Publication record · 18.cifr/1983.kirkpatrick.simulated-annealing

Optimization by Simulated Annealing

v1.0.0

S. Kirkpatrick (IBM Thomas J. Watson Research Center), C. D. Gelatt (IBM Thomas J. Watson Research Center), M. P. Vecchi (Instituto Venezolano de Investigaciones Cientificas)

RAI18.cifr/1983.kirkpatrick.simulated-annealing

Science· 1983· doi:10.1126/science.220.4598.671

There is a deep and useful connection between statistical mechanics (the behavior of systems with many degrees of freedom in thermal equilibrium at a finite temperature) and multivariate or combinatorial optimization (finding the minimum of a given function depending on many parameters). A detailed analogy with annealing in solids provides a framework for optimization of the properties of very large and complex systems.

simulated annealingcombinatorial optimizationstatistical mechanicsMetropolis algorithmtraveling salesman problem

✦ Research context

What this agent contributes to the literature.

Problem solved

Combinatorial optimization problems have exponentially large solution spaces where greedy local search reliably gets trapped in local minima. Prior heuristics had no principled mechanism to escape these traps. Simulated annealing provides a polynomial-time stochastic method with a cooling schedule that gradually focuses the search toward the global optimum.

Novelty

Kirkpatrick et al. introduced the formal analogy between combinatorial optimization and statistical mechanics of annealing, applying the Metropolis algorithm as a general-purpose optimizer. The key innovation is accepting worse solutions with probability exp(-dE/kT), enabling principled escape from local minima. This was the first systematic treatment connecting physical annealing schedules to NP-hard combinatorial problems.

Related research

Computing related research...

Canvas contract1-in / 1-out · unpacked into problem, annealing_params legacy ports

Sample data

Loading sample data...

Total calls

This month

Citations

Last called

—

Image digest

sha256:4bed5b74ae3dc6d8d21f4fa88e81c40126733f5398d677742a7e775a7cd76c5a

Invoke command

python main.py

Inputs

input:application/json

Outputs

output:application/json

Citation

Loading DOI…

Invoke

CPU compute only

How to get GPU access: Your university, lab, or company can become a CIFR institutional member. Members get GPU-accelerated runs for all their researchers. Contact us

Recent invocations(0)

No invocations yet — be the first to call this agent.