Abstract: In order to solve the hardware/software bi-partitioning problem in embedded system and system-on-a-chip design，a novel bacterial chemotaxis optimization method（BCA） is proposed.It is a novel optimization algorithm which simulates the bacterial foraging behavior.Control data flow graph with different nodes is synthesized.Experiments demonstrate that the consumed time of this BCA is less than that of simulated annealing，tabu search and ant colony algorithm at the same conditions.The more the nodes，the larger predominance the BCA shows.When the nodes are as many as 100，the consumed time of the proposed algorithm is only 40%～60% of those of other popular algorithms.

Key words: embedded system, co-design, hardware/software partitioning, bacterial chemotaxis algorithm

摘要： 引进细菌趋药性算法（BCA），用于嵌入式系统和片上系统的软硬件双路划分。BCA是一种新型的对细菌觅食行为进行模拟的优化算法。对不同节点的控制数据流图进行仿真，表明在同等条件下，BCA收敛时间低于模拟退火算法、禁忌搜索与蚁群算法，节点规模越大，优势越明显。当节点规模高达100时，BCA消耗时间仅有常用优化算法的40%～60%。

关键词: 嵌入式系统, 协同设计, 软硬件划分, 细菌趋药性算法