Hull-form optimization using parametric modification functions and particle swarm optimization

Abstract

The focus of this paper is on devising designer-friendly hull-form variations coupled with optimization algorithms. Hull-form variations are carried out through parametric modification functions. Two kinds of representative optimization algorithms are considered here. One is the well-known sequential quadratic programming which is the derivative based. The other is particle swarm optimization which is the derivative free. The results applying these two algorithms to typical hull-form optimization problems are discussed in the paper. The technique using the parametric modification functions has been developed for modifying the ship’s geometry according to the widely recognized naval architect’s design practice. An original geometry can be easily deformed through the change of the variables of the modification functions; and useful information about the effect of the parameters is immediately obtained. Moreover, the variables of the modification functions can be considered as the design variables in the formulation of the optimization problem. For the performance prediction of the hull form, WAVIS version 1.3 is used for the potential-flow and RANS solver. Computational results for both single- and multi-objective problems are presented.