Abstract: To overcome the drawbacks of local optima and instability involved in Shuffled Frog Leaping Algorithm（SFLA）, an improved SFLA is proposed. The proposed algorithm employs Opposition Based Learning（OBL） to generate the initial population, which can obtain better initial candidate solutions. During the course of evolvement, the Differential Evolution（DE） is embedded in SFLA organically to maintain the population diversity. Numerical results show that the proposed SFLA has a better capability to solve complex functions than other algorithms.

Key words: Shuffled Frog Leaping Algorithm（SFLA）, opposition, Differential Evolution（DE）

摘要： 针对混合蛙跳算法在优化过程中受初始值影响较大且容易陷入局部最优的缺陷，提出了一个改进的混合蛙跳算法，该算法利用基于对立学习的策略产生初始种群，提高了产生解的质量；在进化过程中，将差分进化有机地嵌入其中，维持了种群的多样性。数值结果表明，改进的混合蛙跳算法对复杂函数优化问题具有较强的求解能力。

关键词: 混合蛙跳算法, 对立策略, 差分进化