一种多智能体混合蛙跳算法

A Multi-agent Shuffled Frog Leaping Algorithm

提出一种多智能体混合蛙跳算法。将智能体固定在智能体网格上,每个智能体通过与其邻居的竞争与合作,结合混合蛙跳算法的进化机制,不断感知局部环境,并逐渐影响整个智能体网格,以提高自身对环境的适应能力。为更好地适应环境,智能体也可以利用自身的知识进行自学习。仿真实验结果表明,该算法能有效地维持种群的多样性,提高优化精度,同时抑制早熟现象,在高维函数优化方面具有较高的优化性能。

This paper proposes a Multi-agent Shuffled Frog Leaping Algorithm（MSFLA） by introducing the multi-agent system to the Shuffled Frog Leaping Algorithm（SFLA）. This algorithm fixes the agent on grid, with the competition and cooperation with its neighbors, and combining the evolution mechanism of the SFLA. Each agent unceasingly senses local environment, and gradually affects the whole agent grid, so that it enhances its adaptiveness to the environment. The agent also makes self-study by using its knowledge to enhance its adaptiveness to the environment. By the test of high dimension benchmark functions, the results illustrate this algorithm this algorithm can effectively maintain the diversity of the population, increase the precision of optimization, simultaneously, efficiently restrain the prematurity, and has higher optimization performance in the field of high dimension functions optimization.