Simulated Annealing is a probabilistic metaheuristic inspired by the annealing process in metallurgy. Starting from an initial solution and high temperature T, the algorithm: (1) generates a neighbor solution; (2) accepts improvements always; (3) accepts worse solutions with probability exp(-Δf/T) where Δf is the cost increase; (4) gradually decreases temperature according to a cooling schedule. This probabilistic acceptance allows escaping local optima. Key parameters: initial temperature, cooling schedule (exponential, linear, logarithmic), and stopping criterion. SA provides theoretical convergence guarantees to global optima under appropriate conditions and is used in scheduling, routing, circuit design, and molecular modeling.