BIS: A New Swarm-Based Optimisation Algorithm

This paper presents a novel swarm-based search algorithm: the bio-breeding intelligent swarm (BIS) algorithm. BIS agents imitate the offspring and maturity phases of the typical lifecycle of an animal. As in nature, the BIS algorithm makes gender distinction among agents and the main search strategy exploits competition between male agents in an attempt to provide a better location for females. BIS agents embark on various nature-inspired mating strategies and the inspiration for the reproduction model is derived from temperature-dependent sex determination (TSD), a reptilian reproduction system. The BIS algorithm's TSD inspired reproduction model enables female agents to control the gender of offsprings based on guidance provided by their male mates, subsequently resulting in regulation of the male-female ratio in the swarm which in turn auto-controls the balance of exploration and exploitation within the population of agents. The efficiency of the BIS algorithm was tested over a wide range of benchmarks including unconstrained high dimensional and real-world problems. The BIS algorithm performed very well in comparison with a number of leading population-based stochastic search methods, finding the highest number of global optimums.