Multi-objective robust early stage ship design optimisation under uncertainty utilising surrogate models

Tools

Priftis, Alexandros and Boulougouris, Evangelos and Turan, Osman and Atzampos, Georgios (2020) Multi-objective robust early stage ship design optimisation under uncertainty utilising surrogate models. Ocean Engineering, 197. 106850. ISSN 0029-8018 (https://doi.org/10.1016/j.oceaneng.2019.106850)

[thumbnail of Priftis-etal-OE-2020-Multi-objective-robust-early-stage-ship-design-optimisation-under-uncertainty-utilising-surrogated-models]
Preview
 Text. Filename: Priftis_etal_OE_2020_Multi_objective_robust_early_stage_ship_design_optimisation_under_uncertainty_utilising_surrogated_models.pdf
Accepted Author Manuscript
License: Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 logo
Download (1MB)| Preview

Abstract

The paper addresses the parametric design and multi-objective optimisation of ships under uncertainty, applying the Holistic Optimisation Design Approach. The methodology starts with the creation of a detailed parametric model that captures both the external and internal geometric characteristics of the vessel, along with the integration of a number of numerical tools, purposely developed in order to determine each variant’s performance. This allows the evaluation of a multitude of merit functions and design constraints, all part of the optimisation problem. In contrast to previous approaches, the uncertainties introduced by the physical and financial environment are incorporated in the model, providing a more realistic representation of the solution space to the decision maker. The developed methodology is applied on a case study of a Ro-Pax ship, optimised in the concept design phase with regard to the minimisation of the total resistance, required freight rate and building cost.

ORCID iDs

Priftis, Alexandros ORCID logoORCID: https://orcid.org/0000-0002-2773-5846, Boulougouris, Evangelos ORCID logoORCID: https://orcid.org/0000-0001-5730-007X, Turan, Osman and Atzampos, Georgios;
CORE (COnnecting REpositories)