Abstract—Artificial bee colony (ABC) is a relatively new
stochastic algorithm with competitive performance and minimal
tuning parameter. This paper proposes a hybrid ABC algorithm
with differential evolution (DE), but without additional
parameters. DE is a well-known efficient evolutionary algorithm
with proven records but its parameter setting is complicated.
This proposed hybrid algorithm called ABCDE incorporates the
powerful mutation strategies of DE into ABC, in order to
increase convergence while diversity is not compromised. The
performance of ABCDE is evaluated against both original ABC
and opposition-based DE (ODE), a recent DE variant with high
performance. The experiment uses twelve widely accepted
non-linear benchmark functions with various characteristics,
such as difficult landscape, multimodality, shift and rotation, to
evaluate the ABCDE’s performance on many complex functions.
The experimental results demonstrate a superior performance
of ABCDE against original ABC and ODE.
Index Terms—Artificial bee colony, hybridization,
differential evolution.
C. Worasucheep is with the King Mongkut University of Technology
Thonburi, Bangkok, Thailand (e-mail: chukiat.wor@kmutt.ac.th).
Cite: Chukiat Worasucheep, "A Hybrid Artificial Bee Colony with Differential Evolution," International Journal of Machine Learning and Computing vol. 5, no. 3, pp. 179-186, 2015.