摘要
为了平衡网络负载,优化网络资源使用,路由算法必须具备快速规避拥塞的能力。文章借鉴了最短路径算法和蚂蚁算法,在拥塞出现以后,分别以拥塞链路的相邻两端节点为中心、以一个事先设定的路径优劣评价函数为标准,通过两组寻路蚂蚁,逐渐增大网络搜索范围,各自独立寻找绕开拥塞链路到达对方的最优相向路径。直到上述找到的相向路径中途相遇(节点集中出现相同的节点),则意味着算法成功找到了绕开了拥塞链路新路径,算法结束。经仿真证明,该算法由于充分地利用了拥塞前的QoS路由信息,加上具体寻路过程利用了蚂蚁算法的并发性,明显节省了QoS路由搜索时间,有效地规避了网络拥塞,提高了网络服务能力。
For the sake of load balancing and resource use optimiz!ng, the ability of congestion avoidance is absolutely necessary in QoS routing algorithm. In this paper, profited from Dijkstra and ant colony system algorithm, a new algorithm is suggested. Once congestion emerges, encircling respectively by the two nodes connected to the congestion link, two groups of ants are set out to look for another better routings connecting them via an early - set judgmental function. The searching range of network is increased gradually until two routings form different direction having the same node finally, which means that the algorithms succeed in discovering a new routing excluding the congestion link. As proved by the example followed, benefited from making use of the information within the old routing and ant colony system algorithm' s parallel characteristic, the new algorithm succeeds in reducing the time of routing searing and then solving the congestion problem, thereby enhancing the service ability of network..
出处
《计算机仿真》
CSCD
2007年第4期136-138,148,共4页
Computer Simulation
关键词
路由
拥塞
最短路径法
蚂蚁算法
多媒体
Routing
Congestion
Dijkstra
Ant colony system algorithm
Multimedia