GMPLS(generalized multi protocol label switching)网络中的多约束QoS路由问题是要在诸如带宽、代价和延迟的约束条件下找到一条优化的路径.这个问题通常被认为是一个NP-完全问题.在研究共享风险链路组具有的启发信息的基础上,提出了一种具有共享风险链路启发信息的多约束预计算算法.该算法包含预计算和搜索两个部分.预计算主要是能创建和更新每个节点上的路由表.而后,搜索部分则可以在层次化的结构中选择满足约束条件的优化的路径.大量仿真数据表明,相应的方法能够取得满意的结果,可以有效地解决GMPLS网络中多约束的QoS路由问题. The problem of multi-constrained QoS routing in GMPLS is to find an optimal path under constraints such as bandwidth, cost and delay.This problem is usually considered as an NP-complete problem.In the study of sharing Based on the heuristic information of risk chain group, this paper proposes a multi-constraint pre-computation algorithm with shared risk-heuristic information, which contains two parts: pre-computation and searching.Pre-computation is mainly to create and update each Then, the search part can choose the optimal path which satisfies the constraint in the hierarchical structure.A large number of simulation data show that the corresponding method can achieve satisfactory results, which can effectively solve the problem that many of the GMPLS networks Constrained QoS routing problem.