Simulated annealing (hull optimization)

Simulated annealing is a stochastic global optimizer used to refine hull form against an objective such as minimum wave resistance or maximum cargo volume subject to constraints. It mimics metallurgical annealing: from a current design it accepts a worse neighbor with probability exp(-Delta E / T), where T is a temperature lowered on a cooling schedule, so early high T lets the search escape local minima and late low T converges. Coupled to a wave-resistance solver and a hull parameterization, it explores the design space without gradients, which suits the noisy, multi-modal hull-optimization landscape. It is slower than gradient methods but robust to local traps.