Machine learning algorithms in ship design optimization

ABSTRACT

Numerical optimization of complex systems benefits from the technological development of computing platforms in the last 20 years. Unfortunately, this is still not enough, and a large computational time is necessary for the solution of optimization problems when mathematical models that implement rich (and therefore realistic) physical models are adopted. In this paper, we show how the combination of optimization and Artificial Intelligence (AI), in particular Machine Learning algorithms, can help, strongly reducing the overall computational times, making also possible the use of complex simulation systems within the optimization cycle. Original approaches are proposed.

KEYWORDS:

• Design optimization
• global optimization
• machine learning

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 $f\left(x\right)=2+0.01\ast (x_2-x_1^2{)}^2+(1-x_1{)}^2+2\ast (2-x_2{)}^2+7\ast \sin (0.5\ast x_1)\ast \sin (0.7\ast x_1\ast x_2)x_1\in [0,5],x_2\in [0,5]$

2 $f\left(x\right)=\sum _{i=1}^{12}(x_i-0.5{)}^2{x}_i\in [-10:10]$

Additional information

Funding

This research was funded by Italian Minister of Instruction, University and Research (MIUR) to support this research with funds coming from PRIN Project 2017 (No. 2017KKJP4X entitled ‘Innovative numerical methods for evolutionary partial differential equations and applications’).