Microcanonical mean field annealing: A new algorithm for increasing the convergence speed of mean field annealing

Hyuk Jae Lee, Ahmed Louri

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Scopus citations

Abstract

The authors consider the convergence speed of mean field annealing (MFA). They combine MFA with the microcanonical simulation (MCS) method and propose an algorithm called microcanonical mean field annealing (MCMFA). In the proposed algorithm, cooling speed is controlled by current temperature so that computation in the MFA can be reduced without degradation of performance. In addition, the solution quality of MCMFA is not affected by the initial temperature. The properties of MCMFA are analyzed with a simple example and simulated with Hopfield neural networks. In order to compare MCMFA with MFA, both algorithms are applied to graph bipartitioning problems. Simulation results show that MCMFA produces a better solution than MFA.

Original languageEnglish (US)
Title of host publication91 IEEE Int Jt Conf Neural Networks IJCNN 91
PublisherPubl by IEEE
Pages941-946
Number of pages6
ISBN (Print)0780302273
StatePublished - Dec 1 1991
Event1991 IEEE International Joint Conference on Neural Networks - IJCNN '91 - Singapore, Singapore
Duration: Nov 18 1991Nov 21 1991

Publication series

Name91 IEEE Int Jt Conf Neural Networks IJCNN 91

Other

Other1991 IEEE International Joint Conference on Neural Networks - IJCNN '91
CitySingapore, Singapore
Period11/18/9111/21/91

ASJC Scopus subject areas

  • Engineering(all)

Fingerprint Dive into the research topics of 'Microcanonical mean field annealing: A new algorithm for increasing the convergence speed of mean field annealing'. Together they form a unique fingerprint.

  • Cite this

    Lee, H. J., & Louri, A. (1991). Microcanonical mean field annealing: A new algorithm for increasing the convergence speed of mean field annealing. In 91 IEEE Int Jt Conf Neural Networks IJCNN 91 (pp. 941-946). (91 IEEE Int Jt Conf Neural Networks IJCNN 91). Publ by IEEE.