抄録
There are more than 100 design variables and 2000 constraint conditions in designing mid-ship section of ship structure, and hence, it is very difficult to determine all design variables with satisfying all constraint conditions and having a reasonable value of objective function, or construction cost. Since this kind of optimization problem has so many design variables and a huge number of constraint conditions, it takes unbelievably large computational effort for obtaining an optimal solution when all design variables are solved simultaneously. A multi-step optimization method is introduced here with the aim of reducing the computational time. Since the design process of a mid-ship section is divided into three stages, double bottom, bilge hopper tank and topside tank stages, traditionally, three-step optimization may be rational and is introduced. But, when design variables in each step are determined separately, the solution obtained after the final step is far form the optimal solution obtained by solving all design the variables at once. In order to resolve this problem, characteristic design variable method is introduced and investigated here. The proposed method gives good solutions with satisfying all constraint conditions and having reasonable value of the objective function instead of solving all design variables at once. Genetic algorithm is used for optimization procedure in this study.
