Hardware/software partition using adaptive ant colony algorithm
Guo Yong-liangc · Kongzhi yu juece · 2009
In order to solve the hardware/software bi-partitioning problem more efficiently,a novel adaptive ant colony algorithm(AACA) is proposed.The basic idea is to adaptively adjust the state transform probability and the pheromone evaporation factor,which ensures that the randomness of the ant colonies is high enough at the initial for global search and low at the later stage for local search for faster convergence.Experiments synthesize different nodes control data flow graphs,and show that the proposed method has superiority over improved simulated annealing(ISA),improved tabu search(ITS),improved ant colony algorithm(IACA),and DCG3A in terms of global convergence rate and computation time.The more the nodes are,the obvious the superiority is.