基于免疫--蚂蚁算法的多约束QoS路由选择

针对多约束QoS路由选择问题,将其转化为一个多约束赋权图最短路径问题,选择费用、带宽、时延、丢失率为QoS参数。借鉴人体免疫系统的适应能力和蚂蚁算法的全局寻优能力提出了一种新的融合算法即免疫——蚂蚁算法。免疫算法把目标函数和制约条件作为抗原,目标函数的优化解对应为抗体,使得求解过程的收敛方向得以控制;利用蚂蚁算法产生和更新抗体,抗体交叉、变异操作以及对与抗原亲和力高的抗体进行记忆,均能促进快速求解。实验结果表明:免疫——蚂蚁算法表现出了超越免疫算法和蚂蚁算法的优点,大幅度提高了路由选择的效率。

A multiple constrained QoS routing based on immune-ant algorithm

The paper transforms multiple constrained QoS routing problem as the most short path problem of multiple constrained assign-weight chart through aiming at it. It chooses bandwidth, delay, rate of losing as the parameters of QoS. The paper presents a new mixed algorithm, immune-ant algorithm, by using the ability of adaption of human's immune system for reference and using the ability of searching the global optimization of ant algorithm. Immune algorithm makes object function and subject term as antigen, and the object function corresponds to the antibody. It can control the convergence direction of the solving process. Ant algorithm generates and renews antibody. The hybrid and mutation of antibody and the memory of antigen of high affinity can promote the solving process. The experiment results show that the immune-ant algorithm has more advantages than immune algorithm and ant algorithm. And it improves the efficiency of routing greatly.