Multi-objective optimal design of automatic underwater vehicle shell structure

A multi-objective optimal method based on adaptive response surface is proposed for the multi-objective optimization of deep-sea AUV (automatic underwater vehicle) shell structure. The original samples are obtained by the experimental design, the response surface methodology is used to create an approximate model, and the Pareto solution set is obtained after optimization. A spatial sampling method is introduced to get the equally distributed samples from Pareto solution set. After the samples are checked using the finite-element analysis method, they are added into the original sample set. The approximate model can satisfy the accuracy requirements by using an asymptotic global strategy. At last, a grey relational analysis method is used to deal with the Pareto solution sets. A multilevel-scales evaluation method is presented to get the weight of every objective based on the shell structure design of AUV, and then get the satisfactory result. The optimal result is much better than the result got by the empirical formula.