Development Of Artificial Bee Colony (Abc) Variants And Memetic Optimization Algorithms

Sulaiman, Noorazliza (2017) Development Of Artificial Bee Colony (Abc) Variants And Memetic Optimization Algorithms. PhD thesis, Universiti Sains Malaysia.

Download (229kB) | Preview


Bio-inspired optimization algorithms (BIAs) have shown promising results in various diverse realms. One of BIAs, artificial bee colony (ABC) optimization algorithm, has shown excellent performance in many applications compared to other optimization algorithms. However, its performance sometimes deteriorates as the complexity of optimization problems increases. ABC normally has slow convergence rates on unimodal functions and yields premature convergence on complex multimodal functions. Researchers have proposed various ABC variants in order to overcome these problems. Nevertheless, the variants still fail to avoid both limitations simultaneously. Hence, this research work proposes six modified ABC variants and six memetic ABC algorithms with the aim of overcoming the problems of slow convergence rates and premature convergence. The modified ABC variants have been developed by inserting new processing stages into the standard ABC algorithm and modifying the employed-bees and onlooker-bees phases to balance out the exploration and exploitation capabilities of the algorithm. The proposed memetic ABC algorithms have been developed by hybridizing the proposed ABC variants with a local search technique, augmented evolutionary gradient search (EGS). The performances of all modified ABC variants and formulated memetic ABC algorithms have been evaluated on 27 benchmark functions. The best-performed modified ABC variants and memetic ABC algorithms are identified. To validate their robustness, the identified best-performed modified ABC variants and memetic ABC algorithms have been applied in three real-world applications; reactive power optimization (RPO), economic environmental dispatch (EED) and optimal digital IIR filter design. The obtained results have shown the superiority of the proposed optimization algorithms particularly JA-ABC5a, JA-ABC9 and EGSJAABC9 in comparison to the existing ABC variants and memetic ABC algorithm. For example, EGSJAABC9 has produced the most minimum power loss in comparison to other algorithms. Also, EGSJAABC9 has obtained the minimum EED value of 6.5593E+04 ($(lb)) for 6-generatior unit system while JA-ABC9 and EGSJAABC9 acquired the least EED value of 1.1656E+05 ($(lb)) for 10-generator unit system. Meanwhile, EGSJAABC9 has attained the best results at optimizing LP, BP and BS filters with 8.41E-03, 0.00E+00 and 5.70E-01 values of magnitude response error, respectively. As for optimizing HP filter, EGSJAABC9 is the second best. These results show that the proposed ABC variants and memetic ABC algorithms particularly EGSJAABC9 are robust optimization algorithms as they are able to converge faster and avoid premature convergence when dealing with complex optimization problems.

Item Type: Thesis (PhD)
Subjects: T Technology
T Technology > TK Electrical Engineering. Electronics. Nuclear Engineering > TK1-9971 Electrical engineering. Electronics. Nuclear engineering
Divisions: Kampus Kejuruteraan (Engineering Campus) > Pusat Pengajian Kejuruteraaan Elektrik & Elektronik (School of Electrical & Electronic Engineering) > Thesis
Depositing User: Mr Mohamed Yunus Mat Yusof
Date Deposited: 02 Jul 2020 07:11
Last Modified: 17 Nov 2021 03:42

Actions (login required)

View Item View Item