On Distributed , Probabilistic Algorithms for Computer Graphics

When attempting to solve a multi-variate optimisation problem, it is often a wise strategy to sacrifice a locally-optimal solution in the hope of finding a better global solution. Algorithms that solve optimisation problems in this manner are sometimes called hill climbing algorithms . Recently, several new hillclimbing approaches have been proposed and have gained in popularity for solving special classes of optimisation problems . These approaches are probabilistic in nature. We introduce one particular approach called simulated annealing, and show how a distributed version of it can be used to solve a basic problem in computer graphics and image processing: colour q uantisation.