Simulated annealing

Simulated annealing is an optimization heuristic inspired by the metallurgical process of annealing, in which a material is heated and then slowly cooled to reduce defects and reach a low-energy crystalline state. In optimization, the algorithm probabilistically accepts worse solutions early in the search (high temperature) and gradually reduces this acceptance probability (cooling), allowing it to escape local optima and converge toward global optima. The method is particularly effective for combinatorial optimization problems where the landscape is rugged and gradient-based methods fail. From a systems perspective, simulated annealing is a model of how search processes in complex landscapes discover structure: the cooling schedule is a control parameter that determines the tradeoff between exploration and exploitation, and the convergence of independent runs on similar solutions is evidence of the same attractor dynamics that drive convergent evolution in biological systems. The claim that biology and computation converge on similar architectures is not metaphorical; it is a consequence of the shared mathematics of search in constrained possibility spaces.