Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

A gradient system solution to Potts mean field equations and its electronic implementation

K Urahama1, S Ueno

  • 1Department of Computer Science and Electronics, Kyushu Institute of Technology, Fukuoka, Japan.

International Journal of Neural Systems
|March 1, 1993
PubMed
Summary

A novel gradient system method solves combinatorial optimization problems with winner-take-all constraints. This approach guarantees legal local optima and is implemented using analog electronic circuits for efficient problem-solving.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Interaction between catechol-O-methyltransferase (COMT) Val108/158Met and brain-derived neurotrophic factor (BDNF) Val66Met polymorphisms in age at onset and clinical symptoms in schizophrenia.

Journal of neural transmission (Vienna, Austria : 1996)·2006
Same author

1-BP inhibits NF-kappaB activity and Bcl-xL expression in astrocytes in vitro and reduces Bcl-xL expression in the brains of rats in vivo.

Neurotoxicology·2006
Same author

Effects of canopy gaps on the genetic structure of Camellia japonica saplings in a Japanese old-growth evergreen forest.

Heredity·2006
Same author

Preventive effects of CIDR-based protocols on premature ovulation before timed-AI in Ovsynch in cycling beef cows.

Veterinary journal (London, England : 1997)·2006
Same author

Gastric mucosal hyperplasia via upregulation of gastrin induced by persistent activation of gastric innate immunity in major histocompatibility complex class II deficient mice.

Gut·2005
Same author

A carrier of human T-lymphotropic virus type I with periodic limb movements progressing to myelopathy.

Neurology·2005

Area of Science:

  • Computational Mathematics
  • Analog Electronics
  • Operations Research

Background:

  • Combinatorial optimization problems often involve complex constraints, such as winner-take-all mechanisms.
  • Existing methods may struggle with finding guaranteed local optima within feasible solution spaces.
  • Potts mean field equations are frequently used for modeling such problems.

Purpose of the Study:

  • To present a gradient system solution method for Potts mean field equations under winner-take-all constraints.
  • To theoretically prove the method's ability to find legal local optima.
  • To design and simulate an analog electronic circuit implementation of the proposed method.

Main Methods:

  • A gradient descent differential equation approach is employed, confining solution trajectories within the feasible space.

Related Experiment Videos

  • The method is theoretically validated to ensure the generation of legal local optima.
  • An analog electronic circuit based on current-mode subthreshold MOS technologies is designed, incorporating a winner-take-all circuit.

    • Main Results:

    • The gradient system method is proven to consistently yield legal local optimum solutions.
    • An analog circuit implementation was successfully designed and simulated.
    • Simulations demonstrated the circuit's efficacy in solving shortest path problems.

    Conclusions:

    • The proposed gradient system method offers a theoretically sound approach for solving constrained combinatorial optimization problems.
    • The analog electronic circuit provides a practical and efficient hardware implementation for real-time applications.
    • This work bridges theoretical optimization with practical electronic circuit design for complex computational tasks.