Dolphin Pod Optimization

Abstract

A novel nature-inspired, deterministic, global, and derivative-free optimization method, namely the dolphin pod optimization (DPO), is presented for solving simulation-based design optimization problems with costly objective functions. DPO is formulated for unconstrained single-objective minimization and based on a simplified social model of a dolphin pod in search for food. A parametric analysis is conducted to identify the most promising DPO setup, using 100 analytical benchmark functions and three performance criteria, varying pod size and initialization, coefficient set, and box-constraint method, resulting in more than 140,000 optimization runs. The most promising setup is compared with deterministic particle swarm optimization, central force optimization, and DIviding RECTangles and finally applied to the optimization of a destroyer hull form for reduced resistance and improved seakeeping.