A Multiobjective SFLA-Based Technique for Predicting Motifs in DNA Sequences

摘要

In recent years design of new evolutionary techniques for addressing optimization problems is being a booming practice. Furthermore, considering that the vast majority of real optimization problems need to simultaneously optimize more than a single objective function (Multiobjective Optimization Problem - MOP); many of these techniques axe also adapted to this multiobjective context. In this paper, we present a multiobjective adaptation of one of the last proposed swarm-based evolutionary algorithms, the Shuffle Frog Leaping Algorithm (SFLA), named Multiobjective Shuffle Frog Leaping Algorithm (MO-SFLA). To evaluate the performance of this new multiobjective algorithm, we have applied it to solve an important biological optimization problem, the Motif Discovery Problem (MDP). As we will see, the structure and operation of MO-SFLA makes it suitable for solving the MDP, achieving better results than other multiobjective evolutionary algorithms and making better predictions than other well-known biological tools.