A Novel Comparative Study between Dual Population Genetic Algorithm and Artificial Bee colony Algorithm for Function Optimization

This paper conducts a comparative study between an improved variants of genetic algorithm (GA) and a swarm intelligence algorithm (SIA), which are the Dual population Genetic Algorithm (DPGA) and Artificial Bee Colony (ABC) Algorithm. DPGA is a multi-population genetic algorithm (MPGA) that implements two population such as the main population and a complementary population. Since the added population has a totally different fitness function, it is preserved to supply sufficient diversity to the main population by crossover operation. However DPGA employs a dynamic strategy to maintain an appropriate distance between two populations for maintaining diversity. On the contrary ABC, a simple but exceptional derivative of SIA, applies division of labor in single population of artificial bees and allocates some of them to exploration while the others to exploitation. DPGA has its own techniques to sustain the significant balance between exploration vs. exploitation. Thus many such analytical comparisons between DPGA and ABC are the center of attention of this paper. Experiments are conducted on seven benchmark functions using ABC and results are compared with DPGA. The results demonstrate that DPGA performs well for some of the functions but by considering the result of mean absolute error, ABC performs far better than DPGA.