A Survey of Biogeography-Based Optimization
Weian Guo,Ming Chen,Lei Wang,Yanfen Mao,Qidi Wu
DOI: https://doi.org/10.1007/s00521-016-2179-x
2016-01-01
Neural Computing and Applications
Abstract:Optimization is a classical issue and in many areas that are bound up with people’s daily life. In current decades, with the development of human civilization and industry society, many complicated optimization problems are raised. In the meantime, corresponding novel approaches are constantly proposed for solving these problems. One of them is meta-heuristics, which is inspired from natural phenomena and contains many kinds of algorithms. The classical meta-heuristic algorithms have exhibited their superiority in dealing optimization problems, especially for specific problems such as combinatorial optimization. As a novel meta-heuristic algorithm, biogeography-based optimization (BBO), inspired from the science of biogeography, has its own characteristics and exhibits a huge potential in computation and optimization. According to current investigations and analysis on this algorithm, it has not only achieved a great success in numerical optimization problems, but also been implemented in various kinds of applications, and drawn worldwide attentions. In this paper, we present a survey for this algorithm. First, we introduce the basic operators of BBO, including migration and mutation. For migration operator, it mimics species migration among islands, which provides a recombination way for candidate solutions to interact with each other so that the whole population can be improved. Besides linear migration model, several other popular migration models are also introduced and the corresponding performances are analyzed. For mutation operator, the design of BBO is different from other meta-heuristics. In standard BBO, different candidate solutions have different migration rates and the rate assignment is influential to BBO’s performance. Second, we summarized some popular variants of BBO and related hybrid algorithms that significantly enhance BBO’s performance. This part introduces the development of this algorithm and helps readers understand the way to choose a suitable version of BBO for a given problem. The way to improve algorithms’ performances helps readers design new variants of BBO for specific problems. Third, we present the evaluation of BBO’s performance for both numerical and practical problems. The results demonstrate BBO is competent to solve optimization problems. Despite so many achievements of BBO, some open issues that should be considered and solved in future work in order to make this algorithm more competitive in meta-heuristics.