基于矢量测量的多约束路径选择

Normal Measure Based Multi-Constrained Path Selection

多约束路径(multi-constrained path,简称MCP)选择问题是QoS路由问题面临的重要挑战之一.现有的MCP算法不能兼顾降低计算复杂性、提高响应速度和防止可行解丢失等方面的缺点.另外,单纯依靠线性路径长度方程(LPLF)或非线性路径长度方程(NLPLF)都不能有效解决QoS路由问题.定义了崭新的法线测量路径长度方程,并基于该方程提出了解决m约束MCP问题的NMMCP(normal measure based MCP)算法.NMMCP不仅是在线计算与预计算,同时也是LPLF与NLPLF的良好折衷.通过引入Pareto最优理论,NMMCP具有非线性前瞻机制.大量仿真实验表明,NMMCP解决MCP问题是非常有效的.

Multi-constrained path （MCP） selection is one of the great challenges that QoS routing （QoSR） faces. Existing algorithms cannot make a good tradeoff among computation complexity, response speed and preventing from losing feasible solutions. Furthermore, neither linear path length function （LPLF） nor non-linear path length function （NLPLF） alone can solve QoS routing problems. A novel normal measure based path length function is defined and based on it, a normal measure based MCP （NMMCP） algorithm is proposed to solve m-constrained MCP problems. NMMCP makes a good tradeoff not only between on-demand computation and pre-computation, but also between LPLF and NLPLF based algorithms. By introducing Pareto optimal mechanism, NMMCP has nonlinear look-ahead ability. Extensive simulations show that NMMCP is very efficient when both performance and computation cost are taken into account.