Structural Optimization of Ships: Benchmark Study of Metaheuristic Algorithms and Constraint Handling Approaches

Yuecheng Cai,J. Jelovica

DOI: https://doi.org/10.1115/omae2022-79199

2022-06-05

Abstract:Optimization of ship structures can be performed using swarm and evolutionary algorithms in search for global optima, which is useful to maximize performance and minimize costs. There is a wide variety of such algorithms available and some have been proposed recently, but tested on mathematical problems. Thus, a benchmark study is performed here to assess their performance on structural optimization of a 180 m long chemical tanker which needs to fulfil class society’s requirements. Main frame is optimized considering firstly structural weight as a single objective and secondly weight and deck adequacy as two concurrent objectives. Optimization of deck adequacy leads to decrease of stresses in the deck. Artificial neural network is used as surrogate model to reduce the computational time. Following swarm and evolutionary algorithms are considered: PSO, NSGA-II, MOEA/D, MVO, MOMVO, MOGWO, IGWO and GSA. Following constraint handling techniques (CHTs) are used within: static and dynamic penalty, adaptive threshold and repair method. All algorithms are run for 30 times and the statistical results are presented. Results are compared in terms of objective values and speed of convergence. Results reveal that recently proposed swarm algorithms perform worse than well-known evolutionary algorithms in terms of the convergence rate and spread of the non-dominated front. In addition, algorithms’ performance is strongly influenced by CHT used. Among the ones tested, the best CHT is the repair method.

Engineering