10.57647/j.mjee.2024.1802.30

A Two-Criteria Weather Routing Method Based on Neural Network and A-star Algorithm

  1. Faculty of Electrical – Electronics Engineering, Vietnam Maritime University, Hai Phong, Vietnam
Cover Image

Published 2024-05-07

How to Cite

Doan, K. H., & Dinh, A.-T. (2024). A Two-Criteria Weather Routing Method Based on Neural Network and A-star Algorithm. Majlesi Journal of Electrical Engineering, 18(2). https://doi.org/10.57647/j.mjee.2024.1802.30

HTML views: 0

PDF views: 2

Abstract

This paper presents the construction of a method to find the optimal route for ships with two criteria: fuel consumption and sailing time. Unlike most previous studies, the data used in this research was generated from a simulation model using the HIL (Hardward-In-The-Loop) technology instead of real operational data. The HIL simulator is built from equations of the ship's 6 degrees of freedom (6-DOF), models of environmental disturbances, propulsion systems, and technical records of the real ship. In fact, operating data of the real ship is collected from noon reports, which are often incomplete in terms of environmental disturbances acting on the ship, not to mention the large sampling time (usually updated once a day). Meanwhile, the dataset generated from the HIL simulator will fully include the three main environmental components acting on ships, including waves, wind, and currents, with various scenarios. Based on that dataset, an algorithm to find optimal routes with two criteria is proposed using neural networks and the A-star algorithm. Test results show that the proposed algorithm operates reliably and has low errors. This research can be applied to find the optimal routes for small and medium-sized ships in Vietnam before each voyage at a low cost instead of using high-cost weather routing services.

Keywords

  • A-star,
  • HIL,
  • Neural network,
  • Weather routing,
  • weather routing services.