An adaptive approach to fast simulation of traffic groomed optical networks

In this paper, we consider the fast simulation of traffic groomed optical networks, in which multiple subrate traffic streams may be carried on the same wavelength. For real-sized systems, call blocking probabilities may be rare events and require a large amount of CPU power to simulate. We present two importance sampling methods for fast simulation. For a light-load case, we prove that static IS using the standard clock (S-ISSC) method does indeed have bounded relative error (BRE) even in multiclass case. For a nonlight-load case, we suggest, alternatively, adaptive ISSC (A-ISSC) which calculates the relative possibility of reaching each target in every step of simulation. Using A-ISSC, biasing methods which are proven to be optimal or have bounded error can be extended to multidimensional cases while still retaining a very favorable performance.