摘要
针对网络资源管理中的负载均衡与优化问题,提出一种多蚁群网络负载动态均衡方法,采用网络流量工程理论中拥塞控制机制实现信息素随网络流量动态释放与更新。算法通过蚁群间信息素的动态相互作用(蚁群内信息素相互增强,蚁群间信息素相互削弱),将代表网络负载的蚂蚁合理分配到可用路径,避免蚂蚁集中到特定路径而造成网络拥塞。实验结果表明,通过路径信息素控制能够实现网络负载均衡,有效提高网络在路径延时、平均带宽利用率和平均丢包率方面的性能。
A muhiple Ant Colony Algorithm (ACA) path load balancing optimization algorithm for the resources allocation within network sessions was proposed. Combining bio-inspired ant colony algorithm with dynamic load balancing in the network traffic engineering, the algorithm further extended the basic ant colony to multiple ones with dynamic pheromone release design. By fully considering ACA's parallel-computing characteristic, the scheme tried to allocate the traffic load of network sessions to the different paths equally to avoid traffic congestion. Besides analyzing the effective control of the pheromone, compared with basic ant algorithm, simulation experiments illustrate that the proposed scheme achieves better resources allocation in terms of path bandwidth utilization, session delay and session packet loss.
出处
《计算机应用》
CSCD
北大核心
2008年第3期572-574,共3页
journal of Computer Applications
关键词
网络资源管理
蚁群算法
负载均衡
network resource management
Ant Colony Algorithm (ACA)
load balancing