Concept Design Of An Autonomous Underwater Vehicle Using A Multiple-objective Genetic Optimizer

The essence of Concept Design is innovation and revolution. The Concept Design is aprocess which is continuously innovated from logical to divergent thinking. Recently,Autonomous Underwater Vehicle (AUV)’s applicable values on the ocean exploitation havebegun to receive all countries’ attention, and begun to research the AUV. When theenvironment and task requirement is decided, the first thing should do is the Concept Designfor AUV.Traditionally, the AUV design process is largely governed by specific experience andrules of thumb. For it has multi-objective optimization problem in AUV Concept Design, thetraditional designs can’t thoroughly take the interconnection between disciplines into account,can’t make the optimal result achieve optimization of system, and follow with many defaults.With the development of AUV design, the concept design process is further innovative. InAUV concept design process, researchers begin to apply multi-objective intelligentoptimization algorithms in solving multi-objective problems of AUV concept design.This paper researches on multi-discipline and multi-objective optimization problem forthe stage of AUV overall concept design, use NSGA-Ⅱ, which belong to multi-objectiveevolution algorithms, to solve optimal problem in AUV concept design. In this paper, threeobjective attributes are considered: Effectiveness, cost and risk. The genetic algorithm ischosen to search the feasible and optimal solution of the three objective attributes in designspace.Firstly, this paper builds an AUV synthesis model. The AUV synthesis model constitutewith one input module, three main AUV synthesis modules, one constraint module, threeobjective modules and a NSGA-Ⅱ optimization algorithm, the AUV synthesis modulescontains energy module, resistance module and hull module, the objective modules containscost module, risk module and effectiveness module. The effectiveness of AUV is defined onbasis of the nine tasks of US navy and four associated attributes determined by experts. Therisk of AUV constitute with effectiveness, cost and progress. The decision for the weights ofthirteen effectiveness and three risks is a multi attributes decision problem. To solve thisproblem, this paper use Analytic Hierarchy Process.Secondly, this paper has researched on the multi objective optimization of the AUVConcept Design stage by applying the idea of multi discipline optimization, and hasdeveloped an optimized mathematical model, optimized the cost, risk and effectiveness of AUV by using NSGA-Ⅱ. Once the NSGA-Ⅱ generates the last feasible and optimal solution,the decision maker can choose the candidate design to do more detail design, following theirpreference. Moreover, this paper applies technique for order preference by similarity to idealsolution (TOPSIS) to choosing the best one from these Pareto solutions.Finally, four candidate AUVs, which includes lowest cost, lowest risk, largesteffectiveness and the one obtained by using TOPSIS, were analyzed. And selectivelyvalidation for the candidate solutions is accomplished. The result shows that the globaloptimized point of AUV synthesis performance can be effectively obtained by applying thisalgorithm.