基于混沌序列的最大和分散式协调算法

Max-sum decentralised coordination algorithm based on chaos

大规模信息系统的协调问题正成为新兴的研究热点,最大和协调算法显示了在该领域的应用前景,然而其收敛速度和鲁棒性有待进一步提高。针对以上问题,提出一种可用于合作系统中的基于混沌的分散式信息传递算法,并通过图形着色问题验证了该算法的有效性。在该算法中,利用混沌序列丰富的时空动态性,产生最大和协调算法的结点信息。然后依据更新规则对结点信息进行交互,完成算法的迭代。通过与传统最大和协调算法和遗传算法的对比实验结果分析,该算法在收敛速度及鲁棒性方面有着更好的表现。

Coordination problems of large scale information system gradually become a hot spot,and the max-sum product algorithrn shows a bright future of applications with improvements of the convergence speed and the robustness.This paper presents a decentralized message passing algorithm based on chaotic sequence which can be used in cooperative systems and is operated in a canonical coordination problem（graph coloring） performing decentralized coordination.The space-time dynamic messages of factor node are generated using chaotic sequence.Compared to max-sum product algorithm, the solution shows the better performance in both convergence speed and robustness.