1. UKM Learning and Research Repository
2. PHD Thesis / Tesis PHD
3. Faculty of Information Science and Technology / Fakulti Teknologi dan Sains Maklumat

Title:	Modified harmony search algorithms for the aircraft landing problem
Authors:	Omar Salim Abdullah (P74197)
Supervisor:	Salwani Abdullah, Prof. Dr.
Keywords:	Universiti Kebangsaan Malaysia -- Dissertations Dissertations, Academic -- Malaysia Algorithms Landing craft Air traffic controllers
Issue Date:	Sep-2019
Description:	The Aircraft Landing Problem (ALP) is a challenging task for air traffic controllers in an airport. ALP is a non-deterministic polynomial-time hard (NP-hard) problem that deals with assigning an available runway and a landing time to an arriving aircraft. The landing time of each aircraft must be within a stipulated target landing time. If the actual landing time deviates from the target landing time, an additional cost will be imposed. This cost is determined by the amount of earliness or lateness with respect to the actual landing time. ALP can be divided into static and dynamic problems. The static ALP (s-ALP) occurs when all the information on aircraft are fixed and there is no change in the information when the scheduling process commences. On the other hand, the dynamic ALP (d-ALP) considers changes in the information that occur during the scheduling process as new aircraft may appear in the radar range. Solving both static and dynamic ALPs aim to minimize the overall cost i.e., the deviation from a preferred target time of each aircraft. The complexity of the ALP draws the attention of researchers from various research domains to generate a robust system that supports air traffic controllers in making the landing decision. Many heuristic and metaheuristic approaches have been developed to derive a highly effective solution for this complicated problem. The research work presented in this thesis aims to build upon the state-of-the-art search methodologies for aircraft scheduling problems by investigating the use of the Harmony Search Algorithm (HSA) as a population-based algorithm for s-ALP and d-ALP. The research first investigates HSA and incorporates a number of rules to control the neighborhood structures employed during the optimization process. This contribution systematically avoids the slow convergence problem in the traditional HSA algorithm as a result of the random strategy deployed therein. Afterward, a hybridization between HSA and Variable Neighborhood Search (VNS) is proposed to improve the ability of the HSA in exploring an unvisited region in the search space. Moreover, Simulated Annealing (SA) is used to avoid the local optimum. The newly proposed algorithm is coded as VNHSA. The VNS replaces the pitch consideration rate condition in the improvisation step in the original HSA which ensures the algorithm is not stuck in the local optimum. Due to the fact that unpredictable changes might occur during the course of an ongoing scheduling process, the applicability of the VNHSA is tested on d-ALP. The proposed approaches are tested using well-known datasets from OR-library with a range of 10 to 500 aircraft and 1 to 5 runways. Computational results based on standard benchmark datasets demonstrate the effectiveness of the proposed algorithm. Further evaluations are made through comparisons with the best results from other approaches in the scientific literature and statistical tests. The results show that the proposed approaches are able to obtain competitive results and can be deployed in practice.,Ph.D.
Pages:	197
Call Number:	QA76.9.A43A233 2019 3 tesis
Publisher:	UKM, Bangi
Appears in Collections:	Faculty of Information Science and Technology / Fakulti Teknologi dan Sains Maklumat

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.