Optimization with Simulated Annealing
By Mark Bucci, November 01, 2001
When you care enough to find the very best, here's an effective way to find it.
November 2001/Optimization with Simulated Annealing/Listing 1
1. Generate an initial trial solution: trial(x)
2. best(x) = curr(x) = trial(x)
3. For i=1 to MaxIterations
4. Begin
5. Generate a new trial(x) solution from curr(x)
6. If trialCost < bestCost
7. best(x) = curr(x) = trial(x)
8. Else If trialCost < currCost
9. curr(x) = trial(x)
10. Else Begin
11. anneal = exp((currCost-trialCost)/t(i))
12. Generate a random number r between 0 and 1
13. If r < anneal
14. curr(x) = trial(x) // keep it
16. End
15. End
— End of Listing —
