Use this URL to cite or link to this record in EThOS: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.742346
Title: Improved spiral dynamics and artificial bee colony algorithms with application to engineering problems
Author: Hashim, Mohd Ruzaini
ISNI:       0000 0004 7228 4401
Awarding Body: University of Sheffield
Current Institution: University of Sheffield
Date of Award: 2018
Availability of Full Text:
Access from EThOS:
Access from Institution:
Abstract:
The main focus of this research is to develop improved version of spiral dynamic algorithm (SDA) and srtificial bee colony algorithm (ABC) for solving various kinds of optimization problems. There are eight new algorithms developed in this research based on ABC and SDA. The first modification is on the initial distributions based on chaotic maps trajectory, random and opposition based learning. This adaptive initial distribution is used in all proposed algorithms. Second modification is to use spiral radius and chaotic rotational angle in modified SDA. Furher modifications include proposition of three adaptive ABC algorithms based on the modification of step size using exponential, linear and combination of both linear and exponential functions. Investigations have shown that SDA is a fast and simple algorithm but is subject to getting trapped in local optima and it lacks diversity in the search, while ABC is able to get accurate output at the expense of high computational time and slow convergence. Thus, the research further embarks on hybridisation of SDA and ABC, taking advantage of capabilities of both algorithms and three hybrid algorithms are thus proposed. The performances of the proposed algorithms are evaluated and assessed in single objective, multi-objective type and in practical optimization problems. Statistical and significant tests are used in the evaluations. Furthermore, comparative assessments of the performances of the proposed algorithms are carried out with their predecessor algorithms. The results show that the proposed algorithms outperform their predecessor algorithms with high accuracy and fast convergence speed.
Supervisor: Tokhi, Mohammad Osman Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.742346  DOI: Not available
Share: