CN-121997463-A - Ship model optimization method and device, electronic equipment and storage medium

CN121997463ACN 121997463 ACN121997463 ACN 121997463ACN-121997463-A

Abstract

The application provides a ship model optimization device, an electronic device and a storage medium, wherein the method comprises the following steps: by taking the trained BP neural network as the fitness function of the self-adaptive PSO and utilizing the strong nonlinear mapping capability of the BP neural network, the relationship between a plurality of complex factors such as hydrodynamic characteristics, navigation conditions and the like and an optimization target can be deeply learned and accurately modeled, and the optimization precision is greatly improved. The multi-objective problem is converted into single-objective optimization through a reasonable weight distribution or objective normalization method, or a pareto optimal solution set is searched by adopting a multi-objective particle swarm algorithm evolution strategy. The method can comprehensively consider the importance of each target in the searching process, avoid neglecting other key indexes due to single target leading, provide a more comprehensive and reasonable optimal parameter combination scheme for ship design, and enhance the multi-target processing capability.

Inventors

PAN YAN
LI LANQING
LI RONGYANG

Assignees

国家工业信息安全发展研究中心

Dates

Publication Date: 20260508
Application Date: 20260302

Claims (11)

1. A method of optimizing a ship model, the method comprising: Constructing and training a back propagation neural network model based on ship historical data, wherein the back propagation neural network model is used for establishing a nonlinear mapping relation between ship design parameters and at least one performance index; Embedding the trained back propagation neural network model into a fitness function of a self-adaptive particle swarm optimization algorithm, wherein the position of each particle in the particle swarm represents a group of candidate ship design parameters; Initializing the particle swarm, and executing the following steps in each iteration until the set iteration termination condition is met: Inputting the current position of each particle into the trained back propagation neural network model to obtain a corresponding prediction performance index, and determining the current fitness of each particle based on the prediction performance index; according to the current fitness of each particle, updating the individual history optimal position of each particle and the group history optimal position of the particle group; Updating the position of each particle according to the updated individual history optimal position and the updated group history optimal position; and determining a final optimized target ship design parameter combination based on the optimization information recorded in the iterative process.
2. The method of claim 1, wherein determining fitness of each particle based on the predicted performance metrics comprises: assigning a weight value to each prediction performance index; and for each particle, carrying out weighted summation on the numerical values of a plurality of prediction performance indexes corresponding to the particle according to the weight value to obtain the adaptability of the particle.
3. The method of claim 1, wherein updating the individual historical optimal position for each particle and the population historical optimal position for the particle population based on the current fitness of each particle comprises: if the current fitness of the particle is better than the fitness corresponding to the individual historical optimal position of the particle, updating the individual historical optimal position of the particle by using the current position of the particle; And if the current fitness of the particles is better than the fitness corresponding to the group history optimal position of the particle swarm, updating the group history optimal position of the particle swarm by using the current position of the particles.
4. The method of claim 1, wherein updating the location of each particle based on the updated individual historical optimal location and the population historical optimal location comprises: And determining the speed variation of each particle according to the updated individual history optimal position and the updated group history optimal position, and updating the position of each particle according to the speed variation.
5. The method of claim 4, wherein determining the velocity variance of each particle based on the updated individual historical optimal location and the population historical optimal location is accomplished by the following equation: ; Wherein, the Characterization of the first embodiment The particles are at the first Dimension space of The speed at which the iteration is performed, As the weight of the inertia is given, 、 In order for the learning factor to be a function of, 、 In the form of a random number, Characterization of the first embodiment The particles are at the first The individual historic optimal positions of the dimensional space, Characterizing the particle swarm in the first A population history optimal position of the dimensional space; the updating of the position of each particle according to the speed variation is realized by the following formula: ; Wherein, the Characterizing the position of the ith particle in the d-th dimensional space in the t-th iteration, Characterization of the first embodiment The particles are at the first Dimension space of the first degree Speed at the time of iteration.
6. The method of claim 1, wherein determining a final target ship design parameter combination based on optimization information recorded in the iterative process comprises: and determining the candidate ship design parameter combination corresponding to the group history optimal position when the iteration termination condition is met as a final target ship design parameter combination.
7. The method of claim 1, wherein the optimization information is a pareto optimal solution set, and wherein determining the final target ship design parameter combination based on the optimization information recorded in the iterative process comprises: And determining a final target ship design parameter combination from the pareto optimal solution set.
8. The method of claim 6, wherein the determining a final target ship design parameter combination from the pareto optimal solution set comprises: calculating a comprehensive evaluation value for each scheme in the pareto optimal solution set, and selecting a scheme with the optimal comprehensive evaluation value as the final target ship design parameter combination; Or alternatively And receiving a user selection instruction, and selecting a corresponding scheme from the pareto optimal solution set according to the instruction as the final target ship design parameter combination.
9. A ship model optimization device, characterized in that the device comprises: the construction module is used for constructing and training a back propagation neural network model based on the ship historical data, and the back propagation neural network model is used for establishing a nonlinear mapping relation between ship design parameters and at least one performance index; The embedding module is used for embedding the trained back propagation neural network model into the fitness function of the adaptive particle swarm optimization algorithm, wherein the position of each particle in the particle swarm represents a group of candidate ship design parameters; The initialization module is used for initializing the particle swarm and executing the following steps in each generation of iteration until the set iteration termination condition is met: The first determining module is used for inputting the current position of each particle into the trained back propagation neural network model to obtain a corresponding prediction performance index, and determining the current fitness of each particle based on the prediction performance index; The first updating module is used for updating the individual history optimal position of each particle and the group history optimal position of the particle group according to the current fitness of each particle; The second updating module is used for updating the positions of the particles according to the updated individual historical optimal positions and the updated group historical optimal positions; And the second determining module is used for determining a final optimized target ship design parameter combination based on the optimization information recorded in the iterative process.
10. An electronic device comprising a processor and a memory, the processor being configured to execute a ship model optimization program stored in the memory to implement the ship model optimization method of any one of claims 1-8.
11. A storage medium storing one or more programs executable by one or more processors to implement the ship model optimization method of any one of claims 1-8.

Description

Ship model optimization method and device, electronic equipment and storage medium Technical Field The present application relates to the field of ship engineering technologies, and in particular, to a ship model optimization method, a device, an electronic apparatus, and a storage medium. Background The ship model optimization is a core problem in the field of ship engineering, and aims to achieve multi-objective collaborative optimization of propulsion efficiency, hull weight, manufacturing cost, structural strength and the like by adjusting design parameters. When the problem that the traditional BP neural network is easy to be in local optimum in ship parameter prediction is solved, the initial weight and the threshold value of the neural network are optimized through a particle swarm algorithm introducing an adaptive mutation operator, so that the prediction precision and the generalization capability are improved. However, the prior art solution has the limitations of single-target focusing, higher computational complexity, strong data dependence and the like, cannot relate to multi-target collaborative optimization, still depends on the traditional fitness function design, is difficult to accurately map the association of parameters and performance, and the computational complexity is obviously increased along with the increase of the number of targets. Disclosure of Invention The application provides a ship model optimization method, a ship model optimization device, electronic equipment and a storage medium, and aims to solve the problems that in the prior art, the solution scheme has the limitations of single-target focusing, high calculation complexity, high data dependence and the like. In a first aspect, the present application provides a ship model optimization method, including: Constructing and training a back propagation neural network model based on ship historical data, wherein the back propagation neural network model is used for establishing a nonlinear mapping relation between ship design parameters and at least one performance index; Embedding the trained back propagation neural network model into a fitness function of a self-adaptive particle swarm optimization algorithm, wherein the position of each particle in the particle swarm represents a group of candidate ship design parameters; Initializing the particle swarm, and executing the following steps in each iteration until the set iteration termination condition is met: Inputting the current position of each particle into the trained back propagation neural network model to obtain a corresponding prediction performance index, and determining the current fitness of each particle based on the prediction performance index; according to the current fitness of each particle, updating the individual history optimal position of each particle and the group history optimal position of the particle group; Updating the position of each particle according to the updated individual history optimal position and the updated group history optimal position; and determining a final optimized target ship design parameter combination based on the optimization information recorded in the iterative process. In one possible embodiment, the determining the fitness of each particle based on the prediction performance index includes: assigning a weight value to each prediction performance index; and for each particle, carrying out weighted summation on the numerical values of a plurality of prediction performance indexes corresponding to the particle according to the weight value to obtain the adaptability of the particle. In one possible embodiment, the updating the individual historical optimal position of each particle and the population historical optimal position of the particle group according to the current fitness of each particle includes: if the current fitness of the particle is better than the fitness corresponding to the individual historical optimal position of the particle, updating the individual historical optimal position of the particle by using the current position of the particle; And if the current fitness of the particles is better than the fitness corresponding to the group history optimal position of the particle swarm, updating the group history optimal position of the particle swarm by using the current position of the particles. In one possible embodiment, the updating the location of each particle according to the updated individual historical optimal location and the updated population historical optimal location includes: And determining the speed variation of each particle according to the updated individual history optimal position and the updated group history optimal position, and updating the position of each particle according to the speed variation. In one possible implementation manner, the determining the speed variation of each particle according to the updated individual historical optimal position and the updated p