Marine fleet allocation using data mining techniques

Abstract

Nowadays, classification is one of the many fields in Data Mining, also known as Knowledge Discovery in Databases, which aims at extracting information from large data volumes. In order to achieve this, data mining uses different computational techniques from machine learning, statistics and pattern recognition. In this work, a Data Mining techniques is used to help the Decision Maker of a Marin Transportation Firm called “SONOTRAK” to allocate a ship for a trip. The target is to ensure the transportation between “Sfax” city (Tunisia) and a closed island called “Kerkennah”. The fleet of “SONOTRAK” consists of five ships with different passenger and cars capabilities. The obtained classification gives groups of similar trips. Each class will be subject to arrange available ships according to the history of allocated trips in the last year. The most used ship will be the preferred one, and so on. Each ship will have a fitness value calculated according to this arrangement and to its fuel cost. The ship with the better fitness value will be allocated to the trip. The result ensures better management of the fleet of the company, and gives effect not only on the overall traffic but also on the fuel costs.