CONVERGENCE PROPERTIES OF SIMULATED ANNEALING FOR CONTINUOUS GLOBAL OPTIMIZATION

Authors
LOCATELLI M
Citation
M. Locatelli, CONVERGENCE PROPERTIES OF SIMULATED ANNEALING FOR CONTINUOUS GLOBAL OPTIMIZATION, Journal of Applied Probability, 33(4), 1996, pp. 1127-1140
Citations number
19
Categorie Soggetti
Statistic & Probability","Statistic & Probability
Journal title
Journal of Applied Probability → ACNP
ISSN journal
00219002
Volume
33
Issue
4
Year of publication
1996
Pages
1127 - 1140
Database
ISI
SICI code
0021-9002(1996)33:4<1127:CPOSAF>2.0.ZU;2-1
Abstract
In this paper conditions for the convergence of a class of simulated a nnealing algorithms for continuous global optimization are given. The previous literature about the subject gives results for the convergenc e of algorithms in which the next candidate point is generated accordi ng to a probability distribution whose support is the whole feasible s et. A class of possible cooling schedules has been introduced in order to remove this restriction.