Generalized extremal optimization: an application in heat pipe design (Q1776690)

A global-optimization algorithm, called ``generalized extremal optimization'', is presented and applied to one specific problem. The algorithm is a global search meta-heuristic, as the genetic or the simulated-annealing algorithms, but adjusting only one free parameter at each step. No information about gradient of the cost functional is needed, of course. The illustration problem, described in detail, concerns a lumped-parameter model of a heat pipe and its optimization involves only 7 scalar variables and 18 constraints. Numerical results of this case study are exposed. No comparison with other optimization algorithms is presented.