Key links forecast with dynamic weight algorithm for RAW in optical networks

It had been proved that routing and wavelength assignment (RWA) in an arbitrary mesh optical network is an NP-complete problem. Previous works intended to solve this problem through various approaches, such as heuristics algorithm or graph theory. However, the computational time would increase exponentially with the problem size. Especially to a dynamic traffic model, the connection requests arrive and depart in a random manner and light-pathpsilas occupation depends on the current network utilization status, RWA problem become more complicated. Furthermore, the estimation model for link blocking probability is also quite inaccurate. People can find from realistic network where many more connections are requested between some source and destination node pairs than others. We consider the network performance deterioration is more likely caused by certain key links of network, which is related to static factor, such as particular network topology, and dynamic factors, such as wavelength exhausted rate as well as the number of free wavelength. In this paper, we propose an efficiently approximate approach, which comprehensively consider the factors of hops and free wavelengths based on the concept of key links, then dynamically choose and assign the optimal route and wavelength for arrival connection requests. The analysis result shows our algorithm can achieve better load balance and lower blocking probability with less computing complexity comparing to common RWA algorithms, particularly for larger sized networks.