作者: John N. Tsitsiklis
DOI: 10.1007/978-1-4613-8762-6_33
关键词:
摘要: Simulated annealing is a probabilistic algorithm for minimizing general cost function which may have multiple local minima The amount of randomness in this controlled by the “temperature”, scalar parameter decreased to zero as progresses. We consider case where minimization carried out over finite domain and we present survey several results analytical tools studying asymptotic behavior simulated algorithm, time goes infinity temperature approaches zero.