Note
Six metaheuristics in Python, with statistical calibration
Six algorithms, each on a classic problem, with parameters calibrated statistically rather than by hand; public repository.
| Organisation | Universidad de los Andes, graduate course in metaheuristics |
|---|---|
| Sector | Combinatorial optimization |
| Period | 2020 |
| Role | Author |
| Stack | Python |
| Links | github.com/robayo/metaheuristics |
Algorithms
| Algorithm | Problem |
|---|---|
| Path scanning | Capacitated arc routing (CARP) |
| Local search | Asymmetric travelling salesman (ATSP) |
| Simulated annealing | n-queens |
| Genetic algorithm | Capacitated vehicle routing (CVRP) |
| Iterated local search | Multidimensional knapsack (MKP) |
| Large neighbourhood search | Pickup and delivery with time windows (PDPTW) |
Calibration
Each algorithm's parameters were calibrated with designed experiments and statistical tests on benchmark instances, not tuned by hand. The repository is public; a README with a table of gaps against best-known solutions is in progress.