【作者】 杨勇；

【导师】 张晓萍；

【作者基本信息】 兰州大学 ， 无线电物理， 2009， 硕士

【摘要】 波分复用光网络中的路由和波长分配(RWA,routing and wavelengthassignment)问题是一个典型的组合优化问题,主要指在综合考虑网络可用资源及建设成本的前提下,为业务选择合适的路由和合理的分配波长以提高网络的效率。如何设计出良好的路由选择与波长分配算法来充分利用网络资源及降低呼叫阻塞率成为光网络技术中重要的研究课题,近年来有许多优化算法,如遗传算法,蚁群算法,粒子群算法等,都被用于光网络的RWA问题的研究,但上述算法普遍存在实时性不高,计算量大的缺点,特别是在网络规模比较大时。因此,有必要研究一种快速,高效率的光网络RWA算法以适应当今网络规模超大型化的发展趋势。本文将一种改进的脉冲耦合神经网络算法与基于分层图模型的动态波长分配算法相结合求解动态RWA问题,并对WDM网络中的动态选路和波长分配过程进行了模拟仿真,结果显示了改进PCNN算法在解决RWA问题上的有效性和可行性,还通过计算实例将改进的PCNN算法与DIJKSTRA算法所得结果进行了比较,表明了改进PCNN算法具有一定的优越性,降低了网络阻塞率,提高了对网络资源的利用。为进一步研究大规模光网络的路由和波长分配问题提供了新的思路。更多还原

【Abstract】 RWA (routing and wavelength assignment) problem in WDM Optical Network is a classic combinatorial optimization problem. RWA contains selecting appropriate route and contributing rational wavelength for operations. A good RWA scheme should improve network’s efficiency after considering the resource and construction cost of the network. So how to design a good RWA algorithm in order to make full use of network resource and low the call blocking rate becomes an important subject in study of optical network technology. In recent years, there are many optimization algorithms like: genetic algorithm, ant swarm optimization, partial swarm optimization, and so on, all be used in optical network ’s RWA problem, but all the algorithms above have the low real-time response and mass computation defects, especially when the network has a huge scale. Based on this background, it’s necessary to find a rapid, high efficient RWA algorithm of optical network to fit the development trend of super-scale network.Aiming at RWA problem under dynamic traffic flow, an intelligent algorithm based on improved pulse-coupled neural network (PCNN) was introduced into WDM routing in this paper, the wavelength selection made use of layer model. Computer simulation shows the improved PCNN is valid and feasible in solving RWA problem, and the result also shows being compared with DIJKSTRA algorithm, the improved PCNN is ascendant, this provides some new illumination for further study in huge-scale optical network’s RWA problem.更多还原