Application of Artificial Intelligence Methods to Preliminary Design of Ships and Ship Performance Optimization

作者：Abramowski Tomasz^*

来源：Naval Engineers Journal, 2013, 125(3): 101-112.

• artificial intelligence
• ship design

The article presents the methods used for the optimization of design parameters of cargo ships. A solution vector includes seven variables. A model for the determination of ship effective power is formulated using neural networks. Furthermore, the model is used in the optimization procedure where optima are sought using two algorithms: a genetic algorithm, and an algorithm of simulated annealing. Two examples of optimization are presented, the first of them being a single-objective problem in which the solution is sought for a minimum of propulsion power, and the second featuring a multi-objective formulation in which a minimum of propulsion power and a maximum of deadweight are sought simultaneously. The results are analyzed based on the time series of some selected values recorded during the optimization process, and also based on the set of Pareto-optimal solutions obtained in the multi-objective optimization.

• 出版日期2013-9