8. Utilization of Genetic Algorithm for Optimizing the Design of Ship Hull Structures

概要

論文の詳細を見る
The conventional design-spiral approach to basic ship design is replaced by a genetic algorithm in which convergence toward target value is sought of an object function to optimize globally a broad combination of key design variables including the principal hull dimensions, general arrangement and material steel grades, as well as the frame and stiffener spacings and scantlings. The algorithm is trially applied first to minimizing the weight of a hatch cover, and it is found through an exercise that the best version of the algorithm is that which lays weight on searching the immediate vicinity of designs that promise high performance. The refined version of algorithm is then applied to optimizing the design of a double hull tanker, with 40 independent variables adopted for global optimization. It is shown that with appropriate formulation of the object function and of the constraints, accurate design optimization is obtainable already at the initial planning stage of hull design.
社団法人日本船舶海洋工学会の論文