CN117592239B - Multi-objective optimization optical cable network route intelligent planning method and system - Google Patents
Multi-objective optimization optical cable network route intelligent planning method and system Download PDFInfo
- Publication number
- CN117592239B CN117592239B CN202410067768.4A CN202410067768A CN117592239B CN 117592239 B CN117592239 B CN 117592239B CN 202410067768 A CN202410067768 A CN 202410067768A CN 117592239 B CN117592239 B CN 117592239B
- Authority
- CN
- China
- Prior art keywords
- optical cable
- optimized
- planning
- cost
- route
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 230000003287 optical effect Effects 0.000 title claims abstract description 223
- 238000000034 method Methods 0.000 title claims abstract description 71
- 238000005457 optimization Methods 0.000 title claims abstract description 31
- 238000010586 diagram Methods 0.000 claims abstract description 29
- 238000007781 pre-processing Methods 0.000 claims abstract description 22
- 230000002787 reinforcement Effects 0.000 claims abstract description 16
- 230000003993 interaction Effects 0.000 claims abstract description 6
- 239000003795 chemical substances by application Substances 0.000 claims description 22
- 238000012549 training Methods 0.000 claims description 17
- 239000002994 raw material Substances 0.000 claims description 13
- 238000012876 topography Methods 0.000 claims description 12
- 239000000835 fiber Substances 0.000 claims description 11
- 239000013535 sea water Substances 0.000 claims description 11
- 238000013528 artificial neural network Methods 0.000 claims description 7
- 238000010606 normalization Methods 0.000 claims description 7
- 238000012545 processing Methods 0.000 claims description 3
- 238000011478 gradient descent method Methods 0.000 claims description 2
- 230000008569 process Effects 0.000 description 16
- 230000009471 action Effects 0.000 description 13
- 230000007547 defect Effects 0.000 description 7
- 230000008901 benefit Effects 0.000 description 6
- 238000004891 communication Methods 0.000 description 6
- 239000010410 layer Substances 0.000 description 6
- 230000005540 biological transmission Effects 0.000 description 5
- 238000013508 migration Methods 0.000 description 5
- 230000005012 migration Effects 0.000 description 5
- 238000010276 construction Methods 0.000 description 4
- 238000010801 machine learning Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 3
- 125000003275 alpha amino acid group Chemical group 0.000 description 2
- 238000013459 approach Methods 0.000 description 2
- 150000001875 compounds Chemical class 0.000 description 2
- 238000013480 data collection Methods 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 230000001681 protective effect Effects 0.000 description 2
- 239000002356 single layer Substances 0.000 description 2
- 238000012546 transfer Methods 0.000 description 2
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000009412 basement excavation Methods 0.000 description 1
- 210000004027 cell Anatomy 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000003631 expected effect Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/18—Network design, e.g. design based on topological or interconnect aspects of utility systems, piping, heating ventilation air conditioning [HVAC] or cabling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0499—Feedforward networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/096—Transfer learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/06—Multi-objective optimisation, e.g. Pareto optimisation using simulated annealing [SA], ant colony algorithms or genetic algorithms [GA]
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Geometry (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computing Systems (AREA)
- Data Mining & Analysis (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Computer Hardware Design (AREA)
- Mathematical Optimization (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Analysis (AREA)
- Computational Mathematics (AREA)
- Computer Networks & Wireless Communication (AREA)
- Telephonic Communication Services (AREA)
Abstract
The invention discloses a multi-objective optimization optical cable network route intelligent planning method and a system, wherein the method comprises the following steps: obtaining geographic data and demand information of a new optical cable target area; modeling is carried out according to a plurality of targets to be optimized of the geographic data of the new optical cable target area, and a distribution diagram of the targets to be optimized is obtained; preprocessing by combining the demand information and distribution diagrams of a plurality of targets to be optimized; transferring a pre-established and trained experience learning model as an experience decision model to an intelligent planning model of an optical cable network route, inputting the preprocessed data into the intelligent planning model, adopting a reinforcement learning method, completing optical cable route planning and optimization from a starting point to a terminal point through interaction and searching of an intelligent body and a reinforcement learning environment, and outputting a multi-target optimized optical cable route; the experience learning model is trained based on the geographic data of the existing optical cable routing area and the corresponding optical cable routing data and is used for obtaining an optical cable network path planning strategy of a manual expert.
Description
Technical Field
The invention belongs to the field of optical cable network route planning, and particularly provides an intelligent planning method and system for a multi-objective optimized optical cable network route.
Background
The optical cable has the advantages of large transmission capacity, low transmission loss, small transmission delay and the like, and is widely applied to transmission networks of various communication scenes. For example, in a cross-ocean communication scenario, more than 99% of international traffic is transmitted over ocean cables; on land, fiber optic cable communications are widely used as a transmission technology for core backbone networks, convergence networks, and access networks. With the continuous promotion of broadband strategy, optical fiber reaches home to be popularized continuously, optical communication technology is popularized and applied to an access layer continuously, and optical cable communication technology plays an increasingly important role in information communication. Therefore, the method has important significance in planning in detail at the initial stage of the construction of the optical cable network so that each object of the optical cable network can reach the optimal expected effect in the whole life cycle.
In the optical cable network planning work, the routing planning of the optical cable network is one of the key links, the construction cost of the optical cable network in the initial stage, the safety of operation in the whole life cycle and the maintenance cost have profound effects, and the optical cable network cost and the safety have inherent conflict, so that the optimization effect is difficult to achieve at the same time. Taking the routing planning of the submarine optical cable as an example, the shortest path is adopted to lay the submarine optical cable, so that the cost of raw materials and the construction cost can be reduced, however, the shortest path planning mode is adopted to inevitably pass through risk factors such as channels, earthquakes, volcanoes and the like, so that the safety of an optical cable network is reduced; the submarine cable can be effectively improved in the risk resisting capacity by bypassing the high risk area and additionally installing protective measures on the submarine cable, so that the safety of the optical cable network is improved, but the cost of the optical cable network is increased. Therefore, how to realize the multi-objective optimization of the cost and the safety of the optical cable network through the optical cable network route planning has important application value.
Currently, there are mainly two ways of routing optical cable networks. An expert manually determines the path of the cable network segment by segment between the start point and the end point of the cable network according to his own experience. The routing planning method based on expert experience has the characteristic of flexible strategy, and in the actual situation, the path planned according to engineering experience accords with the engineering expected quality, however, the method for planning the optical cable network path manually by the expert has the defects of large workload and low efficiency. Another way is to evaluate various risk factors into cost or risk by numerical modeling on various factors based on Geographic Information System (GIS) data, then transform cost and risk multi-objective into single objective by weighted summation, and finally plan physical route using heuristic algorithm, fast travelling algorithm, etc. GIS-based cable routing has the advantage of high efficiency, however, geographic Information System (GIS) data is based on modeling cost and risk values according to fixed rules, which is represented by the importance of measuring various risk factors, cost and risks by using fixed weights. However, in actual planning, a human expert needs to make a decision according to actual situations, and the importance of various optimization targets is not unchanged along the planned path, but the position of the optical cable route is comprehensively estimated and determined according to various factors of the geographic position. In addition, the route planning method based on GIS data and mathematical modeling searches and plans from the beginning every time, and the existing manual expert experience is not fully utilized, so that the defects of long route planning process and long time consumption are caused. Heretofore, a large number of optical cables are laid on the ocean and the land, the routes of the optical cables are all based on engineering experience of manual experts, the optical cable paths determined after comprehensive evaluation are optical cable paths with a large number of condensed expert experiences, and practical engineering verification proves feasible experience, but the experience is not fully utilized.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provides an intelligent planning method and system for a multi-objective optimized optical cable network route.
In order to achieve the above purpose, the present invention provides a multi-objective optimized optical cable network route intelligent planning method, which includes:
obtaining geographic data and demand information of a new optical cable target area;
modeling is carried out according to a plurality of targets to be optimized of the geographic data of the new optical cable target area, and a distribution diagram of the targets to be optimized is obtained;
preprocessing by combining the demand information and distribution diagrams of a plurality of targets to be optimized;
Transferring a pre-established and trained experience learning model as an experience decision model to an intelligent planning model of an optical cable network route, inputting the preprocessed data into the intelligent planning model, adopting a reinforcement learning method, completing optical cable route planning and optimization from a starting point to a terminal point through interaction and searching of an intelligent body and a reinforcement learning environment, and outputting a multi-target optimized optical cable route;
The experience learning model is trained based on the geographic data of the existing optical cable routing area and the corresponding optical cable routing data and is used for obtaining an optical cable network path planning strategy of a manual expert.
Preferably, the new optical cable target area geographic data includes: altitude, topography, and grade; the demand information includes a start point and an end point of a new cable route.
Preferably, modeling is performed on a plurality of targets to be optimized according to the geographic data of the new optical cable target area, so as to obtain distribution diagrams of the plurality of targets to be optimized; the method specifically comprises the following steps:
when the new optical cable target area is the sea, modeling is carried out according to the cost, and the total cost of the sea cable is obtained through summation operation by the raw material cost and the sea cable deployment cost, wherein the raw material cost is different according to the interval where the sea water depth is positioned, the cost of the light sea cable is multiplied by the different coefficients, and the sea cable deployment cost is different according to the interval where the sea water depth is positioned, and the different deployment costs are corresponding;
According to the elevation diagram of the target area, the total cost of the submarine cable is combined, and a total cost distribution diagram of the submarine cable in the target area is calculated;
Modeling according to risk, and dividing risk factors of gradient of submarine topography on submarine cable :
Wherein,A grade value representing the terrain;
combining risk factors of submarine cables according to terrain gradient of target area And calculating to obtain a submarine topography gradient versus submarine risk factor distribution map.
Preferably, modeling is performed on a plurality of targets to be optimized according to the geographic data of the new optical cable target area, so as to obtain distribution of the plurality of targets to be optimized; the method specifically comprises the following steps:
When the new optical cable target area is land, modeling is carried out according to the cost, and the total cost of the land optical cable is obtained through summation operation by the raw material cost and the deployment cost of the land optical cable; the deployment cost of the land optical cable is different according to the laying areas;
According to the elevation diagram of the target area, combining the total cost of the land optical cable, and calculating to obtain a total cost distribution diagram of the land optical cable of the target area;
Modeling based on risk, risk factors for land cable :
Wherein,Representation normalization,/>Representing the distribution of the kth risk factors in the target area, and P represents the category number of the risk factors;
Combining risk factors of land optical cables according to terrain gradient of target area And calculating to obtain the risk factor distribution affecting the land optical cable.
Preferably, the pretreatment includes a normalization process.
Preferably, the empirical learning model is a neural network, inputs are a total cost distribution diagram and a risk factor distribution diagram of eight adjacent positions of the position i acquired in a nine-grid mode, a linear distance between a current point and a terminal point, and an included angle formed by connecting the current point with the terminal point and connecting the current point with a next planning point, and outputs the included angle as the position of the next planning point.
Preferably, the training step of the empirical learning model includes:
Reading geographic data and routing data of the existing optical cable routing area;
Modeling is carried out according to a plurality of targets to be optimized of the geographic data of the existing optical cable routing area, and distribution of the targets to be optimized is obtained;
preprocessing by combining the route data and distribution graphs of a plurality of targets to be optimized;
And inputting the preprocessed data, the linear distance between the current point and the terminal point and the included angle formed by connecting the current point with the terminal point and connecting the current point with the next planning point into an experience learning model, and training by using a gradient descent method until the training requirement is met, so as to obtain a trained experience learning model.
Preferably, the processing procedure of the intelligent planning model for optical cable network routing comprises the following steps:
For geographical locations Cost after pretreatment/>And risk/>The position/>, is obtained byEnvironmental value/>：
In the method, in the process of the invention,Representing a large value;
In planning, the intelligent agent continuously interacts with the learning environment to maximize the expected rewards obtained by the intelligent agent; wherein, the rule that the agent obtains the reward sets up as: when the intelligent agent successfully searches the terminal, rewarding 5; the intelligent agent searches out boundary rewards-1 and rewards 0 in other cases;
And when more than one feasible path solutions exist, performing dominant relation analysis on all the feasible path solutions to obtain the multi-objective optimized optical cable route.
Preferably, the subjecting all feasible path solutions to dominant relationship analysis includes:
when the total cost and total risk of one feasible path solution is less than the total cost and total risk of the other feasible path solutions, the feasible path solution is a multi-objective optimized cable route;
When neither the total cost nor the total risk of any one feasible path solution can be less than the total cost and total risk of the other feasible path solutions at the same time, then all feasible path solutions are multi-objective optimized cable routes.
On the other hand, the invention provides an intelligent planning system for the multi-objective optimized optical cable network route, which comprises the following components: the system comprises an acquisition module, a modeling module, a preprocessing module, an optimization output module, an empirical learning model and an intelligent planning model for optical cable network routing, wherein,
The acquisition module is used for acquiring geographic data and demand information of a new optical cable target area;
the modeling module is used for modeling according to a plurality of targets to be optimized of the geographic data of the new optical cable target area to obtain distribution diagrams of the targets to be optimized;
the preprocessing module is used for preprocessing the distribution graphs of the demand information and the multiple targets to be optimized;
The optimization output module is used for transferring a pre-established and trained experience learning model as an experience decision model to an intelligent planning model of the optical cable network route, inputting the preprocessed data into the intelligent planning model, adopting a reinforcement learning method, completing the optical cable route planning and optimization from a starting point to a terminal point through interaction and searching of an intelligent body and a reinforcement learning environment, and outputting a multi-objective optimized optical cable route;
The experience learning model is trained based on the geographic data of the existing optical cable routing area and the corresponding optical cable routing data and is used for obtaining an optical cable network path planning strategy of a manual expert.
Compared with the prior art, the invention has the advantages that:
1. The invention provides an empirical learning model, and the empirical learning model is trained by utilizing the optical cable path data which is manually designed by the prior expert experience, so that the empirical learning model has a strategy for simulating the routing decision of an optical cable network by an artificial expert under various cost and risk conditions, and the problem that the relative importance of different optimization targets is measured by adopting fixed rules, so that the local area of a planned route cannot flexibly adapt to engineering requirements is avoided;
2. The invention provides a multi-objective optimization optical cable network route intelligent planning method for transferring expert experience, which utilizes an artificial intelligence method to learn the experience of an artificial expert from optical cable path data which is manually designed by the prior expert experience, and transfers the learned experience of the artificial expert into the optical cable route intelligent planning method. The experience learning model can be reused after one training, so that the problems of slow convergence and long planning time existing in the conventional optical cable network routing algorithm which needs to search the optimal path from the beginning each time are avoided.
Drawings
FIG. 1 is a schematic diagram of a multi-objective optimized cable network route intelligent planning method;
FIG. 2 is a flow chart of a multi-objective optimized cable network route intelligent planning method;
FIG. 3 is an organizational block diagram of an embodiment;
FIG. 4 is a schematic view of a mobile nine-grid window;
FIG. 5 is a schematic diagram of the collection of existing cable routing data;
FIG. 6 is a schematic diagram of an empirical learning model training;
Fig. 7 is a schematic diagram of an intelligent planning model for optical cable network routing.
Detailed Description
Aiming at the defects of the conventional optical cable network route planning method, the invention provides an intelligent multi-objective optimization optical cable network route planning method and system. The invention firstly establishes a model for learning the experience of an optical cable routing planning expert, which is called an experience learning model. Training an experience learning model by utilizing the existing optical cable path data which is manually designed through expert experience; and secondly, establishing an optical cable network route planning model, which is called a route planning model. And transferring the experience learning model to a route planning model so as to form the multi-objective optimized optical cable network route intelligent planning based on the experience of the transfer expert, which is called as an optical cable network route intelligent planning model for convenience. In the training experience learning model, through the routing data of the existing optical cable network, the experience learning model approximates to the optical cable routing path decision strategy of different optimization targets under different scenes, and the importance of the different optimization targets is not required to be determined by using fixed rules. The invention stores the trained experience learning model parameters, and uses the model directly by migration each time when planning a new optical cable network path, without the need of head training. Therefore, the method can effectively accelerate the convergence process of the intelligent planning model of the optical cable network route, and improve the planning efficiency, thereby solving the defects of the conventional optical cable network route planning method.
In order to achieve the above purpose, a multi-objective optimized optical cable network route intelligent planning method is provided, and the principle is schematically shown in fig. 1. The method comprises the steps of geographical data of an existing optical cable routing area, existing optical cable routing data, optical cable network cost and risk modeling, data acquisition and preprocessing, an empirical learning model, a new optical cable routing target area range, new optical cable routing information, an optical cable network routing intelligent planning model and new optical cable network routing information.
More specifically, the geographical data of the existing optical cable routing area refers to the related data of the geographical position of the optical cable still in use, which is manually designed by a manual expert according to experience, and the optical cable has the advantages of good cost benefit and low risk through engineering inspection, and the geographical data of the existing optical cable routing area comprises the height and gradient of the topography;
The existing optical cable routing data refer to the geographical position of the optical cable which is manually designed by a manual expert according to experience, and comprises longitude and latitude information of the optical cable routing;
The optical cable network cost and risk modeling means that various factors influencing the optical cable routing cost and risk are converted into targets to be optimized for cost and risk through data modeling, wherein the various factors comprise gradient and altitude;
the data acquisition and preprocessing refers to acquiring relevant data around a given geographic position according to the geographic position, wherein the relevant data comprise cost and risk. The data preprocessing comprises normalization;
The empirical learning model is a machine learning model trained using the data collection and pre-processed data, the machine learning model comprising a fully connected neural network. Because the training optical cable routing data used by the experience learning model is manually designed according to expert experience, the experience learning model can mathematically fit the approximation of complex relative relation between cost and risk through the training of the data, so that the experience learning model has the capability of making decisions by utilizing the experience of manual experts;
the new optical cable routing target area range data are related data of a target area where a user plans to lay a new optical cable, and the related data comprise altitude and gradient of a geographic position;
The new optical cable demand information refers to information related to the optical cable to be laid by a user and comprises a starting point and an ending point of a new optical cable network route;
The optical cable network routing intelligent planning model is a machine learning model which receives the experience learning model migration data and has artificial expert decision making capability, and the decision-making machine learning model comprises reinforcement learning. The capacity of making decisions by the experience of the manual expert provided by the experience learning model enables the intelligent planning model of the optical cable network route to also have the capacity of making decisions by the experience of the manual expert through migration;
The new optical cable network routing information is the new optical cable network routing information planned by the optical cable network routing intelligent planning model according to the new optical cable demand information and the new optical cable routing target area range data, and comprises the geographic position information of the new optical cable network and the longitude and latitude of the new optical cable network routing.
It should be noted that the empirical learning model can be reused after one training, and the intelligent planning model for optical cable network routing does not need to be trained from scratch each time a new optical cable routing is planned.
The invention provides a multi-objective optimization optical cable network route intelligent planning method and system, which comprise an expert experience learning model, wherein on the basis of the existing optical cable route information and the related geographic information, various risk factors in the existing optical cable information and the geographic area are converted into the cost and risk targets to be optimized of the optical cable route through numerical modeling, and the cost and risk targets are used for training the expert experience learning model through data acquisition and preprocessing. The expert experience learning model learns strategies of manual expert on optical cable routing planning decisions under different cost and risk conditions from the existing optical cable routing information and the related geographic information. Meanwhile, the cost and risk influence of various factors in the area on the new optical cable network are calculated through numerical modeling by reading the information of the new optical cable demand of the user and the information of the geographical range of the relevant area, and the data are extracted through data acquisition and preprocessing and used for the routing planning of the new optical cable network. And transferring the expert experience learning model to an intelligent planning model of the optical network route, and planning information of the new optical network route by combining the acquired information of the new optical network route and the geographic range of the relevant area. The working flow chart of the multi-objective optimized optical cable network route intelligent planning method based on migration expert experience is shown in figure 2.
The technical scheme of the invention is described in detail below with reference to the accompanying drawings and examples.
Example 1
The embodiment 1 of the invention provides an intelligent planning method for a multi-objective optimized optical cable network route.
In an embodiment, the empirical learning model employs a deep neural network and the new cable routing model employs a reinforcement learning framework. A schematic of the structural framework of an embodiment is shown in fig. 3.
Step one, reading data
The existing optical cable routes geographic data of an area, including GIS data such as altitude, gradient and the like of the area, and the data can be acquired by GIS data websites disclosed by a network; and reading the routing data of the existing optical cable network, wherein the data can be obtained through a public website.
Step two, optical cable network cost and risk modeling
Optical cable network cost and risk modeling is mainly to model various risk factors of a geographical area so as to quantify the influence of the various risk factors on the cost and risk of the optical cable network. The method is suitable for the scene of land and ocean optical cable network routing planning, and the emphasis of land and ocean in optical cable network routing planning is slightly different. Therefore, the step is summarized by modeling the cost and risk of the optical cable network for subdividing the land and ocean scenes, the area where the new optical cable is planned to be laid is marked as D, and a geographic information system is modeled by adopting a grid method.
1. Marine cable network cost and risk modeling
1) Cost modeling
The marine cable network cost modeling includes raw material cost and deployment cost modeling. Different sea depths require the use of different types of fiber optic cable, so the raw material cost of the sea cable is closely related to the sea depth, assuming the price of the light sea cable for ease of illustrationIs/>. When the sea water depth exceeds 1000 meters, a light sea cable is used; when the water depth is within the range of 200-1000 meters, a single-layer armored submarine cable is used; when the water depth is less than 200 meters, a double-layer armored protection type submarine cable is used. Assuming that the prices of single-layer and double-layer armored submarine cables are 1.3 and 1.6 times that of light submarine cables, respectively, the raw material cost of submarine cables can be modeled using the following formula:
in the formula (1), Represents the cost of raw materials, and H represents the depth of seawater.
The difficulty of laying submarine cables in different submarine environments is different, so that the deployment cost of the submarine cables is different. Such as submarine cables in shallow sea areas near landing sites, have a high probability of being affected by human activity. Thus, the submarine cable needs to be buried for protection, resulting in higher radiation costs than in deep sea areas. It is assumed that the cost when submarine cables are laid in areas with depths exceeding 1200 meters isWhen the sea water depth is less than 1200 meters, the deployment cost of the sea cable is/>，/>Indicating the buried depth of the submarine cable. Thus, deployment cost of submarine cable/>Can be expressed as:
In formula (2), H represents the sea water depth. When the sea water depth is within the range of 200-1200 meters, Taking 1 meter; when the sea water depth is within the range of 100-200 meters,/>Taking 2 meters; when the sea water depth is less than 100,/>Taking 3 meters.
Total cost of submarine cableCan be expressed as:
using the elevation map in the target area D and equations (1) - (3), the overall cost profile of the sea cable in the target area D can be calculated.
2) Risk factor modeling
As the gradient of the submarine topography can affect the safety of the submarine cable, the greater the submarine topography relief, the higher the risk of causing the submarine cable to break. Influence of submarine topography gradient on submarine cable risk classification can be expressed as the following formula:
In the formula (4) of the present invention, Representing the risk factor of the gradient of the submarine topography to the submarine cable. And (4) calculating the distribution of the gradient of the submarine topography to the submarine risk factors by using the gradient of the submarine cable and the formula (4).
In this step, modeling is performed on various risk factors affecting the submarine cable, so as to obtain a target to be optimized of the submarine cable, for example, the target distribution of the cost to be optimized is obtained according to the modeling of the water depth elevation map. The invention uses cost and risk as two optimization targets, wherein the cost targets consider two factors of raw material cost and deployment cost, and the risk uses terrain gradient as an influencing factor. In practical applications, multiple optimization objectives may be set as needed, and each optimization objective may take into account multiple factors. Modeling of each objective may be optimized according to actual process needs.
2. Land cable network cost and risk modeling
1) Cost modeling
Land cables do not involve cables using different protective armor, so the raw material cost difference is small, and the land cables are set in the target area during modeling. The deployment modes of the land optical cable are pipeline laying, excavation laying and overhead, and the corresponding deployment modes are respectively/>、/>And/>. In addition, land cables involve land right problems, and when the cable construction causes loss of land owners, compensation needs to be given, and deployment modes may be caused, for example, when the cable passes through a farmland area, the cable cannot be laid in a buried manner due to frequent ploughing of the farmland, but is in an overhead manner, and loss of crops needs to be compensated. Assume the cost of compensation is/>. Depending on the land characteristics at the geographic location of the GIS, the deployment cost of the fiber optic cable can be expressed as:
the total cost of a land cable can be expressed as:
By means of equations (5) - (6), the total cost of the land cable in the target area, the cost target profile to be optimized, i.e. the profile of the cost optimization target, can be calculated.
(2) Risk factor modeling
The land optical cable is commonly subjected to natural disasters such as landslide, flood and the like in the mountain, and the vehicle breaks, scrapes, pulls and the like the optical cable. The greater the altitude or grade, the higher the risk of the mountain landslide. The risk that a mountain landslide would create for a fiber optic cable can be expressed as:
in the formula (7), the amino acid sequence of the compound, Is the altitude of the geographic location,/>Represented as a slope.
Assume that the geographic location is a distance from the river. According to engineering experience, the optical cable is deployed more than 5 km from the river, and the optical cable is little affected by flood, so the risk of the river to the cable can be expressed as:
the impact of a vehicle on the cable can be expressed as:
Combining several risk factors, the risk factors for a land cable can be expressed as:
in the formula (10), the amino acid sequence of the compound, Representing normalization. In practical application, various risk factors can be expanded, assuming that/>Seed risk factor, wherein/>Represents the/>The distribution of risk factors in the region, the final risk factor distribution can be expressed as:
Step three, existing submarine cable data acquisition and pretreatment
In this embodiment, a grid method is used to model the geographic information system, and the calculated cost and risk distribution are also in the form of a grid. The invention adopts a mode of moving a nine-grid window to collect data, as shown in figure 4. Each small cell in the nine-grid represents a geographic location. At the time of data acquisition, the 5 th small lattice indicates the current position, and the remaining 8 small lattices are adjacent positions thereof. In order to more specifically describe the parameter acquisition process, the range selected in the step is D #，/>) By way of example, and assuming that the area already has a fiber optic cable, this step will be exemplified by collecting risk data for the existing fiber optic cable route. Modeling the risk of the region by using a formula (4) to obtain the distribution of risk values. The risk profile and existing submarine cable routing is shown in fig. 5. And acquiring risk data of the existing submarine cable at each point along the path of the existing submarine cable, wherein the No. 5 of the nine-grid in the acquisition process is always on the path of the existing optical cable. Let the position in number 5 be/>Then the adjacent risk value is collected as/>. The cost distribution of region D can be calculated according to formulas (1) - (3), and the cost data at that location is collected in the same way as risk data collection, labeled/>。
At the same time, three parameters are required to be collected, and the included angle formed by the connecting line of the current point and the end point and the connecting line of the current point and the next planning point is shown as figure 5. At geographical location/>The included angle of the points is expressed as/>. Assume that the plane position coordinates of the current point are (/ >)) The plane position coordinates of the end point are (/ >) The straight line distance between the two points is. These two parameters are the same for the risk profile and the cost profile. It is to be noted that,And/>Are two important super parameters, such as when/>Is about to reach the emphasis, in general/>The variation range of the algorithm is reduced, and the algorithm is prevented from bypassing blindly. In addition, there is a next point of the actual cable routing, i.e. Thus, the acquired dataset may be expressed as:
Input: 。
and (3) outputting: 。
the data preprocessing involves normalization.
Step four, training an experience learning model
Since the path of the existing optical cable is expert complex、/>、/>、/>And determining the next planning point segment by segment under the condition, the different factors are not simple and fixed, and are a complex and nonlinear process. The present invention employs neural networks for learning such complex, non-linearities, which have proven to approach any complex process indefinitely by increasing the number of hidden layers, increasing the number of neurons, and selecting appropriate activation functions. Thus, the present invention employs a neural network learning expert decision process, which uses a neural network of hidden layers as an example, but not limiting the number of hidden layers. As shown in fig. 6. Taking the data acquired and normalized in the step three as/>、/>、/>、/>Input of an empirical learning model, wherein the model empirical learning model selects one point from 8 candidate points as a necessary point of a next path and marks the selected point as/>. Will/>/>, With the actual fiber optic cableIn contrast, using a gradient descent-containing approach, the empirical learning model is continually trained so that it can learn the strategy of the existing cable decision, which is noted as/>. Through training, the experience learning model can predict the actual/>, according to the related information of the geographic positionThe anastomotic points act as the necessary points of the path.
Step five, reading new optical cable and related data thereof and modeling numerical values
And reading the new optical cable and related data thereof according to the first step, and carrying out numerical modeling on the new optical cable area according to the second step to obtain a cost distribution diagram and a risk distribution diagram of the new optical cable target area.
Step six: new optical cable path planning
And (3) planning a new optical cable route by using an optical cable network route intelligent planning model, wherein the internal structure of the optical cable network route intelligent planning model is shown in figure 7. Step five, calculating to obtain the geographical positionCost and risk of (a) are expressed as/>, respectivelyAnd/>The learning environment of Agent is at geographic location/>The value of (2) may be calculated according to the following formula:
In the method, in the process of the invention, Representation takes a large value,/>And/>Respectively the set of costs and risks for all geographic locations within the target area D. In planning, agents constantly interact with the learning environment to maximize their expected rewards. The rule for the Agent to get rewards is set as follows: when the Agent successfully searches the end point, rewarding 5; agents search out boundary rewards-1 (i.e., penalty 1); the remaining cases are awarded 0.
And step five, calculating to obtain cost and risk distribution at the geographic position to be used for data acquisition, preprocessing acquired data, and inputting the preprocessed acquired data into an empirical decision module, wherein the acquisition process is shown in step three. The experience decision module has the same structure as the experience learning model, and uses model parameters of the experience learning model to collect positionsIs input into the empirical decision module to output the next action, i.e./>. The parameters used by the empirical decision-making module in the intelligent planning model for the optical cable network routing are migrated from the empirical learning model. The decision strategy of the empirical decision module is thus the same as the manual expert in planning the existing cable routing usage planning strategy.
In order to restrict the Agent to utilize expert strategy and exploration process in the planning process, the invention respectively uses by setting a random parameter and a decision thresholdAnd/>Representation of/>. In the planning process, by comparing/>And/>The magnitude relation of (2) determines whether the agent takes an action selected by the empirical decision module or randomly selects a candidate action as the next action. Such as when/>Setting to 0.95, the Agent will select the action decided by the experience decision module with a 95% probability as the next action, namely the necessary point of the target optical cable path; and one action is selected from the 8 candidate actions with a probability of 5% as the next action, i.e. the must-pass point of the target cable path. In other words, the Agent explores with 95% probability using the experience of a human expert, and with 5% probability. It should be noted that the present invention extracts the expert experience of the existing optical cable, and introduces the expert experience into the traditional reinforcement learning framework through the migration learning mode, as the main mode of decision making, which is completely different from the strategy adopted by the traditional reinforcement learning. The Agent selects final action according to the action, acts the action on the environment, changes the environment, feeds back a rewarding signal or punishment signal to the Agent, and finally designs an optical cable network route through repeated action and learning with the environment.
It should be noted that the parameters of the empirical learning model may be trained once and used repeatedly. Therefore, when the intelligent planning model of the optical cable network route performs new optical cable route planning, the new route does not need to be trained from the beginning every time when planning, the time for designing the optical cable network route can be saved, and the design efficiency of the optical cable network route is improved.
Step seven: feasible route dominance solution analysis
In step 5, by setting upIn the process, the Agent exploration capability is reserved, so that the planned feasible paths are not one, and a plurality of feasible solutions can exist. So after all feasible solutions are obtained, all feasible solutions are subjected to dominant relationship analysis. For example, assume that the total cost and total risk of two viable paths are denoted/>, respectively、/>、And/>If/>And/>If the cost and risk of the path 1 are better than those of the feasible path 2, the feasible path 2 is completely governed by the feasible path 1, and the best solution is the feasible path 1; if/>And/>That is, the target to be optimized of one feasible path is not better than that of the other feasible path, the feasible path 2 and the feasible path 1 are in incomplete dominance, and the feasible path 2 and the feasible path 1 are paths optimized by multiple targets
Step eight: outputting the multi-objective optimized optical cable network path.
Through step seven, cable network routes with optimized costs and risks can be selected from all possible path solutions.
Example 2
The embodiment 2 of the invention provides a multi-objective optimized optical cable network route intelligent planning system, which is realized based on the method of the embodiment 1, and comprises the following steps: comprising the following steps: the system comprises an acquisition module, a modeling module, a preprocessing module, an optimization output module, an empirical learning model and an intelligent planning model for optical cable network routing, wherein,
The acquisition module is used for acquiring geographic data and demand information of a new optical cable target area;
the modeling module is used for modeling according to a plurality of targets to be optimized of the geographic data of the new optical cable target area to obtain distribution diagrams of the targets to be optimized;
the preprocessing module is used for preprocessing the distribution graphs of the demand information and the multiple targets to be optimized;
The optimization output module is used for transferring a pre-established and trained experience learning model as an experience decision model to an intelligent planning model of the optical cable network route, inputting the preprocessed data into the intelligent planning model, adopting a reinforcement learning method, completing the optical cable route planning and optimization from a starting point to a terminal point through interaction and searching of an intelligent body and a reinforcement learning environment, and outputting a multi-objective optimized optical cable route;
The experience learning model is trained based on the geographic data of the existing optical cable routing area and the corresponding optical cable routing data and is used for obtaining an optical cable network path planning strategy of a manual expert.
Overview:
The invention provides an intelligent planning method and system for multi-objective optimized optical cable network routing, which are used for learning artificial expert experience from the existing optical cable path subjected to engineering inspection and are used for optical cable network routing planning. According to the invention, an experience learning model is trained by utilizing the optical cable path data which is manually designed by the prior expert experience, and the experience learning model can simulate the routing decision of an optical cable network by an artificial expert under various geographic environment conditions, so that the defects of distributing fixed weights to various optimization targets are avoided. The trained experience learning model is transferred to the route planning model, and the advantage that the experience learning model can be repeatedly utilized is utilized. Thus, the defects of the current optical cable network route planning method are overcome.
Finally, it should be noted that the above embodiments are only for illustrating the technical solution of the present invention and are not limiting. Although the present invention has been described in detail with reference to the embodiments, it should be understood by those skilled in the art that modifications and equivalents may be made thereto without departing from the spirit and scope of the present invention, which is intended to be covered by the appended claims.
Claims (9)
1. An intelligent planning method for multi-objective optimized optical cable network routing, the method comprising:
obtaining geographic data and demand information of a new optical cable target area;
modeling is carried out according to a plurality of targets to be optimized of the geographic data of the new optical cable target area, and a distribution diagram of the targets to be optimized is obtained;
preprocessing by combining the demand information and distribution diagrams of a plurality of targets to be optimized;
Transferring a pre-established and trained experience learning model as an experience decision model to an intelligent planning model of an optical cable network route, inputting the preprocessed data into the intelligent planning model, adopting a reinforcement learning method, completing optical cable route planning and optimization from a starting point to a terminal point through interaction and searching of an intelligent body and a reinforcement learning environment, and outputting a multi-target optimized optical cable route;
The experience learning model is trained based on the geographic data of the existing optical cable routing area and the corresponding optical cable routing data and is used for obtaining an optical cable network path planning strategy of an artificial expert;
The training step of the experience learning model comprises the following steps:
Reading geographic data and routing data of the existing optical cable routing area;
Modeling is carried out according to a plurality of targets to be optimized of the geographic data of the existing optical cable routing area, and distribution of the targets to be optimized is obtained;
preprocessing by combining the route data and distribution graphs of a plurality of targets to be optimized;
And inputting the preprocessed data, the linear distance between the current point and the terminal point and the included angle formed by connecting the current point with the terminal point and connecting the current point with the next planning point into an experience learning model, and training by using a gradient descent method until the training requirement is met, so as to obtain a trained experience learning model.
2. The intelligent planning method for a multi-objective optimized fiber optic cable network route of claim 1, wherein the new fiber optic cable objective region geographic data comprises: altitude, topography, and grade; the demand information includes a start point and an end point of a new cable route.
3. The intelligent planning method for the multi-objective optimization optical cable network route according to claim 1, wherein modeling is performed on a plurality of objects to be optimized according to geographic data of a new optical cable object area to obtain distribution diagrams of the plurality of objects to be optimized; the method specifically comprises the following steps:
when the new optical cable target area is the sea, modeling is carried out according to the cost, and the total cost of the sea cable is obtained through summation operation by the raw material cost and the sea cable deployment cost, wherein the raw material cost is different according to the interval where the sea water depth is positioned, the cost of the light sea cable is multiplied by the different coefficients, and the sea cable deployment cost is different according to the interval where the sea water depth is positioned, and the different deployment costs are corresponding;
According to the elevation diagram of the target area, the total cost of the submarine cable is combined, and a total cost distribution diagram of the submarine cable in the target area is calculated;
Modeling according to risk, and dividing risk factors of gradient of submarine topography on submarine cable :
Wherein/>A grade value representing the terrain;
combining risk factors of submarine cables according to terrain gradient of target area And calculating to obtain a submarine topography gradient versus submarine risk factor distribution map.
4. The intelligent planning method for the multi-objective optimization optical cable network route according to claim 1, wherein modeling is performed on a plurality of objects to be optimized according to geographic data of a new optical cable object area to obtain distribution diagrams of the plurality of objects to be optimized; the method specifically comprises the following steps:
When the new optical cable target area is land, modeling is carried out according to the cost, and the total cost of the land optical cable is obtained through summation operation by the raw material cost and the deployment cost of the land optical cable; the deployment cost of the land optical cable is different according to the laying areas;
According to the elevation diagram of the target area, combining the total cost of the land optical cable, and calculating to obtain a total cost distribution diagram of the land optical cable of the target area;
Modeling based on risk, risk factors for land cable :/>Wherein/>Representation normalization,/>Representing the distribution of the kth risk factors in the target area, and P represents the category number of the risk factors;
Combining risk factors of land optical cables according to terrain gradient of target area And calculating to obtain the risk factor distribution affecting the land optical cable.
5. The method for intelligent planning of a multi-objective optimized fiber optic cable network route of claim 3 or 4, wherein said preprocessing comprises normalization processing.
6. The intelligent planning method for the multi-objective optimized optical cable network route according to claim 1, wherein the empirical learning model is a neural network, inputs a total cost distribution map and a risk factor distribution map of eight adjacent positions of a position i acquired by adopting a nine-grid-based mode, and outputs the total cost distribution map and the risk factor distribution map of the eight adjacent positions as a position of a next planning point, wherein the straight line distance between a current point and a destination point, and an included angle formed by connecting the current point with the destination point and connecting the current point with the next planning point.
7. The intelligent planning method for the multi-objective optimized cable network route according to claim 1, wherein the processing procedure of the intelligent planning model for the cable network route comprises the following steps:
For geographical locations Cost after pretreatment/>And risk/>The position/>, is obtained byEnvironmental value/>：In the/>Representing a large value;
In planning, the intelligent agent continuously interacts with the learning environment to maximize the expected rewards obtained by the intelligent agent; wherein, the rule that the agent obtains the reward sets up as: when the intelligent agent successfully searches the terminal, rewarding 5; the intelligent agent searches out boundary rewards-1 and rewards 0 in other cases;
And when more than one feasible path solutions exist, performing dominant relation analysis on all the feasible path solutions to obtain the multi-objective optimized optical cable route.
8. The intelligent planning method for a multi-objective optimized cable network route according to claim 7, wherein said performing a dominant relationship analysis on all feasible path solutions comprises:
when the total cost and total risk of one feasible path solution is less than the total cost and total risk of the other feasible path solutions, the feasible path solution is a multi-objective optimized cable route;
When neither the total cost nor the total risk of any one feasible path solution can be less than the total cost and total risk of the other feasible path solutions at the same time, then all feasible path solutions are multi-objective optimized cable routes.
9. A system based on the multi-objective optimized fiber optic cable network routing intelligent planning method of claim 1, comprising: the system comprises an acquisition module, a modeling module, a preprocessing module, an optimization output module, an empirical learning model and an intelligent planning model for optical cable network routing, wherein,
The acquisition module is used for acquiring geographic data and demand information of a new optical cable target area;
the modeling module is used for modeling according to a plurality of targets to be optimized of the geographic data of the new optical cable target area to obtain distribution diagrams of the targets to be optimized;
the preprocessing module is used for preprocessing the distribution graphs of the demand information and the multiple targets to be optimized;
The optimization output module is used for transferring a pre-established and trained experience learning model as an experience decision model to an intelligent planning model of the optical cable network route, inputting the preprocessed data into the intelligent planning model, adopting a reinforcement learning method, completing the optical cable route planning and optimization from a starting point to a terminal point through interaction and searching of an intelligent body and a reinforcement learning environment, and outputting a multi-objective optimized optical cable route;
The experience learning model is trained based on the geographic data of the existing optical cable routing area and the corresponding optical cable routing data and is used for obtaining an optical cable network path planning strategy of a manual expert.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410067768.4A CN117592239B (en) | 2024-01-17 | 2024-01-17 | Multi-objective optimization optical cable network route intelligent planning method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410067768.4A CN117592239B (en) | 2024-01-17 | 2024-01-17 | Multi-objective optimization optical cable network route intelligent planning method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117592239A CN117592239A (en) | 2024-02-23 |
| CN117592239B true CN117592239B (en) | 2024-04-26 |
Family
ID=89918648
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410067768.4A Active CN117592239B (en) | 2024-01-17 | 2024-01-17 | Multi-objective optimization optical cable network route intelligent planning method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117592239B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112861442A (en) * | 2021-03-10 | 2021-05-28 | 中国人民解放军国防科技大学 | Multi-machine collaborative air combat planning method and system based on deep reinforcement learning |
| CN113139529A (en) * | 2021-06-21 | 2021-07-20 | 北京科技大学 | Linear cultural heritage exploration method and system, storage medium and electronic equipment |
| CN114254837A (en) * | 2021-12-28 | 2022-03-29 | 西安交通大学 | Travel route customizing method and system based on deep reinforcement learning |
| CN115550233A (en) * | 2021-06-30 | 2022-12-30 | 中兴通讯股份有限公司 | Distributed route determination method, electronic device, and storage medium |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11358004B2 (en) * | 2019-05-10 | 2022-06-14 | Duke University | Systems and methods for radiation treatment planning based on a model of planning strategies knowledge including treatment planning states and actions |
-
2024
- 2024-01-17 CN CN202410067768.4A patent/CN117592239B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112861442A (en) * | 2021-03-10 | 2021-05-28 | 中国人民解放军国防科技大学 | Multi-machine collaborative air combat planning method and system based on deep reinforcement learning |
| CN113139529A (en) * | 2021-06-21 | 2021-07-20 | 北京科技大学 | Linear cultural heritage exploration method and system, storage medium and electronic equipment |
| CN115550233A (en) * | 2021-06-30 | 2022-12-30 | 中兴通讯股份有限公司 | Distributed route determination method, electronic device, and storage medium |
| CN114254837A (en) * | 2021-12-28 | 2022-03-29 | 西安交通大学 | Travel route customizing method and system based on deep reinforcement learning |
Non-Patent Citations (3)
| Title |
|---|
| Multi-Objective Optimization for Submarine Cable Route Planning Based on the Ant Colony Optimization Algorithm;Zanshan Zhao 等;《Photonics》;20230802;第10卷(第8期);全文 * |
| 区域生态网络精细化空间模拟及廊道优化研究――以汾河流域为例;李欣鹏;李锦生;侯伟;;地理与地理信息科学;20200915(第05期);全文 * |
| 自动化技术、计算机技术;中国无线电电子学文摘;20100425(第02期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117592239A (en) | 2024-02-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Gigović et al. | Application of the GIS-DANP-MABAC multi-criteria model for selecting the location of wind farms: A case study of Vojvodina, Serbia | |
| CN109359350B (en) | Wind power plant road intelligent design method for optimizing fine construction cost | |
| CN111598473B (en) | Multi-platform combined task planning method for complex observation task | |
| CN109492756B (en) | Multi-element wire galloping early warning method based on deep learning and related device | |
| CN103218660B (en) | A kind of airway selection method based on extensive fuzzy competition nerve net | |
| CN113554849A (en) | Air-ground monitoring slope system and instability risk assessment method thereof | |
| CN113505510B (en) | Ecological safety pattern recognition method fusing landscape index and random walk model | |
| CN106772695A (en) | A kind of Along Railway wind speed forecasting method for merging many air measuring station measured datas | |
| CN112560215B (en) | Electric power line selection method based on deep reinforcement learning | |
| CN109992923A (en) | A kind of transmission line of electricity paths planning method stage by stage based on variable resolution cost surface | |
| CN115759883B (en) | Ecological management partitioning method based on network group characteristics | |
| CN112182951B (en) | Land utilization planning method considering multiple uncertainties | |
| CN114418215A (en) | Smart city power transmission line planning method based on artificial intelligence | |
| CN110929952B (en) | Optical cable fault probability prediction method based on circuit surrounding environment and laying type | |
| CN110378527A (en) | A kind of water station site selecting method | |
| Li et al. | Multi-mechanism swarm optimization for multi-UAV task assignment and path planning in transmission line inspection under multi-wind field | |
| CN117592239B (en) | Multi-objective optimization optical cable network route intelligent planning method and system | |
| CN105069027A (en) | Method used for calculating actual tower material strength of power transmission pole tower and capable of simplifying indexes | |
| CN117688844A (en) | Urban waterlogging real-time simulation method and system based on deep neural network | |
| CN113447762A (en) | Offshore wind farm submarine cable fault position positioning system and working method thereof | |
| CN117170408A (en) | Photovoltaic panel site inspection path intelligent planning system and method based on unmanned aerial vehicle | |
| CN115686070B (en) | Autonomous line patrol path planning method for unmanned aerial vehicle of power distribution/transmission line | |
| JPH0949884A (en) | Predicting method for local weather using neural network | |
| CN112183805B (en) | Prediction method for online inspection result state of power transmission line | |
| CN117579533A (en) | Optical cable network route intelligent planning method capable of being used for multi-objective optimization |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |