Authors:
Ruba Khurma
1
;
Moutaz Alazab
2
;
J. Merelo
3
and
Pedro Castillo
3
Affiliations:
1
Department of Computer Science, Al-Ahliyya Amman University, Amman, Jordan
;
2
Department of Artificial Intelligence, Al-Balqa University, Al-Salt, Jordan
;
3
Department of Computer Architecture and Computer Technology, ETSIIT and CITIC, University of Granada, Granada, Spain
Keyword(s):
Snake Optimizer, Evolutionary Operators, Selection Schemes.
Abstract:
Evolutionary algorithms (EA) adopt a Darwinian theory which is known as ”survival of the fittest”. Snake Optimizer (SO) is a recently developed swarm algorithm that inherits the selection principle in its structure. This is applied by selecting the fittest solutions and using them in deriving new solutions for the next iterations of the algorithm. However, this makes the algorithm biased towards the highly fitted solutions in the search space, which affects the diversity of the SO algorithm. This paper proposes new selection operators to be integrated with the SO algorithm and replaces the global best operator. Four SO variations are investigated by individually integrating four different selection operators: SO-roulettewheel, SO-tournament, SO-linearrank, and SO-exponentialrank. The performance of the proposed SO variations is evaluated. The experiments show that the selection operators have a great influence on the performance of the SO algorithm. Finally, a parameter analysis of t
he SO variations is investigated.
(More)