Utilizing time-linkage property in DOPs: An information sharing based Artificial Bee Colony algorithm for tracking multiple optima in uncertain environments

An information sharing artificial bee colony (ABC) algorithm has been proposed for locating and tracking multiple peaks in non-stationary environments. The niching method has been adapted by hybridizing two techniques. A modified variant of the fitness sharing has been used for detecting multiple peaks simultaneously and a speciation based technique is employed to keep the better individuals of the previous generation. The base algorithm used here is a modified variant of ABC that helps to synchronize the employer and onlooker forager swarms by synergizing the local information. The main crux of our algorithm is its independency of the problem dependent control parameters, like niche radius, and the absence of any hard-partitioning technique that leads to high computational burden. Our framework aims at bringing about a simple, robust approach that can be applied to a variety of dynamic functional landscapes. Experimental investigations are undertaken on standard benchmarks focussing on the competitive performance of our algorithm in contrast to the existing state-of-the-art to highlight the significance of our work.