Fruit Fly Algorithm Based on Extremal Optimization

Aiming at the problem that solutions obtained by simply using FOA (Fruit fly optimization algorithm) would fall into local optimum, with slow convergence and low accuracy, we therefore propose an algorithm EOFOA that combines EO (Extremal optimization algorithm) with FOA. EOFOA brings the basic idea of EO, an algorithm of extremal dynamics, into FOA, so as to prevent local optimum and increase convergence accuracy by taking advantages of changing the initial distribution strategy, improving the diversity of the fruit fly population through randomly replacing the individuals that are not adaptive to the population, and making use of its strong volatility that enables the algorithm to search continuously. Simulation results show that: EOFOA is more accurate than FOA and the convergence is also quickened, which is efficient in avoiding premature convergence.