BBO as a Markov Process
Haiping Ma, Dan Simon · 2017
This chapter explores on Markov theory that provides insight into biogeography-based optimization (BBO) behavior. It develops a Markov model for the basic BBO algorithm. The chapter analyzes the convergence properties of BBO for binary problems. It develops and discusses Markov models of BBO extensions. Markov theory is a good way to answer theoretical questions about bio-inspired optimization algorithms, and it might also lead to unexpected and new avenues of research. The study of bio-inspired optimization algorithms has often been ad hoc, simulation-based and non-analytic. Markov models have been a valuable theoretical tool to analyze bio-inspired optimization algorithms, including simple genetic algorithms and simulated annealing. In BBO, two main steps are significant, migration and mutation, so the transition probability includes the migration probability and the mutation probability for one generation. The mutation rate affects the convergence rate of BBO.