Asymptotics in Time, Temperature and Size for Optimization by Simulated Annealing: Theory, Practice and Applications
Abstract
This project was concerned with simulated annealing, a Monte Carlo method for obtaining globally optimal or nearly globally optimal solutions to a variety of optimization problems. Results were obtained in two main areas: 1) Characterizing the cooling rate necessary and sufficient for simulated annealing to hit the global minimum and ii) obtaining an upperbound for the time-constant of convergence of simulated annealing at a fixed temperature to its equilibarium distribution and studying the growth of this bound as the temperature approaches zero asymptotically.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 19, 1990
- Accession Number
- ADA217680
Entities
People
- P. R. Kumar
- Vasant B. Rao
Organizations
- University of Illinois Urbana–Champaign