Abstract :
In this paper, we investigate the static multicast advance reservation (MCAR) problem for all-optical wavelength-routed WDM networks. Under the advanced reservation traffic model, connection requests specify their start time to be some time in the future and also specify their holding times. We investigate the static MCAR problem where the set of advance reservation requests is known ahead of time. We prove the MCAR problem is NP-complete, formulate the problem mathematically as an integer linear program (ILP), and develop three efficient heuristics, seqRWA, ISH, and SA, to solve the problem for practical size networks. We also introduce a theoretical lower bound on the number of wavelengths required. To evaluate our heuristics, we first compare their performances to the ILP for small networks, and then simulate them over real-world, large-scale networks. We find the SA heuristic provides close to optimal results compared to the ILP for our smaller networks, and up to a 33% improvement over seqRWA and up to a 22% improvement over ISH on realistic networks. SA provides, on average, solutions 1.5-1.8 times the cost given by our conservative lower bound on large networks.
Keywords :
computational complexity; heuristic programming; integer programming; linear programming; multicast communication; optical fibre networks; telecommunication network routing; telecommunication traffic; wavelength assignment; wavelength division multiplexing; NP-complete problem; advanced reservation traffic model; all-optical wavelength-routed WDM networks; integer linear programming; multicast advance reservation; seqRWA heuristics; static MCAR problem; static multicast advance reservation; static routing and wavelength assignment; theoretical lower bound; Color; Correlation; Image color analysis; Optical fiber networks; Optical switches; Routing; WDM networks; Advance reservation; NP-complete; heuristics; integer linear program (ILP); multicast; routing and wavelength assignment (RWA); scheduled demands; simulated annealing;