Ship scheduling method, device, equipment and medium in multi-project marine operation

Abstract

The invention discloses a ship scheduling method, device and equipment in multi-project marine operation and a medium, and belongs to the field of marine operation scheduling. The method comprises the following steps: acquiring detailed information of an ocean operation project by identifying at least one construction design, inputting the information of the ocean operation project into a first operation scheduling model, generating first scheduling information, generating a first scheduling graph according to the first scheduling information, and acquiring real-time information of an operation area, and inputting the real-time information of the operation area and the first scheduling graph into a second operation scheduling model to generate second scheduling information, generating a second scheduling graph according to the second scheduling information, and finally performing ship scheduling according to the second scheduling graph. According to the method, the first operation scheduling model and the second scheduling operation model are constructed for the marine operation scheduling task in the multi-ship and multi-project mode, and the first scheduling graph and the second scheduling graph are generated according to the operation project information and the real-time information of the operation area, so that the real-time performance and the accuracy of marine operation scheduling are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of ship operation scheduling, and in particular to a ship scheduling method, device, equipment and medium in multi-project marine operations. Background Art

[0002] At present, the execution of marine operation projects involves single-ship single-project, single-ship multiple-project, multiple-ship single-project and multiple-ship multiple-project modes. For the multiple-ship multiple-project mode, multiple ships are usually equipped with tools and equipment to execute the corresponding projects, and the executable projects between ships have overlapping and dependent relationships.

[0003] In current actual operations, there are usually the following methods for job scheduling in a multi-ship and multi-project mode: The first is to rely mainly on the experience of the dispatcher, and the dispatcher obtains the scheduling plan manually. This method relies heavily on the subjective experience of the dispatcher, and it is difficult to obtain the optimal scheduling plan. This method requires manual calculation and review, which is inefficient. The second is to obtain a job scheduling model based on machine learning, neural networks or deep learning based on the experience of the comprehensive dispatcher and historical data training, and generate a scheduling plan from the job scheduling model. This method only uses simple machine learning, neural networks or deep learning models, and the accuracy of the scheduling plan generated by it needs further discussion. At the same time, this method does not take into account the real-time data of the operation area, so that the generated scheduling plan cannot meet the needs of real-time scheduling.

[0004] The preceding description is intended to provide general background information and does not necessarily constitute prior art. Summary of the invention

[0005] In view of this, an embodiment of the present invention provides a method, device, equipment and medium for scheduling ships in multi-project marine operations. A first aspect of the present invention provides a method for scheduling ships in multi-project marine operations, comprising the following steps:

[0006] Obtain information on marine operation projects;

[0007] Inputting the marine operation project information into a first operation scheduling model to generate first scheduling information;

[0008] generating a first scheduling graph according to the first scheduling information;

[0009] Obtain real-time information of the operation area;

[0010] Inputting the real-time information of the operation area and the first scheduling diagram into a second operation scheduling model to generate second scheduling information;

[0011] A second scheduling diagram is generated according to the second scheduling information; and the ship is scheduled according to the second scheduling diagram.

[0012] Furthermore, the acquisition of marine operation project information specifically includes the following steps:

[0013] Obtain at least one construction design plan, wherein the construction design plan has a plurality of sub-projects;

[0014] Identify the construction design plan and obtain the marine operation project information in the construction design plan; the marine operation project information includes multiple sub-project information and dependency information between sub-projects, and the sub-project information at least includes sub-project identification, operation area and sub-project workload.

[0015] Furthermore, the first job scheduling model includes an encoding network, a generation network and a decision network;

[0016] The inputting of the marine operation information into the first operation scheduling model to generate the first scheduling information specifically includes:

[0017] Inputting the plurality of sub-item information into an encoding network to obtain a plurality of sub-item vectors;

[0018] Input the plurality of sub-item vectors into a generation network, and generate a plurality of first scheduling sub-information corresponding to the plurality of sub-items through the generation network; the first scheduling sub-information at least includes operation vessel information, operation day information and window label information;

[0019] The plurality of first scheduling sub-information and dependency information between the sub-items are input into a decision network to generate first scheduling information.

[0020] Furthermore, the decision network adopts a multi-network fusion structure based on dependency, including an input layer, an embedding layer, a feature extraction layer, a first gating layer, a multi-block shared MLP layer, a linear layer and an output layer;

[0021] The step of inputting the plurality of scheduling sub-information and the dependency relationship information between the sub-items into the decision network to generate the first scheduling information specifically includes:

[0022] Inputting a plurality of first scheduling sub-information and dependency relationship information between sub-items through an input layer;

[0023] The first scheduling sub-information and the corresponding dependency information are converted into feature vectors by using an embedding layer to obtain a sub-information feature vector and a dependency feature vector;

[0024] The sub-information feature vector and the dependency feature vector are fused using the fusion layer to obtain a fused feature vector;

[0025] Using multiple feature extraction networks in the feature extraction layer to extract features from the fused feature vector, multiple initial feature vectors are obtained;

[0026] Using the first gating layer to generate first weight data corresponding to each feature extraction network, performing weighted summation on the initial feature vectors output by each feature extraction network to obtain a first combined feature vector;

[0027] Performing task migration processing on the first combined feature vector using multiple shared MLP layers to obtain a first migration feature vector;

[0028] The first migration feature vector is transformed by using a linear layer to obtain first scheduling information;

[0029] The output layer is used to output the first scheduling information.

[0030] Furthermore, the multi-block shared MLP layer is a multi-level block network structure, each block includes m MLPs and a shared layer, and the shared layer of the i-th block is used to extract the output features of the m MLPs in the current block to obtain m knowledge features, and MLP i k1 The output features of MLP i k2 The extracted knowledge features are fused as MLP i+1 k1 The input of , where m is the number of operating ships, MLP represents the multi-layer perceptron model, k2=1:m and k2≠k1.

[0031] Furthermore, the first scheduling information includes an operation vessel identifier, an operation project identifier, a project sub-identifier, an operation area, and an execution sequence corresponding to the project sub-identifier;

[0032] The generating of the first scheduling graph according to the first scheduling information is specifically:

[0033] The first scheduling graph is constructed with the operating vessel as the root node, the triple consisting of the operating project ID, the project sub-ID, and the operating area as the leaf node, and the execution sequence corresponding to the project sub-ID as the constraint condition.

[0034] Further, the second job scheduling model includes a first encoding layer, a second encoding layer, a feature fusion layer and a decoding layer;

[0035] The step of inputting the real-time information and the first scheduling diagram into the second job scheduling model to generate the second scheduling information specifically includes:

[0036] Encode the real-time information in the operation area using the first encoding layer to obtain a first encoding feature;

[0037] Encoding the first scheduling graph using a second encoding layer to obtain a second encoding feature;

[0038] The first coding feature and the second coding feature are concatenated using a feature fusion layer to obtain a fused feature;

[0039] The fused features are decoded using the decoding layer to obtain the second scheduling information.

[0040] The second aspect of the present invention discloses a ship scheduling device in multi-project marine operations, the device comprising:

[0041] The first acquisition module is used to acquire marine operation project information;

[0042] A first generating module, used for inputting the marine operation project information into a first operation scheduling model to generate first scheduling information;

[0043] A second generating module, used to generate a first scheduling graph according to the first scheduling information;

[0044] The second acquisition module is used to obtain real-time information of the operation area;

[0045] A third generating module, used for inputting the real-time information and the first scheduling diagram into the second job scheduling model to generate second scheduling information;

[0046] The fourth generating module is used to generate a second scheduling graph according to the second scheduling information.

[0047] A third aspect of the present invention discloses an electronic device, comprising a processor and a memory;

[0048] The memory is used to store programs;

[0049] The processor executes the program to implement the method for scheduling ships in multi-project marine operations.

[0050] A fourth aspect of the present invention discloses a computer-readable storage medium, wherein the storage medium stores a program, and the program is executed by a processor to implement the method for scheduling ships in multi-project marine operations.

[0051] The embodiment of the present invention also discloses a computer program product or a computer program, which includes a computer instruction stored in a computer-readable storage medium. A processor of a computer device can read the computer instruction from the computer-readable storage medium, and the processor executes the computer instruction, so that the computer device executes the above method.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] (1) The present invention proposes a two-stage method for obtaining an operation scheduling solution for marine operation scheduling tasks under a multi-ship and multi-project mode. The first stage is to obtain marine operation project information based on construction design identification, input the marine operation project information into the first operation scheduling model, and generate a first scheduling diagram. Since the first operation scheduling model is trained based on historical data, the first scheduling diagram generated in this stage is a static scheduling diagram. The second stage is to generate a second scheduling diagram through a second operation scheduling model based on real-time information in the operation area and the first scheduling diagram. This stage is essentially to use real-time information to make real-time corrections to the first scheduling diagram, so the second scheduling diagram generated is a dynamic scheduling diagram. Therefore, the present invention proposes a method of combining static scheduling diagrams and dynamic scheduling diagrams for marine operation scheduling tasks under a multi-ship and multi-project mode, which improves the real-time and accuracy of operation scheduling, and also improves the experience of scheduling personnel.

[0054] (2) When constructing the first operation scheduling model, the present invention takes into account that different operation ships can execute the same sub-project, and the execution of different sub-projects has dependencies, so a decision network based on a multi-network fusion structure of dependencies is constructed to obtain the first scheduling information. The decision network uses multiple blocks of shared MLP layers, and the shared layer in each block learns the knowledge features of other MLPs in the block, and fuses the knowledge features of other MLPs with the output features of the current MLP as the input of the MLP in the next block. Because of the integration of prior knowledge, the obtained scheduling plan is more in line with the multi-ship multi-project model.

[0055] The present invention constructs multiple generation sub-networks to generate first scheduling sub-information, and the first scheduling sub-information includes operation window label information. Multiple operation window label information is obtained by fine-grained division of operation time, so that the generated first scheduling sub-information is more in line with the needs of marine operation scheduling.

[0056] Additional aspects and advantages of the present invention will be given in part in the following description, and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0058] Figure 1 The present invention is a schematic diagram of the implementation process of a method for dispatching ships in multi-project marine operations.

[0059] Figure 2 The present invention is a schematic diagram of the architecture of marine operation scheduling in a method for scheduling ships in multi-project marine operations.

[0060] Figure 3 The present invention is a schematic diagram of an architecture for generating first scheduling information in a method for scheduling ships in multi-project marine operations.

[0061] Figure 4 The figure is a schematic diagram of a decision network in a method for dispatching ships in multi-project marine operations according to the present invention.

[0062] Figure 5 The schematic diagram of the architecture of multiple shared MLP layers in a method for scheduling ships in multi-project marine operations of the present invention is shown.

[0063] Figure 6 It is a schematic diagram of a first scheduling diagram in a ship scheduling method for multi-project marine operations of the present invention.

[0064] Figure 7 It is a schematic diagram of a second scheduling diagram in a ship scheduling method for multi-project marine operations of the present invention.

[0065] Figure 8 The present invention is a schematic structural diagram of a ship dispatching device for multi-project marine operations.

[0066] Fig. 9 The figure is a schematic diagram of an electronic device architecture of the present invention.

[0067] Fig.10 A schematic diagram of a computer-readable storage medium architecture of the present invention. DETAILED DESCRIPTION

[0068] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0069] At present, the execution of marine operation projects involves single-ship single-project, single-ship multiple-project, multiple-ship single-project and multiple-ship multiple-project modes. For the multiple-ship multiple-project mode, multiple ships are usually equipped with tools and equipment for executing corresponding projects, and the executable projects between ships have overlapping and dependent relationships. For example, ship A is equipped with equipment for gravity, magnetic, earthquake, and topographic detection, ship B is only equipped with equipment for earthquake detection, and ship C is equipped with equipment for geological sampling and gravity and magnetic detection. Since ships A and B can both perform earthquake detection, and ships A and C can both perform gravity and magnetic detection, the executable projects of ships A, B, and C have overlapping relationships. At the same time, in order to improve the safety and accuracy of sampling, it is usually necessary to conduct topographic detection before sampling, that is, before ship C performs the sampling task, ship A needs to complete topographic detection in the same operation area. This makes it very difficult to schedule operations in the multiple-ship multiple-project mode.

[0070] like Figure 1 As shown, the first embodiment of the present invention provides a method for scheduling ships in multi-project marine operations, comprising the following steps:

[0071] S1. Obtain information on marine operation projects.

[0072] Marine operation project information is usually recorded in the form of construction design documents, which record key information of the project, such as detailed project planning and technical specifications. Construction design documents may include but are not limited to project plans, technical specifications, drawings, charts, engineering specifications, etc., which record all aspects of the project in detail, including but not limited to project name, sub-project information, operation area, sub-project workload and dependencies between sub-projects. For example, whether sub-projects can be executed simultaneously, sub-project A is executed before sub-project B, or sub-project B is executed after sub-project A, etc. The present invention proposes to obtain marine operation project information by identifying multiple construction designs, thereby improving the efficiency of obtaining marine operation project information.

[0073] Specifically, Figure 2 As shown, the embodiment of the present invention obtains marine operation project information through a recognition network. The recognition network adopts an OCR network and a deep neural network. Among them, the OCR network is used to recognize PDF and image versions as text; the deep neural network is used to extract sub-project information in the text and the dependency information between sub-projects. The deep neural network can specifically adopt BILSTM, BIGRU network, etc. The extracted sub-project information includes sub-project name, operation area, sub-project workload and other information, and the dependency information includes predecessor tasks, successor tasks and the like.

[0074] By identifying at least one construction design, the marine operation project information obtained is shown in Table 1:

[0075] Table 1

[0076]

[0077]

[0078] S2. Input the marine operation project information into the first operation scheduling model to generate first scheduling information.

[0079] In the embodiment of the present invention, the first operation scheduling model is trained by storing historical marine operation project information and ship scheduling information; the trained first operation scheduling model can accept new marine operation project information as input and output reasonable first scheduling information.

[0080] Reference Figure 3 After the first operation scheduling model is trained, the first operation scheduling model and the identified marine operation project information are used to generate first scheduling information, which specifically includes:

[0081] S2-1. Input multiple sub-project information into the encoding network to obtain multiple sub-project vectors;

[0082] S2-2. The vectors of multiple sub-projects are input to the generation network, and the multiple first scheduling sub-information corresponding to the multiple sub-projects is generated by the generation network; the first scheduling sub-information includes the operation ship information, the operation days information and the window label information;

[0083] S2-3. Input the dependency data between the plurality of first scheduling sub-information and sub-items into the decision network to generate the first scheduling information.

[0084] Specifically, S2-1. Input multiple sub-project information into the coding network to obtain multiple sub-project vectors. In the embodiment of the present invention, the coding network is used to receive marine operation project information, including the sub-project information and dependency relationships extracted in step S1, and convert the input marine operation project information into one or more sub-project vectors representing key attributes of the sub-project information.

[0085] In the actual production process, considering the influence of various factors such as weather and waves, the operation cycle is divided into multiple continuous operation windows by month. For example, March to September of the current year is divided into the first window, and September of the current year to March of the next year is divided into the second window. The first window label and the second window label are used to identify the first window and the second window respectively.

[0086] Specifically, S2-2 inputs a plurality of sub-item vectors into a generation network, and generates a plurality of first scheduling sub-information corresponding to a plurality of sub-items through the generation network, including:

[0087] S2-2-1. Obtain the sub-project identifier corresponding to the sub-project vector and generate the sub-network label;

[0088] S2-2-2. Match the sub-project identifier and the label of the generated sub-network to obtain the generated sub-network;

[0089] S2-2-3. Input the sub-item vector into the generation sub-network to generate the first scheduling sub-information.

[0090] Exemplarily, the generation network includes multiple generation sub-networks, each of which is responsible for generating a specific type of scheduling information, such as start time, end time, resource allocation, etc.

[0091] In an embodiment of the present invention, when constructing a generative network, multiple generative subnetworks are constructed for multiple work boats. For example, a first generative subnetwork for work boat V1, a second generative subnetwork for work boat V2, and a third generative subnetwork for work boat V3 are constructed respectively, and labels are set for multiple generative subnetworks according to the sub-project categories executable by the work boats. For example, a label a is set for the first generative subnetwork, labels b, c, and d are set for the second generative subnetwork, and labels a, b, and d are set for the third generative subnetwork. The generative subnetwork adopts an adversarial generative network model.

[0092] Specifically, S2-3 inputs the dependency data between the plurality of first scheduling sub-information and sub-items into the decision network to generate the first scheduling information, including:

[0093] According to the sub-project identifier and the label of the generated sub-network, the corresponding generated sub-network is matched; for example, according to sub-project category a, the first generated sub-network and the third generated sub-network are matched, according to sub-project category b, the second generated sub-network and the third generated sub-network are matched, and according to sub-project category c, the second generated sub-network is matched.

[0094] Input the sub-item vector to the matching generative sub-network. For example, input the sub-item vector corresponding to the sub-item information (A, a, Ar1, 5000) to the first generative sub-network and the third generative sub-network, input the sub-item information (B, b, Ar2, 300) to the second generative sub-network and the third generative sub-network, and input the sub-item information (B, c, Ar2, 3000) to the second generative sub-network.

[0095] The first scheduling sub-information is outputted by generating a sub-network. The first scheduling sub-information includes a project identifier, a sub-project identifier, an operation area information, an operation ship information, a first window label and first operation day information, a second window label and second operation day information.

[0096] The first scheduling sub-information generated according to the marine operation information shown in Table 1 is shown in Table 2:

[0097] Table 2

[0098]

[0099]

[0100] After obtaining the first scheduling sub-information, the dependencies between the multiple first scheduling sub-information and the sub-projects are input into the decision network, and the first scheduling information is output through the decision network; the first scheduling information includes the execution sequence corresponding to the operation vessel, operation project identifier, project sub-identifier, operation area and project sub-identifier.

[0101] Reference Figure 4 The decision network proposed in the embodiment of the present invention adopts a multi-network fusion structure based on dependency, including an input layer, an embedding layer, a feature extraction network, a first gating layer, a multi-block shared MLP layer, a linear layer and an output layer, wherein:

[0102] The input layer is used to input a plurality of first scheduling sub-information and dependency relationship data between sub-items;

[0103] The embedding layer is used to convert the first scheduling sub-information and the corresponding dependency relationship data into a feature vector to obtain a sub-information feature vector and a dependency relationship feature vector;

[0104] The fusion layer is used to fuse the sub-information feature vector and the dependency feature vector to obtain a fused feature vector;

[0105] The fused feature vector is subjected to feature extraction using multiple feature extraction networks in the feature extraction layer to obtain multiple initial feature vectors; wherein the feature extraction layer may adopt a deep neural network model;

[0106] The first gating layer is used to generate first weight data corresponding to each feature extraction network, and perform weighted summation on the initial feature vectors output by each feature extraction network to obtain a first combined feature vector;

[0107] The multiple shared MLP layers are used to perform task migration processing on the first combined feature vector to obtain a first migration feature vector;

[0108] The linear layer is used to transform the first migration feature vector to obtain first scheduling information;

[0109] The output layer is used to output the first scheduling information.

[0110] Reference Figure 5 The multi-block shared MLP layer is a multi-level block network structure. Each block includes m MLPs and a shared layer. The shared layer of the i-th block is used to extract the output features of the m MLPs in the current block, obtain m knowledge features, and convert the MLP ik1 The output features of MLP i k2 The extracted knowledge features are fused as MLP i+1 k1 The input of , where m is the number of operating ships, MLP represents the multi-layer perceptron model, k2=1:m and k2≠k1.

[0111] Assume that the output of the MLPs in the i-th block is: X_MLP i 1 , ..., X_MLP i k1 , ..., X_MLP i m ,

[0112] The input is shared to the shared layer, which extracts knowledge and obtains the knowledge feature data of the i-th level as follows:

[0113] Y_MLP i 1 =EXTRACT(X_MLP i 1 ), ..., Y_MLP i k1 =EXTRACT(X_MLP i k1 ), ..., Y_MLP i m =EXTRACT(XMLP i m );

[0114] Then the feature data input to the MLPs in the i+1th level is:

[0115] {Input_MLP i+1 1} = concat(X_MLP i 1 ,Y_MLP i 2 ,……,Y_MLP i k1 ,……,Y_MLP i m );

[0116] {Input_MLP i+1 k1} = concat(X_MLP i k1 ,Y_MLP i 1 ,……,Y_MLP ik1-1 ,Y_MLP i k1+1 ,……,Y_MLP i m );

[0117] {Input_MLP i+1 m} = concat(X_MLP i m ,Y_MLP i 1 ,……,Y_MLP i k1 ,……,Y_MLP i m-1 ).

[0118] The decision network makes a decision on the first scheduling sub-information in Table 2, and obtains the first scheduling information as shown in Table 3:

[0119] Table 3

[0120] Ship number V1 V2 V3 (V1, A, a, Ar1, 1) (V2, C, c, d, Ar3, 1) (V3, D, d, Ar4, 1) (V1, C, a, Ar3, 2) (V2, B, c, Ar2, 2) (V3, B, a, Ar2, 1) (V3, B, b, Ar2, 3)

[0121] The decision network constructed in the embodiment of the present invention can learn the prior knowledge of the MLP in the previous block through multiple shared MLP layers, so that in the decision-making process, the computer results can be constrained by prior knowledge, making the calculated results more accurate and improving the accuracy of the result calculation.

[0122] S3. Generate a first scheduling diagram according to the first scheduling information;

[0123] In the embodiment of the present invention, after the first scheduling information is generated, the first scheduling diagram is constructed with the operating vessel as the root node, the triple consisting of the operating project identifier, the project sub-identifier, and the operating area as the leaf node, and the execution sequence corresponding to the project sub-identifier as the constraint condition, such as Figure 6 As shown, Figure 6 The implementation in represents the sub-item execution order for the same ship, and the dotted line represents the sub-item execution order between different ships, e.g. Figure 6 In the example, the operation vessel V3 first completes the operation area Ar2, project identifier B, and sub-project identifier a, then the operation vessel V3 completes the operation area Ar2, project identifier B, and sub-project identifier c, and finally the operation vessel V3 completes the operation area Ar2, project identifier B, and sub-project identifier b. By constructing the first scheduling diagram, the operation scheduling plan can be clearly reflected.

[0124] S4. Obtain real-time information of the operating area;

[0125] In the embodiment of the present invention, during the offshore operation, factors such as typhoons, wave heights, fisheries, and exercises will affect the smooth progress of the offshore operation. Since the first scheduling diagram is generated based on the marine operation project information and the first operation scheduling model, and the first operation scheduling model is obtained by training based on historical data, the essence of the first scheduling diagram is generated based on static data, which does not take into account the real-time information of the actual operation area, and therefore the first scheduling diagram cannot fully meet the needs of real-time scheduling.

[0126] In view of this problem, the present invention also uses the real-time information of the operation area and the second operation scheduling model to correct the first scheduling diagram to obtain a real-time second scheduling diagram. The real-time information of the operation area obtained specifically includes one or more of typhoon information, wave information, fishery information and exercise information in the operation area.

[0127] S5. Input the real-time information of the operation area and the first scheduling diagram into the second operation scheduling model to generate the second scheduling information;

[0128] In an embodiment of the present invention, the second job scheduling model includes a first coding layer, a second coding layer, a feature fusion layer and a decoding layer, wherein the first coding layer is used to encode real-time information in the operating area to obtain a first coding feature FeatureReal. During the encoding process, the operating area information is represented by an area identifier, for example, the operating area Ar1 is represented by 1, the operating area Ar2 is represented by 2, and the operating area Ar3 is represented by 3; the wave information is represented by the average wave height information in the operating area, for example, if the average wave height is 3m, it is represented by 3, and if the average wave height is 7m, it is represented by 7; typhoon information, fishery information, and exercise information are represented by 0 or 1, 0 indicates that there is no typhoon in the operating area, 1 indicates that there is a typhoon in the operating area, 0 indicates that there is no fishing activity in the operating area, 1 indicates that there is fishing activity in the operating area, 0 indicates that there is no exercise activity in the operating area, and 1 indicates that there is an exercise activity in the operating area;

[0129] For example, the real-time information obtained is: operating area Ar1, no typhoon, average wave height 1m, no fishing activities, no exercise activities, operating area Ar2, typhoon, average wave height 6m, no fishing activities, no exercise activities, operating area Ar3, no typhoon, average wave height 1m, no fishing activities, exercise activities, then the first coding feature FeatureReal after encoding by the first coding layer is: {(1,0,1,0,0); (2,1,6,0,0); (3,0,0,0,1)};

[0130] The second encoding layer is used to encode the first scheduling graph to obtain the second encoding feature. The second encoding layer can use a graph convolutional neural network based on an attention mechanism. The graph convolutional neural network based on the attention mechanism encodes each node in the first scheduling graph to obtain the hidden vector of each node, and then obtains the feature of the first scheduling graph based on the hidden vector of each node, and determines the feature of the first scheduling graph as the second encoding feature as Featuregraph;

[0131] The feature fusion layer is used to concatenate the first coded feature FeatureReal and the second coded feature Featuregraph to obtain the fused feature, FeatureConcat = FeatureReal⊕FeatureGraph;

[0132] The decoding layer is used to decode the fused features to obtain the second scheduling information. The decoding layer can adopt a multi-objective prediction model based on multiple multi-layer perceptrons, and obtain the second scheduling information by inputting the fused feature FeatureConcat into the decoding layer for decoding. The second scheduling information includes the execution order corresponding to the operation vessel, operation project identifier, project sub-identifier, operation area and project sub-identifier.

[0133] Assuming that when the operating vessel V3 performs sub-project a of the operating area Ar2, the real-time information obtained is that there will be fishing activities in the operating area Ar2, the second scheduling information obtained after adjustment based on the real-time information is shown in Table 4:

[0134] Table 4

[0135] Ship number V1 V2 V3 (V1, A, a, Ar1, 1) (V2, C, c, d, Ar3, 1) (V3, D, d, Ar4, 1) (V1, C, a, Ar3, 2) (V2, B, C, Ar2, 3) (V3, B, a, Ar2, 1) (V3, B, b, Ar2, 2)

[0136] S6. Generate a second scheduling diagram according to the second scheduling information; and perform scheduling operations on the ship according to the second scheduling diagram.

[0137] Reference Figure 7 After generating the second scheduling information, the operating vessel is taken as the root node, the triple consisting of the operating project identifier, the project sub-identifier, and the operating area is taken as the leaf node, and the execution sequence corresponding to the project sub-identifier is taken as the constraint condition to construct the second scheduling diagram; the ship is scheduled through the second scheduling diagram.

[0138] The ship scheduling method in multi-project marine operations provided by the present invention utilizes the first operation scheduling model and the second scheduling operation model to generate a scheduling plan that integrates real-time information, which is beneficial for dispatchers to accurately schedule marine operation tasks. At the same time, the present invention proposes a method for generating an operation scheduling plan suitable for a multi-ship multi-project mode, which further improves the generation accuracy of the scheduling plan by learning the knowledge characteristics of other MLP layers during the decision-making process through a decision network containing multiple shared MLP layers.

[0139] Reference Figure 8 The second embodiment of the present invention provides a marine operation scheduling device, the device comprising:

[0140] The first acquisition module 801 is used to acquire marine operation project information;

[0141] The first generating module 802 is used to input the marine operation project information into the first operation scheduling model to generate first scheduling information;

[0142] The second generating module 803 is used to generate a first scheduling graph according to the first scheduling information;

[0143] The second acquisition module 804 is used to acquire real-time information of the operation area;

[0144] The third generating module 805 is used to input the real-time information and the first scheduling diagram into the second job scheduling model to generate second scheduling information;

[0145] The fourth generating module 806 is used to generate a second scheduling diagram according to the second scheduling information, and perform scheduling operations on the ship according to the second scheduling diagram.

[0146] The first acquisition module 801, when acquiring the marine operation project information, includes:

[0147] Obtain at least one construction design;

[0148] At least one construction design is input into the identification network to identify and obtain marine operation project information, which includes information of multiple sub-projects and dependency information between sub-projects. The sub-project information includes at least sub-project identification, operation area and sub-project workload.

[0149] The first job scheduling model includes an encoding network, a generation network and a decision network;

[0150] The first generating module 802 inputs the marine operation project information into the first operation scheduling model to generate first scheduling information, including:

[0151] Inputting multiple sub-item information into the encoding network to obtain multiple sub-item vectors;

[0152] Inputting multiple sub-project vectors into a generation network, and generating multiple first scheduling sub-information corresponding to the multiple sub-projects through the generation network; the first scheduling sub-information includes operation ship information, operation day information and window label information;

[0153] The dependency relationship data between the plurality of first scheduling sub-information and the sub-items are input into the decision network to generate the first scheduling information.

[0154] The generation network includes a plurality of generation sub-networks, and the plurality of sub-item vectors are input into the generation network, and a plurality of first scheduling sub-information corresponding to the plurality of sub-items is generated through the generation network, including:

[0155] Get the sub-project identifier corresponding to the sub-project vector and generate the sub-network label;

[0156] Match the sub-project identifier and the label of the generated sub-network to obtain the generated sub-network;

[0157] The sub-item vector is input into the generation sub-network to generate the first scheduling sub-information.

[0158] The decision network adopts a multi-network fusion structure based on dependency, including input layer, embedding layer, feature extraction layer, first gating layer, multi-block shared MLP layer, linear layer and output layer;

[0159] Inputting the dependency relationship data between the plurality of scheduling sub-information and the sub-items into the decision network to generate the first scheduling information includes:

[0160] Using the input layer to input a plurality of first scheduling sub-information and dependency relationship information between sub-items;

[0161] The first scheduling sub-information and the corresponding dependency information are converted into feature vectors by using an embedding layer to obtain a sub-information feature vector and a dependency feature vector;

[0162] The sub-information feature vector and the dependency feature vector are fused using the fusion layer to obtain a fused feature vector;

[0163] Using multiple feature extraction networks in the feature extraction layer to extract features from the fused feature vector, multiple initial feature vectors are obtained;

[0164] Using the first gating layer to generate first weight data corresponding to each feature extraction network, performing weighted summation on the initial feature vectors output by each feature extraction network to obtain a first combined feature vector;

[0165] Performing task migration processing on the first combined feature vector using multiple shared MLP layers to obtain a first migration feature vector;

[0166] The first migration feature vector is transformed by using a linear layer to obtain first scheduling information;

[0167] The output layer is used to output the first scheduling information.

[0168] The multi-block shared MLP layer is a multi-level block network structure. Each block includes m MLPs and a shared layer. The shared layer of the i-th block is used to extract the output features of the m MLPs in the current block, obtain m knowledge features, and convert the MLP ik1 The output features of MLP i k2 The extracted knowledge features are fused as MLP i+1 k1 The input of , where m is the number of operating ships, MLP represents the multi-layer perceptron model, k2=1:m and k2≠k1.

[0169] The second generation module 803 generates a first scheduling diagram according to the first scheduling information, including: taking the operating vessel as the root node, the triple consisting of the operating project identifier, the project sub-identifier, and the operating area as the leaf node, and the execution sequence corresponding to the project sub-identifier as the constraint condition to construct the first scheduling diagram.

[0170] The second acquisition module 804 acquires real-time information of the operation area, including: acquiring one or more of typhoon information, wave information, fishery information and exercise information in the operation area.

[0171] The third generation module 805 inputs the real-time information and the first scheduling diagram into the second job scheduling model to generate second scheduling information, including:

[0172] Encode the real-time information in the operation area using the first encoding layer to obtain a first encoding feature;

[0173] Encoding the first scheduling graph using a second encoding layer to obtain a second encoding feature;

[0174] The first coding feature and the second coding feature are concatenated using a feature fusion layer to obtain a fused feature;

[0175] The fused features are decoded using the decoding layer to obtain the second scheduling information.

[0176] The fourth generation module 806 generates a second scheduling diagram according to the second scheduling information, including: taking the operating vessel as the root node, the triple consisting of the operating project identifier, the project sub-identifier, and the operating area as the leaf node, and the execution sequence corresponding to the project sub-identifier as the constraint condition to construct the second scheduling diagram.

[0177] The contents of the method in the first embodiment of the present invention are all applicable to the present device embodiment. The functions specifically implemented by the present device embodiment are the same as those of the above method embodiment, and the beneficial effects achieved are also the same as those achieved by the above method.

[0178] Fig. 9: is a schematic diagram of the structure of the electronic device proposed in the third embodiment of the present invention. The memory of this embodiment stores program instructions for implementing the ship scheduling method in multi-project marine operations of any of the above embodiments. The processor is used to execute the program instructions stored in the memory to perform ship scheduling for multi-project marine operations. Among them, the processor can also be called a CPU (Central Processing Unit). The processor may be an integrated circuit chip with signal processing capabilities. The processor can also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc.

[0179] The contents of the method in the first embodiment of the present invention are all applicable to the electronic device embodiment. The functions specifically implemented by the electronic device embodiment are the same as those of the above method embodiment, and the beneficial effects achieved are also the same as those achieved by the above method.

[0180] Fig.10 Schematic diagram of the structure of the computer-readable storage medium of the fourth embodiment of the present invention. The computer-readable storage medium of the fourth embodiment of the present invention stores program instructions that can implement the above-mentioned ship scheduling method in multi-project marine operations, wherein the program instructions can be stored in the above-mentioned storage medium in the form of a software product, including a number of instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) or a processor (processor) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned computer-readable storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or terminal devices such as a computer, a server, a mobile phone, and a tablet.

[0181] The contents of the method in the first embodiment of the present invention are all applicable to the computer-readable storage medium embodiment. The functions specifically implemented by the computer-readable storage medium embodiment are the same as those of the above method embodiment, and the beneficial effects achieved are also the same as those achieved by the above method.

[0182] This embodiment also provides a computer program product. When the computer program product is run on a computer, the computer is caused to execute the above-mentioned related steps to implement the ship scheduling method in multi-project marine operations provided by the above embodiment.

[0183] Those skilled in the art will appreciate that the modules in the devices in the embodiments of the present invention may be adaptively changed and arranged in one or more devices different from the embodiments. The modules or units or components in the embodiments of the present invention may be combined into one module or unit or component, and in addition they may be divided into a plurality of submodules or subunits or subcomponents. Except that at least some of such features and / or processes or units are mutually exclusive, all features disclosed in this specification (including corresponding claims, abstracts and drawings) and all processes or units of any method or device disclosed in this manner may be combined in any combination. Unless otherwise expressly stated, each feature disclosed in this specification (including corresponding claims, abstracts and drawings) may be replaced by an alternative feature providing the same, equivalent or similar purpose.

[0184] Through the description of the above implementation methods, technical personnel in the relevant field can understand that for the convenience and simplicity of description, only the division of the above-mentioned functional modules is used as an example. In actual applications, the above-mentioned functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0185] It should be noted that the logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection portion with one or more wirings (electronic device), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways if necessary, and then stored in a computer memory.

[0186] In addition, each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other. In particular, for embodiments such as devices and equipment, since they are basically similar to method embodiments, the relevant parts can refer to the partial description of the method embodiments. The embodiments of the devices and equipment described above are merely schematic, wherein the modules, units, etc. described as separate components may or may not be physically separated, that is, they may be located in one place, or they may be distributed in multiple places, such as nodes of a system network. Specifically, some or all of the modules and units may be selected according to actual needs to achieve the purpose of the above-mentioned embodiment scheme. Those skilled in the art can understand and implement it without paying creative labor.

[0187] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0188] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0189] In addition, the terms "first", "second", etc. used in the embodiments of the present invention are only used for descriptive purposes and should not be understood as indicating or implying relative importance, or implicitly indicating the number of technical features indicated in the present embodiment. Therefore, the features defined by the terms "first", "second", etc. in the embodiments of the present invention can explicitly or implicitly indicate that the embodiment includes at least one of the features. In the description of the present invention, the word "multiple" means at least two or two or more, such as two, three, four, etc., unless otherwise clearly and specifically defined in the embodiments.

[0190] In the embodiments of the present invention, the terms "include", "comprises" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element. In addition, components, features, and elements with the same name in different embodiments of the present invention may have the same meaning or different meanings, and their specific meanings need to be determined by their explanation in the specific embodiment or further combined with the context in the specific embodiment.

[0191] Although embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and cannot be construed as limitations of the present invention, and those of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present invention. Those skilled in the art will readily come to think of other embodiments of the present invention after considering the specification and practicing the present invention. This application is intended to cover any variations, uses or adaptations of the present invention, which follow the general principles of the present invention and include common knowledge or customary technical means in the art that are not disclosed in the present invention. The specification and embodiments are intended to be exemplary only, and the true scope and spirit of the present invention are indicated by the following claims.

Claims

1. A method for dispatching ships in multi-project marine operations, characterized in that: The following steps are involved: Obtain information on marine operation projects; Inputting the marine operation project information into a first operation scheduling model to generate first scheduling information; generating a first scheduling graph according to the first scheduling information; Obtain real-time information of the operation area; Inputting the real-time information of the operation area and the first scheduling diagram into a second operation scheduling model to generate second scheduling information; A second scheduling diagram is generated according to the second scheduling information; and the ship is scheduled according to the second scheduling diagram.

2. A method for dispatching ships in multi-project marine operations according to claim 1, characterized in that: The acquisition of marine operation project information specifically includes the following steps: Obtain at least one construction design plan, wherein the construction design plan has a plurality of sub-projects; Identify the construction design plan and obtain the marine operation project information in the construction design plan; the marine operation project information includes multiple sub-project information and dependency information between sub-projects, and the sub-project information at least includes sub-project identification, operation area and sub-project workload.

3. A method for dispatching ships in multi-project marine operations according to claim 2, characterized in that: The first job scheduling model includes an encoding network, a generation network and a decision network; The inputting of the marine operation information into the first operation scheduling model to generate the first scheduling information specifically includes: Inputting the plurality of sub-item information into an encoding network to obtain a plurality of sub-item vectors; Input the plurality of sub-item vectors into a generation network, and generate a plurality of first scheduling sub-information corresponding to the plurality of sub-items through the generation network; the first scheduling sub-information at least includes operation vessel information, operation day information and window label information; The plurality of first scheduling sub-information and dependency information between the sub-items are input into a decision network to generate first scheduling information.

4. A method for dispatching ships in multi-project marine operations according to claim 3, characterized in that: The decision network adopts a multi-network fusion structure based on dependency, including an input layer, an embedding layer, a feature extraction layer, a first gating layer, a multi-block shared MLP layer, a linear layer and an output layer; The step of inputting the plurality of scheduling sub-information and the dependency relationship information between the sub-items into the decision network to generate the first scheduling information specifically includes: Inputting a plurality of first scheduling sub-information and dependency relationship information between sub-items through an input layer; The first scheduling sub-information and the corresponding dependency information are converted into feature vectors by using an embedding layer to obtain a sub-information feature vector and a dependency feature vector; The sub-information feature vector and the dependency feature vector are fused using the fusion layer to obtain a fused feature vector; Using multiple feature extraction networks in the feature extraction layer to extract features from the fused feature vector, multiple initial feature vectors are obtained; Using the first gating layer to generate first weight data corresponding to each feature extraction network, performing weighted summation on the initial feature vectors output by each feature extraction network to obtain a first combined feature vector; Performing task migration processing on the first combined feature vector using multiple shared MLP layers to obtain a first migration feature vector; The first migration feature vector is transformed by using a linear layer to obtain first scheduling information; The output layer is used to output the first scheduling information.

5. A method for dispatching ships in multi-project marine operations according to claim 4, characterized in that: The multi-block shared MLP layer is a multi-level block network structure, each block includes m MLPs and a shared layer, the shared layer of the i-th block is used to extract the output features of the m MLPs in the current block, obtain m knowledge features, and convert the MLP i k1 The output features of MLP i k2 The extracted knowledge features are fused as MLP i+1 k1 The input of , where m is the number of operating ships, MLP represents the multi-layer perceptron model, k2=1:m and k2≠k1.

6. A method for dispatching ships in multi-project marine operations according to claim 1, characterized in that: The first scheduling information includes the operation vessel identifier, the operation project identifier, the project sub-identifier, the operation area and the execution sequence corresponding to the project sub-identifier; The generating of the first scheduling graph according to the first scheduling information is specifically: The first scheduling graph is constructed with the operating vessel as the root node, the triple consisting of the operating project ID, the project sub-ID, and the operating area as the leaf node, and the execution sequence corresponding to the project sub-ID as the constraint condition.

7. A method for dispatching ships in multi-project marine operations according to claim 1, characterized in that: The second job scheduling model includes a first encoding layer, a second encoding layer, a feature fusion layer and a decoding layer; The step of inputting the real-time information and the first scheduling diagram into the second job scheduling model to generate the second scheduling information specifically includes: Encode the real-time information in the operation area using the first encoding layer to obtain a first encoding feature; Encoding the first scheduling graph using a second encoding layer to obtain a second encoding feature; The first coding feature and the second coding feature are concatenated using a feature fusion layer to obtain a fused feature; The fused features are decoded using the decoding layer to obtain the second scheduling information.

8. A ship dispatching device for multi-project marine operations, characterized in that: The device comprises: The first acquisition module is used to acquire marine operation project information; A first generating module, used for inputting the marine operation project information into a first operation scheduling model to generate first scheduling information; A second generating module, used to generate a first scheduling graph according to the first scheduling information; The second acquisition module is used to obtain real-time information of the operation area; A third generating module, used for inputting the real-time information and the first scheduling diagram into the second job scheduling model to generate second scheduling information; The fourth generating module is used to generate a second scheduling diagram according to the second scheduling information; and perform scheduling operations on the ship according to the second scheduling diagram.

9. An electronic device, characterized in that: including a processor and a memory; The memory is used to store programs; The processor executes the program to implement a method for scheduling ships in multi-project marine operations as described in any one of claims 1-8.

10. A computer-readable storage medium, characterized in that: The storage medium stores a program, and the program is executed by a processor to implement a method for scheduling ships in multi-project marine operations according to any one of claims 1 to 8.

Images