Learning and Evolution: An Introduction to Non-darwinian Evolutionary Computation

摘要: The field of evolutionary computation has drawn inspiration from Darwinian evolution in which species adapt to the environment through random variations and selection fittest. This type found wide applications, but suffers low efficiency. A recently proposed non-Darwinian form, called Learnable Evolution Model or LEM, applies a learning process guide processes. Instead mutations recombinations, LEM performs hypothesis formation instantiation. Experiments have shown that may speed-up an by two more orders magnitude over Darwinian-type algorithms terms number births (or fitness evaluations). price is higher complexity instantiation mutation recombination operators. appears be particularly advantageous problem domains evaluation costly time-consuming, such as design, complex optimization problems, fluid dynamics, evolvable hardware, drug others.