Using a Bayesian network for selecting the optimal unit in maritime transport

A Bayesian network, a directed acyclic graph, is a computationally and spatially efficient method for representing a joint probability distribution. It is used in medical and technical diagnostics, information retrieval, fraud detection, control systems, etc. The article proposes the use of this tool in the field of maritime transport.The presented solution allows for the ranking of commercial vessels that can assist in maritime rescue operations. The priority was given to using basic ship parameters obtainable from automatic information exchange systems used at sea. A similarly constructed Bayesian network can be used to create a ranking of maritime transport units in order to increase the reliability of the transport chain.
INTRODUCTION
Bayesian methods are applied in a very wide scientific area. They are used in economic sciences to assess investment risk, in technical sciences to diagnose the reliability of a technical system, as well as in biological sciences and medicine, where they are increasingly used in technical diagnostics. [4] In particular, the Bayesian network, as a directed acyclic graph, is a convenient tool that can clearly present the problem being solved. It is a spatially and computationally efficient method. Additionally, access to tools that allow calculating conditional probability based on an entered network is very simple.
Free programs for preliminary research can be found on the Internet. This article uses results obtained with the help of the GeNIe 2.0 program.[3]In the literature, one can find many works using Bayesian trees for diagnostics in medicine or for assessing investment risk. In terms of use in the field of maritime technologies, proposals have appeared to use this tool for: optimal planning of inspections on offshore structures, monitoring ship engine operation, assessing the risk of helicopter landing on passenger ships, or for preliminary ship design assessment. [2] There are also models describing the causes of ship accidents, but so far I have not encountered the use of a Bayesian network to analyze the accuracy of selecting a maritime unit for a given task. (…)

The article contains 17,200 characters.

Source: Logistyka Magazine 4/2011

Using a Bayesian network for selecting the optimal unit in maritime transport – PDF file for download