TY  - MANSCPT
AU  - Nurul Asyikin Binti Zainal
TI  - Enhancing the symbiotic organisms search algorithm with fuzzy adaptive selection
SN  - THE0010431 (Local)
PY  - 2025///
CY  - Kuantan, Pahang
PB  - UMPSA
KW  - Faculty of Computing
KW  - Dissertations
KW  - Universities and colleges
KW  - Theses
N1  - Faculty of Computing; Thesis (Doctor of Philosophy) -- Universiti Malaysia Pahang Al Sultan Abdullah - 2025; Include bibliographical reference
N2  - The Symbiotic Organism Search (SOS) algorithm has recently attracted significant attention due to its superior performance compared to other metaheuristic algorithms. A key advantage of SOS is its reliance on three distinct search phases (based on three update operators) and does not rely on any parameter controls. On the one hand, the three update operators work well (i.e., mutualism, commensalism, and parasitism) among each other and relieve the need to do significant tuning of any control parameters. On the other hand, not having any control parameters can also be a potential drawback. In the current form, SOS has a pre-set sequence for exploration (searching through the entire solution space) and exploitation (utilizing the known local best) beginning from commensalism, mutualism, and parasitism update operators. Having a pre-set sequence can be counter-productive because the search process becomes insensitive to the current needs of the optimization, which could lead to premature convergence or getting trapped in local optima by not adapting dynamically to the problem's requirements. To address these issues, this thesis proposes an adaptive fuzzy-based SOS variant called Fuzzy Symbiotic Optimization Search (FSOS). Based on the use of Mamdani fuzzy, FSOS can select one or a combination of the mutualism, commensalism, and parasitism update operators by adjusting their individual probabilities through fuzzy decision-making. FSOS was tested on 23 benchmark functions and applied to a t-way test generation case study to assess its effectiveness. The experimental results show that FSOS delivers competitive performance, surpassing its predecessor (SOS) and other metaheuristic algorithms with a 95% confidence level. Overall, the findings demonstrate the potential of a parameter-free adaptive algorithm in solving general optimization problems, making them relevant for scientific research and practical applications
ER  -