Non-linear Simplex Shuffled Frog Leaping Algorithm

Abstract

Shuffled frog leaping algorithm (SFLA) is a recent addition to the family of memetic algorithms that takes its inspiration from the natural foraging behavior of frogs. In SFLA the colony of frogs is divided into memeplexes of equal size. SFLA gathered the interest of research fraternity to solve many real world complex optimization problems. The basic structure of SFLA posses some inherent limitations. In order to overcome the limitation, in this study an enhanced variant of SFLA is proposed and named as NL-SFLA. Generally the initial population is generated using a traditional pseudo-random numbers which may not be much efficient. In NL-SFLA an attempt has been made to initialize the population of frog by integrating the concept of nonlinear simplex method of Nelder and Mead. Later modification is done in the frog distribution scheme in memeplexes to handle continuous optimization problems. Numerical results of NL-SFLA are compared with the state-of-art algorithms over a set of benchmark problems. Also the efficiency of the proposal is investigated on four real world problems. Simulated results signify the efficacy of the proposal.