Simulated Annealing in C#
The simulated annealing is a metaheuristic, a random search algorithm inspired from physics sciences. In this example, we will doing a simple thing : adjusting one coefficent for having a better results for the algorithm to found the global minimum of the function :
f(X)=0.11(0.5 X² + Cos[Pi X] - 2 Sin[2 Pi X] + Cos[3 Pi X]*Sin[Pi X] in [-5; 5].
// {...}
// COEF is the Most important value to adjust.
// energy = COEF x myFunction(x)
// Change this value until having a quality results for Simulated Annealing.
COEF = 1; // <- Try with 1, 10, 100, 1000.....
// After running the playground :
// When ct0 =~ ct1 or ct2 =~ ct3, its a bad adjustement.
// A possible good situation is when ct2 =~ 2 x ct1 and ct3 =~ K x ct2 where K depends of number of local optimum.
// The number of simulations is now fixed to 200.
// Next time we will see how to change for better precisions.
// {...}
Join the CodinGame community on Discord to chat about puzzle contributions, challenges, streams, blog articles - all that good stuff!
JOIN US ON DISCORD Online Participants