Optimization design of hatch corners in a container ship based on the neural network and multi-objective particle swarm optimization

摘要

Multi-objective optimization method is used to optimize the design ofthe ship’s hatch corner. The sample space of design variables isobtained according to the Latin hypercube sampling design method.With the analysis results, BP neural network is applied to construct theproxy models of the objective functions, and then the proxy models areoptimized with multi-objective particle swarm optimization (MOPSO)and then the improved minimum distance method is used to quicklyselect an optimal solution from the set of non-inferior solutions. Theresult shows that the elliptical hatch corner is the optimal shape for thisposition.