A new bat algorithm based on a novel topology and its convergence

The basic bat algorithm (BA) is a swarm intelligent algorithm, which simulates the echolocation behavior of micro bats. Although BA is easy to implement and efficient in solving some continuous and discrete problems, it has some shortcomings such as premature convergence and low precision. To overcome these shortcomings, this paper presents a new bat algorithm (denoted as CABA) based on a novel topology. In this algorithm, by integrating the idea of cellular automaton, a novel topology is designed first, which is used to update the velocity of each bat. This topology can help bats to use the information provided by some good neighbors, which means it can help BA to reduce the risk of falling into local optimum. Second, in order to alleviate the difficulty of parameter adjustment in local search equations, a new local search equation is presented based on the number of iterations rather than loudness. Third, the convergence analysis is given based on martingale theory. The computational complex of the proposed algorithm is also discussed. Final, three comparison experiments and two practical optimization problems are conducted to test the performance of the proposed algorithm.