Distributed Evolutionary Algorithm for Path Planning in Navigation Situation

• Main Authors: Roman Smierzchalski, Lukasz Kuczkowski, Piotr Kolendo, Bartosz Jaworski
• Format: Article
• Language: English
• Published: Gdynia Maritime University 2013-06-01
• Series: TransNav: International Journal on Marine Navigation and Safety of Sea Transportation
• Subjects: Evolutionary Algorithm; Path Planning; Navigation Situation; Multi-Population Algorithm; Single-Population Algorithm; Collision Avoidance; Simulation Environment; Multi-Population Distributed Evolutionary Algorithm
• Online Access: http://www.transnav.eu/files/Distributed Evolutionary Algorithm for Path Planning in Navigation Situation,438.pdf

This article presents the use of a multi-population distributed evolutionary algorithm for path planning in navigation situation. The algorithm used is with partially exchanged population and migration between independently evolving populations. In this paper a comparison between a multi-population and a classic single-population algorithm takes place. The impact on the ultimate solution has been researched. It was shown that using several independent populations leads to an improvement of the ultimate solution compared to a single population approach. The concept was checked against a problem of maritime collision avoidance.