Premium
Convergence of the generalized simulated annealing method with independent parameters for the acceptance probability, visitation distribution, and temperature functions
Author(s) -
de Andrade M. D.,
Mundim K. C.,
Malbouisson L. A. C.
Publication year - 2008
Publication title -
international journal of quantum chemistry
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.484
H-Index - 105
eISSN - 1097-461X
pISSN - 0020-7608
DOI - 10.1002/qua.21736
Subject(s) - simulated annealing , convergence (economics) , probability distribution , mathematics , function (biology) , distribution function , distribution (mathematics) , statistical physics , mathematical optimization , statistics , physics , mathematical analysis , quantum mechanics , evolutionary biology , economics , biology , economic growth
In their original form, the Generalized Simulated Annealing (GSA), proposed by Tsallis and Stariolo, was defined with two independent parameters, q a and q v , used in the definition of the acceptance probability, visitation distribution, and temperature functions. In the posterior applications of this algorithm, however, another independent parameter has been introduced, replacing q v in the definition of the temperature function, becoming more efficient and allowing a convergence with a small number of cycles. Nevertheless, there is no convergence proof of the GSA algorithm to the absolute minimum in this case. In this work it is presented a convergence proof of the GSA method to the absolute minimum, with three independent parameters, q a , q v , and q T , to define the acceptance probability, visitation distribution, and temperature functions, using a modified form of the distribution function, ′ g q v ,q T, in the formulation of the algorithm. © 2008 Wiley Periodicals, Inc. Int J Quantum Chem, 2008