- A simple genetic algorithm for the Travelling salesman problem.
- < 100 lines of code
- Achieves reasonable performance for 25, 50, 100 cities
- Optimal solution isn't guaranteed as the algorithm can get stuck in local minima / be slow to converge, but empirically was found to work well
-
Convergence is fast
-
Snapshot of each generation (with freeze on finish)
-
Convergence is fast
-
Snapshot of every 10th generation (with freeze on finish)
-
Convergence is reasonable
-
Snapshot of every 10th generation (with freeze on finish)
