Abstract

In the initial period, the peculiarity of artificial fish swarm algorithm is of fast searching speed and high optimization accuracy, but in the later period, the convergence speed is always slow, and artificial fish tend to gather around the local optimum. Therefore, the solving ability of the algorithm becomes weak and the global optimal value is hard to obtain. Considering the introduction of RNA computation based on biomolecular operations, the optimization capability of traditional algorithm can be enhanced effectively. Therefore, RNA computing is introduced to artificial fish swarm algorithm, and a modified artificial fish swarm algorithm is presented on the grounds of RNA computing. In the later period of artificial fish swarm algorithm, the transformation, replacement and recombination operations in RNA computation are applied to increase diversity of artificial fish, so as to further the convergence speed and optimization capability of the algorithm. In the meantime, the improved algorithm, RNA-AFSA, is tested by four typical functions, and the results prove that the modified artificial fish swarm algorithm has better optimization effects in search accuracy, stability, and other aspects.

中文翻译：

---

RNA计算优化的人工鱼群算法

摘要

在初期，人工鱼群算法的特点是搜索速度快，优化精度高，但后期收敛速度总是很慢，人工鱼往往聚集在局部最优值附近。因此，算法的求解能力变弱，难以得到全局最优值。考虑到基于生物分子操作的RNA计算的引入，可以有效增强传统算法的优化能力。因此，将RNA计算引入人工鱼群算法中，并在RNA计算的基础上提出了一种改进的人工鱼群算法。人工鱼群算法后期，经过改造，应用RNA计算中的替换和重组操作来增加人工鱼的多样性，从而进一步提高算法的收敛速度和优化能力。同时对改进后的RNA-AFSA算法进行了四个典型函数的测试，结果证明改进后的人工鱼群算法在搜索精度、稳定性等方面具有较好的优化效果。