attention coefficients to obtain ship operation prediction information;

[0026] The calculation formula of the attention coefficient is:

[0027]

[0028] In the formula, Wα represents the attention weight matrix; represents the hidden state sequence, is the forward hidden state vector at time t is the backward hidden state vector, softmax represents the normalized exponential function, and α t is the attention coefficient.

[0029] According to one aspect of the embodiments of the present application, an electronic device is provided, including a memory, a processor, and a computer program stored on the memory, and the processor executes the computer program to implement the steps of the above method.

[0030] According to one aspect of the embodiments of the present application, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0031] According to one aspect of the embodiments of the present application, a computer program product is provided, which includes a computer program, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0032] The beneficial technical effects brought by the technical solutions provided by the embodiments of the present application include:

[0033] A ship offshore operation decision-making method provided by the present application has the beneficial effects that operation data of the ship is obtained; the operation data is input into a decision-making model to obtain ship operation prediction information output by the decision-making model, and a ship operation strategy is generated according to the ship operation prediction information. The decision-making model includes a convolutional neural network, a recurrent neural network, and an attention module connected in sequence. The hyperparameters of the convolutional neural network are determined based on a preset optimization algorithm. The ship operation strategy includes at least one of an operation time window, a cable tension, and an attitude adjustment. The present application collects the operation data of the ship, uses the decision-making model to output the ship operation prediction information corresponding to the operation data, and obtains the ship operation strategy according to the ship operation prediction information, which can accurately predict the ship operation information in the future period, so as to quickly provide targeted and adaptive decision-making suggestions, enabling the ship to flexibly respond under various complex and unexpected offshore operation conditions, with high decision-making efficiency and not affected by personal experience and feelings, effectively reducing the occurrence of wrong decisions, and thus improving the operation safety.

[0034] Additional aspects and advantages of the present application will be given in part in the following description, which will become apparent from the following description, or can be understood through the practice of the present application. Brief Description of the Drawings

[0035] The above and / or additional aspects and advantages of the present application will become apparent and easy to understand from the following description of the embodiments in conjunction with the drawings, where:

[0036] Figure 1 is a flowchart of the ship offshore operation decision-making method provided by the embodiment of the present application;

[0037] Figure 2 is a structural diagram of the electronic device provided by the embodiment of the present application. Detailed Embodiments

[0038] Embodiments of the present application will be described below with reference to the drawings in the present application. It should be understood that the embodiments described below in conjunction with the drawings are exemplary descriptions for explaining the technical solutions of the embodiments of the present application, and do not constitute limitations on the technical solutions of the embodiments of the present application.

[0039] Those skilled in the art can understand that, unless specifically stated otherwise, the terms "the" and "said" used herein may also include the plural form. It should be further understood that the term "including" used in the description of the present application means the presence of the described features, integers, steps, operations, elements, and / or components, but does not exclude the presence of other features, information, data, steps, operations, elements, components, and / or combinations thereof supported by the art. It should be understood that when we say an element is "connected" or "coupled" to another element, the element can be directly connected or coupled to the other element, or it can mean that the element and the other element establish a connection relationship through an intermediate element. In addition, the "connection" or "coupling" used herein may include wireless connection or wireless coupling. The term "and / or" used herein means at least one of the items defined by the term. For example, "A and / or B" can be implemented as "A", or as "B", or as "A and B".

[0040] To make the objectives, technical solutions, and advantages of the present application clearer, the following will further describe the embodiments of the present application in detail with reference to the accompanying drawings.

[0041] The embodiment of the present application provides a method for making decisions on ship offshore operations. This method can be used in mobile phones, computers, servers, the cloud, and other terminals that can obtain the operation data of the ship and generate ship operation strategies based on this operation data.

[0042] Optionally, the ship can be a dredger, a crane ship, a pile driving ship, a cable laying ship, an offshore rescue and salvage ship, a floating dock, a diving work ship, etc., which can be used for offshore operations.

[0043] As Figure 1 shown, the method for making decisions on ship offshore operations of the present application includes:

[0044] S101: Obtain the operation data of the ship.

[0045] Optionally, the type of data to be obtained, the acquisition frequency of the data, the data acquisition method, etc. can be determined according to the type of the ship, the work requirements, and the functions of the ship.

[0046] In one embodiment, the ship is a crane ship. When the crane ship is operating, the wave data, the ship motion response data, and the cable force data of the crane ship in the working state are obtained.

[0047] Optionally, the ship is a crane ship. Obtaining the operation data of the ship includes: determining that the ship is a crane ship, collecting the operation data of the crane ship in the working state, the operation data including wave data, ship motion response data, and cable force data; preprocessing the operation data and adding position information to the operation data.

[0048] Optionally, the wave data includes wave environment parameters such as wave height, wave direction, and wave period, and the ship motion response data includes attitude data of the ship such as pitch, roll, and yaw. The ship motion response data can be obtained through an attitude instrument, and the cable force data of the ship can be obtained by testing the cables of the ship with a cable force tester.

[0049] Optionally, the preprocessing of the operation data includes at least one of denoising, outlier removal, time synchronization, and normalization.

[0050] Optionally, the collected operation data is time-series data, and the position information can be added to the operation data in a position encoding manner so that each time step in the operation data corresponds to position information.

[0051] S102: Input the operation data into the decision model, obtain the ship operation prediction information output by the decision model, and generate a ship operation strategy based on the ship operation prediction information.

[0052] Optionally, the decision model includes a convolutional neural network, a recurrent neural network, and an attention module connected in sequence. The hyperparameters of the convolutional neural network are determined based on a preset optimization algorithm. The ship operation strategy includes at least one of an operation time window, cable tension, and attitude adjustment.

[0053] Optionally, the preset optimization algorithm includes at least one of simulated annealing algorithm (SSA), genetic algorithm (GA), particle swarm optimization algorithm (PSO), differential evolution algorithm (DE), gray wolf optimization algorithm (GWO), and PLO (Polar Lights Optimization) algorithm.

[0054] Optionally, the preset optimization algorithm is the PLO algorithm. The determination of the hyperparameters includes: randomly generating an initial particle swarm in the solution space, calling the objective function to evaluate the fitness of each particle in the initial particle swarm to obtain the initial fitness value; iteratively updating the velocity and position of the particles to generate a new particle swarm, obtaining the fitness update value corresponding to each updated particle, determining the optimal particle position based on the initial fitness value and the fitness update value, and determining the hyperparameters of the convolutional neural network based on the optimal particle position. The hyperparameters include learning rate, convolutional kernel size, number of neurons, and number of convolutional kernels.

[0055] Optionally, the scalable dimension of the solution space can be determined according to the number of hyperparameters. The number of parameters included in each particle in the initial particle swarm corresponds to the value of the scalable dimension. Among them, the initial particle swarm can be generated by pseudo-random numbers or other random methods.

[0056] In one embodiment, the initial particle swarm is represented in the form of a matrix with a size of N rows and D columns, where N represents the size of the candidate solutions included in the initial particle swarm ("the size of the candidate solutions" is also referred to as "the population size", indicating how many candidate solutions are randomly generated during initialization. For example, if the population size is N, it means there are N particles (N candidate solutions)), and D represents the scalable dimension of the solution space. The expression of this initial particle swarm can be:

[0057]

[0058] In the formula, X(N, D) represents the initial particle swarm, UB and LB are the boundaries of the solution space, R is a sequence of random numbers with values in the range [0, 1], and X(N, D) represents the parameter in the Nth row and Dth column of the matrix.

[0059] Optionally, at the beginning of the iteration, a fitness evaluation is performed on each particle in the generated initial population (i.e., the initial particle swarm) (the fitness evaluation is performed at the beginning of the iteration). Among them, during the iteration process, each particle is brought into the process of building and training the convolutional neural network in the decision model. The parameters of the particles in the initial particle swarm are used one by one to build the convolutional neural network. After building the convolutional neural network, the test samples are input into the convolutional neural network, and the predicted values output by the decision model based on the built convolutional neural network are obtained. According to the predicted values, the test samples, and the true values corresponding to the test samples, the objective function is called to obtain the fitness value corresponding to each particle.

[0060] In one embodiment, the objective function can be:

[0061]

[0062] Wherein, is the predicted value corresponding to the ith test sample, yi is the true value corresponding to the ith test sample, and n is the number of test samples; the smaller the fitness value, the lower the prediction error and the better the model performance.

[0063] Optionally, after obtaining the fitness value corresponding to each particle in the initial particle swarm, record this fitness value and iteratively update the particles. Among them, the velocity and position of the particles are iteratively updated, including: updating the velocity and position of the particles according to the rotational motion and the auroral ellipse walking strategy; performing a particle collision strategy based on a preset probability to update the particle positions.

[0064] Optionally, the expression of the rotational motion can be:

[0065]

[0066] In the formula, C is the integration constant, and the charge q, mass m of the charged particle, and magnetic field strength B do not change. For simplicity, the values of C, q, and B can be 1, and m can be 100. The damping factor a can take a random value in [1, 1.5], and v(t) represents the velocity of the particle at time t. First, the rotational motion is manifested as the particle rotating along the Earth's magnetic field lines and moving slowly along a fixed trajectory. This motion pattern emphasizes local development and fine-tuning, aiming to explore the local solution space more deeply to find the local optimum or optimize the local structure of the current solution. This is consistent with the local search phase, in which minor adjustments and small steps are taken to improve the quality of the solutions and make them closer to the optimal solution.

[0067] Optionally, the aurora ellipse walk involves rapid movement around the candidate point to obtain the best solution or local optimum. This pattern emphasizes the global exploration feature, and the particle explores the solution space in larger steps to discover more valuable regions. This corresponds to the global exploration phase, in which the solution space is explored with larger step sizes to find the global optimum or a better solution.

[0068] Optionally, when updating the velocity and position of the particle using the combination of rotational motion and aurora ellipse walk, a new matrix is obtained based on the updated particle, and the relevant expression is:

[0069] X new (i, j) = X(i, j) + r2 × (W1 × v(t) + W2 × Ao)

[0070] Ao = Levy(d) × (X avg (j) - X(i, j)) + LB + r1 × (UB - LB) / 2;

[0071]

[0072] In the formula, X new (i, j) is the updated position of the parameter in the i-th row and j-th column of the matrix, X(i) represents the i-th particle, X(i, j) is the position of the parameter in the i-th row and j-th column of the matrix before update, r2 is the interference brought by factors such as the uncontrollable environment to the particle, and its value can be in [0, 1], r1 is a random number, t is the current iteration number, X avg represents the center of gravity of the particle swarm, and this center of gravity position reflects the distribution of the entire particle swarm in the solution space, which is of great significance for adjusting the movement direction and step size of the particle, X avg (j) represents the value of the center of gravity of the particle swarm in the j-th dimension, T is the maximum number of iterations, Levy(d) is the value sampled from the stable distribution, d is the step size, W1 and W2 are adaptive weights, and Ao represents the change of the aurora ellipse.

[0073] Optionally, to jump out of the local optimal solution, a particle collision strategy is used to change the positions of the particles. Among them, the particle collision strategy is executed based on a preset probability to update the particle positions, including: updating the particle positions through a particle collision formula, and the particle collision formula is:

[0074]

[0075] In the formula, X new (i, j) is the parameter in the i-th row and j-th column of the updated particle swarm, X(i, j) is the parameter in the i-th row and j-th column of the particle swarm before update, r3, r4, and r5 are random numbers with values in [0, 1], X(a, j) represents any parameter in the particle swarm, and K represents the collision probability between particles.

[0076] Optionally, multiple iterations are performed according to the above optimization method. After detecting that the maximum number of iterations is reached or a predetermined fitness threshold is reached, the iteration is stopped, and the optimal solution is determined according to the iteration result. After obtaining the optimal solution, the optimal solution is output, and this optimal solution is applied to the hyperparameter configuration in the convolutional neural network, thereby completing the construction of the convolutional neural network. The convolutional neural network is used for the construction and training of the decision-making module. After the training is completed, the job data encoded by the position is input into the decision-making model to obtain the ship operation prediction information output by the decision-making model. Among them, obtaining the ship operation prediction information output by the decision-making model includes: inputting the job data into the convolutional neural network to obtain the job vector sequence output by the convolutional neural network; obtaining the job prediction data output by the recurrent neural network based on the job vector sequence, and the job prediction data includes the hidden state sequence at different time steps; based on the weighted processing of the job prediction data by the attention module, obtaining the ship operation prediction information.

[0077] Optionally, the convolutional neural network adopts the optimal convolutional kernel size and the number of convolutional kernels determined by a preset optimization algorithm in the convolutional layer to perform high-dimensional and non-linear feature extraction on the job data. In the convolutional operation, each convolutional kernel slides and calculates the input features respectively and outputs multiple feature maps. These feature maps are then compressed and dimension-reduced by the pooling layer to obtain a one-dimensional vector, and the obtained one-dimensional vector is input into the fully connected layer. The fully connected layer integrates the key information extracted in the previous convolutional and pooling links, updates the network weights through backpropagation, and accelerates convergence by using the optimal learning rate obtained by the preset optimization algorithm. It can maximize the representation ability of the convolutional neural network, and at the same time ensure that the model parameter configuration and hyperparameter selection are optimal globally, providing a feature basis for the subsequent recurrent neural network and attention module to perform time series prediction and decision support.

[0078] Optionally, the convolutional neural network extracts features from the input job data, generates a sequence of job vectors based on the extracted features, and outputs the sequence of job vectors to the recurrent neural network. Among them, the recurrent neural network can be a BiGRU (Bidirectional Recurrent Neural Network), which consists of two independent GRUs (Gated Recurrent Unit). One processes data forward in time series, and the other processes data backward in time series. Through this bidirectional structure, the recurrent neural network can capture both the forward and backward information of the sequence data, thereby better understanding and predicting the patterns in the sequence. GRU is a gated recurrent neural network unit, and its formula includes an update gate, a reset gate, and a new candidate state. The calculation process of GRU can be as follows:

[0079] z t =σ(W z ·[h t-1 ,x t +b z ;

[0080] r t =σ(W r ·[h t-1 ,x t +b r ;

[0081]

[0082]

[0083] In the formula, z t is the update gate at time t, r t is the reset gate at time t, x t is the input at time t, h t-1 is the hidden state at time t - 1, h t is the output at time t, is the candidate hidden state, σ is the sigmoid activation function, tanh is the hyperbolic tangent function, W z 、W r 、W h are the weight matrices of GRU, b z 、b r 、b h are the bias vectors of GRU. BiGRU is a neural network structure composed of two GRU hidden layers in opposite directions, consisting of the forward hidden state vector of time t backward hidden state vector Composition. Each hidden layer receives the same input data at each moment, so it can extract feature vectors more comprehensively, improve the accuracy of feature extraction, and fully analyze historical data and feature factors to output job prediction data.

[0084] Optionally, the attention module obtains the importance of different time steps based on the attention mechanism, and weights the hidden state sequences of different time steps in the job prediction data according to this importance.

[0085] Optionally, the importance can be expressed as an attention coefficient. Among them, the weighting process includes: obtaining the attention coefficients of different time steps according to the hidden state sequence, and performing a weighting process based on the attention coefficients to obtain ship operation prediction information; the calculation formula of the attention coefficient is:

[0086]

[0087] In the formula, Wα represents the attention weight matrix; represents the hidden state sequence, is the forward hidden state vector at time t is the backward hidden state vector, softmax represents the normalized exponential function, and α t is the attention coefficient.

[0088] Optionally, after obtaining the ship operation prediction information through the weighting process, determine the ship operation strategy according to the ship operation prediction information and operation requirements (under what wave data can operate, the limit of cable force, the limit ship attitude under the operation state, etc.). The ship operation strategy includes strategy suggestions in aspects such as the operation time window, cable tension (the cable tension or pre-tension can be determined according to the predicted wave height or ship attitude), attitude adjustment (the center of gravity can be moved in advance and ballast water can be adjusted during periods with large waves), etc.

[0089] The ship offshore operation decision-making method of the present application has the following advantages:

[0090] (1) By collecting and processing offshore environment and ship state data in real time, this method can provide instant decision support in a dynamically changing environment and help the ship quickly adapt to different operation conditions.

[0091] (2) It can efficiently extract and fuse multi-source data, improve the perception and prediction ability of the decision-making model for complex environments, and thus improve the accuracy of decision-making.

[0092] (3) Through the dynamic analysis of environmental changes and ship states, it provides targeted adaptive decision-making suggestions, enabling the ship to respond flexibly under various complex and unexpected offshore operation conditions.

[0093] (4) By predicting potential risks in advance and formulating contingency plans, accidents and equipment failures that may occur during operations can be reduced, thereby improving operational safety and ensuring the safety of personnel and equipment.

[0094] (5) Since it can analyze in real time and make quick decisions, this method significantly reduces waiting time and adjustment time, improves the overall efficiency of ship operations, and reduces operating costs.

[0095] (6) It can utilize a large amount of operation data for decision-making model training and optimization, continuously improve the performance of the decision-making model, and ensure that the best decision-making support can be provided under different operating conditions.

[0096] In an alternative embodiment, an electronic device is provided, as Figure 2 shown, Figure 2 The electronic device 4000 shown includes: a processor 4001 and a memory 4003. Among them, the processor 4001 and the memory 4003 are connected, such as connected through a bus 4002. Optionally, the electronic device 4000 may further include a transceiver 4004, and the transceiver 4004 can be used for data interaction between this electronic device and other electronic devices, such as data sending and / or data receiving, etc. It should be noted that in practical applications, the transceiver 4004 is not limited to one, and the structure of the electronic device 4000 does not constitute a limitation to the embodiments of the present application.

[0097] The processor 4001 may be a CPU (Central Processing Unit, central processor), a general-purpose processor, a DSP (Digital Signal Processor, data signal processor), an ASIC (Application Specific Integrated Circuit, application-specific integrated circuit), an FPGA (Field Programmable Gate Array, field programmable gate array) or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. It can implement or execute various exemplary logic blocks, modules and circuits described in combination with the disclosure of the present application. The processor 4001 may also be a combination that implements computing functions, such as a combination including one or more microprocessors, a combination of a DSP and a microprocessor, etc.

[0098] The bus 4002 may include a path for transmitting information among the above components. The bus 4002 can be a PCI (Peripheral Component Interconnect) bus, an EISA (Extended Industry Standard Architecture) bus, or the like. The bus 4002 can be divided into an address bus, a data bus, a control bus, etc. For the sake of convenience in representation, only a thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0099] The memory 4003 can be a ROM (ReadOnlyMemory), or other types of static storage devices that can store static information and instructions, a RAM (RandomAccessMemory), or other types of dynamic storage devices that can store information and instructions. It can also be an EEPROM (Electrically Erasable ProgrammableReadOnlyMemory), a CD-ROM (CompactDiscReadOnlyMemory), or other optical disc storage, optical disc storage (including compact discs, laser discs, optical discs, digital versatile discs, Blu-ray discs, etc.), magnetic disk storage media, other magnetic storage devices, or any other medium that can be used to carry or store computer programs and can be read by a computer, which is not limited herein.

[0100] The memory 4003 is used to store the computer program for implementing the embodiments of the present application and is controlled by the processor 4001 to execute. The processor 4001 is used to execute the computer program stored in the memory 4003 to implement the steps shown in the foregoing method embodiments.

[0101] Among them, the electronic device can be any kind of electronic product that can perform human-computer interaction with an object. For example, a personal computer, a tablet computer, a smart phone, a personal digital assistant (PDA), a game console, an Internet Protocol Television (IPTV), a smart wearable device, etc.

[0102] The electronic device may further include a network device and / or an object device. Among them, the network device includes, but is not limited to, a single network server, a server group composed of multiple network servers, or a cloud composed of a large number of hosts or network servers based on cloud computing.

[0103] The network where the electronic device is located includes, but is not limited to, the Internet, wide area network, metropolitan area network, local area network, virtual private network (VPN), etc.

[0104] The embodiment of the present application provides a computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, the steps and corresponding contents of the foregoing method embodiment can be implemented.

[0105] The embodiment of the present application further provides a computer program product including a computer program, and when the computer program is executed by a processor, the steps and corresponding contents of the foregoing method embodiment can be implemented.

[0106] Those skilled in the art of the present technology can understand that the steps, measures, and solutions in the various operations, methods, and processes discussed in the present application can be alternated, changed, combined, or deleted. Further, the other steps, measures, and solutions in the various operations, methods, and processes discussed in the present application can also be alternated, changed, rearranged, decomposed, combined, or deleted. Further, the steps, measures, and solutions in the related technologies that are the same as those disclosed in the present application can also be alternated, changed, rearranged, decomposed, combined, or deleted.

[0107] In the description of the present application, the directions or positional relationships indicated by the words "center", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. are the exemplary directions or positional relationships based on the drawings, which are for the convenience of describing or simplifying the embodiments of the present application, rather than indicating or implying that the device or component referred to must have a specific orientation or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation to the present application.

[0108] The terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the present application, unless otherwise specified, the meaning of "plurality" is two or more.

[0109] In the description of the present application, it should be noted that unless otherwise clearly specified and defined, the terms "installed", "connected", and "coupled" should be understood in a broad sense. For example, it may be a fixed connection, a detachable connection, or an integral connection; it may be directly connected, or indirectly connected through an intermediate medium, and it may be the internal communication of two components. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to specific circumstances.

[0110] In the description of this specification, specific features, structures, materials, or characteristics may be combined in any one or more embodiments or examples in a suitable manner.

[0111] The above are only some implementation manners of this application. It should be noted that for those of ordinary skill in the art in this technical field, without departing from the technical concept of the solution of this application, other similar implementation means based on the technical idea of this application also fall within the protection scope of the embodiments of this application.

Claims

1. A decision-making method for ship offshore operations, characterized in that, The method includes: Obtaining the operation data of the ship; Inputting the operation data into a decision-making model, obtaining the ship operation prediction information output by the decision-making model, and generating a ship operation strategy according to the ship operation prediction information. The decision-making model includes a convolutional neural network, a recurrent neural network, and an attention module connected in sequence. The hyperparameters of the convolutional neural network are determined based on a preset optimization algorithm. The ship operation strategy includes at least one of an operation time window, a cable tension, and an attitude adjustment.

2. The shipboard offshore operation decision-making method according to claim 1, wherein The ship is a crane ship. The obtaining of the operation data of the ship includes: Determining that the ship is a crane ship, and collecting the operation data of the crane ship in the working state. The operation data includes wave data, ship motion response data, and cable force data; Preprocessing the operation data and adding position information to the operation data.

3. The ship offshore operation decision-making method according to claim 1, characterized in that, The preset optimization algorithm is the PLO algorithm. The determination of the hyperparameters includes: Randomly generating an initial particle swarm in the solution space, calling a target function to evaluate the fitness of each particle in the initial particle swarm, and obtaining an initial fitness value; Iteratively updating the velocity and position of the particles to generate a new particle swarm, obtaining the fitness update value corresponding to each updated particle, determining the optimal particle position according to the initial fitness value and the fitness update value, and determining the hyperparameters of the convolutional neural network based on the optimal particle position. The hyperparameters include a learning rate, a convolutional kernel size, the number of neurons, and the number of convolutional kernels.

4. The ship offshore operation decision-making method according to claim 3, wherein The iteratively updating the velocity and position of the particles includes: Updating the velocity and position of the particles according to the rotational motion and the aurora ellipse walking strategy; Performing a particle collision strategy based on a preset probability to update the particle position.

5. The ship offshore operation decision-making method according to claim 4, characterized in that The performing a particle collision strategy based on a preset probability to update the particle position includes: Updating the particle position through a particle collision formula. The particle collision formula is: X new (i, j) = X(i, j) + sin(r3 × π) × (X(i, j) - X(a, j)) where X new (i, j) is the parameter in the i-th row and j-th column of the updated particle swarm, X(i, j) is the parameter in the i-th row and j-th column of the particle swarm before update, r3 is a random number, and X(a, j) represents any parameter in the particle swarm.

6. The ship offshore operation decision-making method according to claim 1, wherein The obtaining of the ship operation prediction information output by the decision-making model includes: Inputting the operation data into the convolutional neural network and obtaining an operation vector sequence output by the convolutional neural network; Obtaining the operation prediction data output by the recurrent neural network based on the operation vector sequence. The operation prediction data includes a hidden state sequence at different time steps; Based on the weighted processing of the attention module on the operation prediction data, obtaining the ship operation prediction information.

7. The ship sea operation decision-making method according to claim 6, wherein, The weighted processing includes: Obtaining attention coefficients at different time steps according to the hidden state sequence, and performing weighted processing on the ship operation prediction data based on the attention coefficients to obtain the ship operation prediction information; The calculation formula of the attention coefficient is: Where, \(W_{\alpha}\) represents the attention weight matrix; represents the hidden state sequence, is the forward hidden state vector at time \(t\), is the backward hidden state vector, softmax represents the normalized exponential function, and \(\alpha\) t is the attention coefficient.

8. An electronic device, comprising a memory, a processor, and a computer program stored on the memory, characterized in that The processor executes the computer program to implement the steps of the method according to any one of claims 1-7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1-7.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1-7.

Images