Dynamical neural schema for quadratic discrete optimizations problems

E. Goles, G. Hernandez, M. Matamala

Research output: Contribution to journalConference articlepeer-review

2 Scopus citations


We present a Dynamical Neural Schema (DNS) that gives very good approximation at global minimum of discrete quadratic problem (P). We test DNS with classical schema and for fixed sigmoid functions. For the computational experiments we take 20×20 symmetric matrices, where we known the global minimum on {0,1}n, and also 40×40 random symmetric matrices. In almost all the cases DNS is better. Experimental comparisons will be given and also theoretical results concerning relations between fixed points of DNS and global optimum at problem (P).

Original languageEnglish
Pages (from-to)96
Number of pages1
JournalNeural Networks
Issue number1 SUPPL
StatePublished - 1988
Externally publishedYes
EventInternational Neural Network Society 1988 First Annual Meeting - Boston, MA, USA
Duration: 6 Sep 198810 Sep 1988


Dive into the research topics of 'Dynamical neural schema for quadratic discrete optimizations problems'. Together they form a unique fingerprint.

Cite this