ABSTRACT
This chapter shows that heuristic techniques, often applied in Computational Intelligence, Artificial Intelligence, and Operational Research optimization problems, can be used in optical networks to provide near-optimal and fast solutions for efficient network design and management. It describes Simulated Annealing, a heuristic technique that can eficiently deal with optimization problems that involve both integer and real variables. Modern heuristic methods have evolved significantly during the last decade and have been applied to a number of optimization problems. Another argument in favor of heuristics is their flexibility and capability “of coping with more complicated objective functions and/or constraints than exact algorithms” does. Genetic Algorithms were originally introduced in 1975. Tabu Search has its origin in 1970 when it was used in combinatorial procedures applied to nonlinear covering problems. The Artificial Neural Network method emulates the structure of biological neural networks and their natural way to encode and solve a given problem.
