基于BP人工神经网络与遗传算法的航速优化

Speed optimization based on BP artificial neural network and genetic algorithm

为了进一步提高船舶能耗效率,本文提出一种基于BP人工神经网络与遗传算法的航速优化技术路线。首先,介绍常见油耗模型的构建方法;其次,利用BP人工神经网络建立目标船舶的油耗模型。模型预测的平均绝对误差为2.3%,准确度和泛化能力基本满足工程应用要求。最后,利用遗传算法,并基于历史气象数据对目标船舶的航线做分段航速优化。计算结果表明,航速优化后目标船舶的航行时长不仅能减少1.35天,燃油损耗还可节省10.1%,由此说明对航行船舶做分段航速优化是一种可行方案。

In order to further improve the energy efficiency of ships,a speed optimization technology route based on BP artificial neural network and genetic algorithm is proposed.First of all,some methods for constructing fuel consumption models are briefly introduced.Secondly,the BP artificial neural network was used to establish the fuel consumption model of the target ship,and the average absolute error of the model prediction was 2.3%,and the accuracy and generalization ability basically met the requirements of engineering applications.Finally,genetic algorithm is used to optimize the segmented speed of the target ship based on historical meteorological data.The calculation results show that the sailing time of the target ship can not only be reduced by 1.35 days,but also save fuel loss by 10.1%after speed optimization,which indicates that segmented speed optimization of sailing ships is a feasible solution.