基于遗传算法的船舶推力分配方法研究

Research on the method of ship thrust distribution based on genetic algorithm

近年来,海洋资源开发需求的日益增加对船舶推进器系统的性能提出越来越高的要求。当前,推力系统的主流配置方法为过驱动配置法,然而其存在求解效率低、全局范围内性能差等缺点。针对这一问题,本文提出了一种基于改进遗传算法的船舶推力动态分配方法,采用多步优化算法和径向基神经网络进行求解和训练。为提高全局可行域内的计算效率,本文对遗传算法中的交叉变异概率函数及适应度函数进行了优化。仿真结果表明,本文提出的推力分配方法功率消耗小且性能优异。

In recent years,the development of marine resources has put forward higher and higher requirements for the performance of the marine propulsion system.At present,the main configuration method of the thrust system is overdrive configuration,however,it has the disadvantages of low efficiency and poor performance in the global range.Aiming at solving this problem,this paper proposes a dynamic thrust allocation method based on improved genetic algorithm.It adopts multi-step optimization algorithm and radial basis function neural network to solve and train,and optimize the genetic algorithm to improve the computation efficiency in the global feasible area.The simulation results show that the thrust distribution method proposed in this paper has small power consumption and excellent performance simultaneously.