Hull-form optimization of a luxury yacht under deterministic and stochastic operating conditions via global derivative-free algorithms

Abstract

Simulation-based design optimization (SBDO) techniques are used in the shape

design of complex engineering systems. SBDO methods integrate an optimization algorithm, a tool for the design modification, and analysis tools. In the context of ship/ocean applica- tions the objective function is often noisy, its derivatives are not directly provided and local minima cannot be excluded, therefore global derivative-free algorithms are widely used. The objective of this work is to investigate the efficiency of three global deterministic derivative-free optimization algorithms for the deterministic and stochastic hull-form optimization of a luxury yacht. Particle Swarm Optimization (DPSO), Dolphin Pod Optimization (DPO), and DIviding RECTangles (DIRECT) are applied to reduce the total resistance over a variety of conditions. The approach includes a comparison of the performances of the optimization algorithms, based on deterministic results for two separate operating conditions. DPSO is identified as the most promising optimization algorithm and is used for the robust design optimization (RDO) per- formed considering a stochastic variation of the cruise speed with uniform distribution within a speed range from 8 to 16 kn. The resistance curve of deterministic and robust solutions is finally presented.