Archimedean spiral based artificial bee colony algorithm

Sonal Sharma,Sandeep Kumar,Kavita Sharma
DOI: https://doi.org/10.1080/09720510.2019.1616392
2019-10-03
Journal of Statistics and Management Systems
Abstract:Swarm intelligence techniques are popular meta-heuristics which are developed to solve very complex problems that are not decipherable using traditional optimization approaches. In this class, artificial bee colony is latest and efficient algorithm that is able to solve complex problems. During solution search process it perform exploration in the early iterations, while should utilize the identified area in later stage. By analyzing the ABC, it is clear that the position update process depends on the difference of two random vector multiplied by a random component. So the step size of ABC is exceedingly hinge on these arbitrary components that leads to skipping of good solutions. For that reason, in this paper, Archimedean Spiral inspired Local Search (ASLS) algorithm is united with ABC and the new algorithm is labelled as Archimedean Spiral based ABC (ASABC) algorithm. In ASABC, in the current swarm only good solutions get chance to exploit the nearby search space using ASLS local search approach in every iteration. To demonstrate the performance of ASABC, experiment performed over 10 recognized benchmarks and compared with basic ABC, Best-so-far ABC and Modified ABC. Outcomes prove that ASABC algorithm may be considered as a viable algorithm for optimization problems.
What problem does this paper attempt to address?