Autonomous Ships: A Velocity Obstacle Particle Swarm Optimization Hybrid Algorithm

Abstract

Hovedmålet med denne oppgaven var å designe en kollisjons-unngående hybrid-algoritme basert på Velocity Obstacle og Multi Objective Particle Swarm Optimization. Formålet var å forbedre ytelsen i områder som anses om svakheter i de originale algoritmene. Både original- og hybrid-algoritmene ble kjørt gjennom en rekke tester, som viste at hybridalgoritmen viste bedre ytelse i visse områder i forhold til algoritmene den var basert på.

The main goal of this thesis was to design a hybrid collision avoidance algorithm from components of the Velocity Obstacle and Multi Objective Particle Swarm Optimization algorithms. The purpose was to improve the performance in areas that were weaknesses in the original algorithms. The original and hybrid algorithms were run to a selection of tests, which found that the hybrid algorithm improved performance in certain areas compared to the algorithms it was based on.