μABC: A Micro Artificial Bee Colony Algorithm for Large Scale Global Optimization

Abstract

In this paper, we propose a new variant of Artificial Bee Colony Algorithm termed as μABC: Micro Artificial Bee Colony algorithm, which evolves with a very small population compared to its traditional version. In this approach the bees are ranked via their fitness. Best bee is kept unaltered, whereas the other bees are reinitialized with help of some modifications based on the food source obtained by best bee. This type of raking system will always help bees (apart from best bee) to exploit areas in the vicinity of food source corresponding to best bee. μABC is validated over a benchmark suite of shifted functions suggested in CEC’2008 competition and compared with the methods like EPSPSO, CCPSO2, etc. Various comparisons with dimensions greater than 100 show the performance of μABC in solving higher dimensional problems with less computational effort.