Research On Shuffled Frog Leaping Algorithm

In recent years, with the rapid development of computer technology and the need of actual problems, a class of algorithms used to simulate biological communities have made rapid development. Current mainstream swarm intelligence optimization algorithms including genetic algorithm, ant colony algorithm, PSO and frog leaping algorithm, which uses the advantages of group mathematical description of the problem and is not required to meet differentiable and other conditions, is a completely distributed management. Shuffled Frog Leaping Algorithm(SFLA) is a relatively new meta-heuristic global optimization algorithm which simulated frog foraging and has been developed to solve combinatorial optimization problems. SFLA has the advantage of a few parameters, easy to understand and robustness which has been successfully used in many practical problems. However, SFLA is a relatively new algorithm, and is still in an immature stage both in theory and practice, which is prone to premature convergence, slow search speed, reduce the diversity of frog populations and easy to fall into local optimal solution and some other shortcomings for some of the complex issues.The optimization principle and mechanism of SFLA is studied in this paper. An improved Shuffled Frog Leaping Algorithm(ISFLA) is proposed to enhance the performance of computing. The Good Point set theory is used to generate the initial population of the improved algorithm, which makes the initial frog populations be more distributed in the feasible region more evenly, ensures the diversity of the initial population, and accelerate the entire optimization process of searching to the global optimal value. In the sub-populations change the search strategy which all frogs are searching for new places instead of just let the worst frog searching in the standard algorithm and design an adaptive mobile factor to adjust the moving step, accelerate the convergence speed and avoid premature convergence of the algorithm problems better.Experimental results show that the improved algorithm has faster convergence speed and higher precision which can effectively avoid the problem of premature convergence of the algorithm.In further research, the SFLA will be studied further. The work will focus on the analysis of parameter settings to improve the accuracy of the solution and execution speed. The SFLA combine with other groups of intelligent optimization algorithms can be used on more practical engineering problems.