Development of a negative selection algorithm for ship collision risk detection

Γιακουμάκης, Κωνσταντίνος; Giakoumakis, Konstantinos

dc.contributor.author	Γιακουμάκης, Κωνσταντίνος	el
dc.contributor.author	Giakoumakis, Konstantinos	en
dc.date.accessioned	2023-08-31T10:34:46Z
dc.date.available	2023-08-31T10:34:46Z
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/57940
dc.identifier.uri	http://dx.doi.org/10.26240/heal.ntua.25637
dc.rights	Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ελλάδα	*
dc.rights.uri	http://creativecommons.org/licenses/by-nc-nd/3.0/gr/	*
dc.subject	Negative Selection Algorithm	en
dc.subject	Ship Collision Risk Detection	en
dc.subject	Algorithm	en
dc.subject	Risk	en
dc.title	Development of a negative selection algorithm for ship collision risk detection	en
heal.type	bachelorThesis
heal.classification	Computer Programming	en
heal.language	en
heal.access	free
heal.recordProvider	ntua	el
heal.publicationDate	2023-06-27
heal.abstract	Maritime transportation safety is threatened by the occurrence of ship collisions which may result in catastrophic consequences. Researchers have developed collision prevention methods that can be broadly classified into collision risk detection and collision avoidance. Collision risk detection aims to identify potential collisions prior to their occurrence, whereas collision avoidance aims to determine evasive actions after a potential collision has been detected. This thesis focuses on the development of a novel collision risk detection method utilizing a negative selection algorithm. In order to achieve our goal of developing a novel collision risk detection method, we drew upon knowledge from both the ship collision risk detection and artificial immune system sectors. The initial stages of our thesis focused on reviewing and synthesizing existing research in these fields, identifying key insights that inspired our approach. From there, we developed a secondary algorithm to construct the dataset needed for training and evaluating the negative selection algorithm. This simulation algorithm incorporated an expert-based method of ship domain as a collision criterion. With the dataset in place, we then proceeded to train our detectors and develop the method used by the algorithm to identify potential threats once a detector set has been acquired. To evaluate the proposed negative selection algorithm for collision risk detection, we conducted a thorough sensitivity analysis of its key parameters. This analysis served a dual purpose: firstly, it allowed us to gain a deeper understanding of how these parameters affect both the accuracy and the execution time of the algorithm. Secondly, through this analysis we were able to optimize these parameters, ultimately enabling us to select the optimal values for each parameter and improve the overall performance of the algorithm. After finalizing the development of our algorithm, we evaluated its performance in terms of accuracy and execution time through a series of experiments. Additionally, we conducted several case studies to demonstrate the practical application of our algorithm. The outcomes of our proposed algorithm exhibit promising results. The low execution time, even with limited computational resources and its high accuracy demonstrate the potential of negative selection algorithms in addressing the collision risk detection problem. Although there is room for optimization and further advancement, we believe that our algorithm can serve as a foundation for future studies in this emerging field.	en
heal.advisorName	Βεντίκος, Νικόλαος	el
heal.committeeMemberName	Ηλιοπούλου, Ελευθερία	el
heal.committeeMemberName	Θεμελής, Νικόλαος	el
heal.academicPublisher	Εθνικό Μετσόβιο Πολυτεχνείο. Σχολή Ναυπηγών Μηχανολόγων Μηχανικών. Τομέας Μελέτης Πλοίου και Θαλάσσιων Μεταφορών	el
heal.academicPublisherID	ntua
heal.numberOfPages	75 σ.	el
heal.fullTextAvailability	false