An implementation of the Held-Karp dynamic programming algorithm to solve the Traveling Salesman Problem
- Parallelization via OpenMP (see Makefile)
- Linux (e. g.
pacman -S openmp
) and macOS (brew install llvm
) ACTUAL_N
is the number of vertices to incorporate. Adapt yourdist_matrix
accordingly.
On my machines, 28 vertices are processed in 1:20 to 1:30 mins with < 8 GB memory usage.