The aim of this work is to study the Tabu Search algorithm as a graph coloring technique for wavelength assignment in optical networks, a crucial function in optical network planning. The performance of the Tabu Search is assessed in terms of the number of wavelengths and computation time and is compared with the one of the most common Greedy algorithm. It is concluded that for real networks with a large number of nodes
and a higher variance node degree of the path graph relatively to its average node degree value, the Greedy algorithm is preferable to the Tabu Search algorithm since it returns the same number of colors of Tabu Search, but in a shorter computation time.