Bacterial foraging optimization with memory and clone schemes for dynamic environments

Abstract

Dynamic optimization problems (DOPs) are prevailingly addressed because of their origins from real-world issues. In addition to existing methods that have been developed for evolutionary algorithms to solve DOPs, this paper provides a hybrid memory and clone scheme, called memory-based clone selection, for bacterial foraging optimization algorithms in dynamic environments. Meanwhile, two gene libraries (Random and Non-random) are involved to clone outstanding individuals and dynamically manage the gene hall and memory. This approach not only results in greater diversity and better global search ability, but also enables the algorithm higher adaptability to environmental dynamics and changes. The simulation result generated by a dynamic rotation peak benchmark confirms that proposed memory and clone schemes-based BFO (MCBFO) outperforms standard BFO and PSO in terms of population diversity, convergence rate and searching ability.