Traveling Salesman Problem on Implosion TSP-1000
0Optimality Gap (Initial)(Near-)Optimality
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| (Near-)OptimalityFine-tuning split (FT)=TSP-1002025.01 | 0 | — | — | 17.5 | |
| INViT-1V+ASAP w/ MAMLPolicy=INViT-1V+ASAP w/ MAML, Fine-tuning split (FT)=TSP-1002025.01 | 9.67 | 6.99 | 2.68 | 5.3 | |
| INViT-1V+ASAP w/ MAML(Tune All)Policy=INViT-1V+ASAP w/ MAML(Tune All), Fine-tuning split (FT)=TSP-1002025.01 | 9.67 | 7.53 | 2.14 | 5.3 | |
| INViT-1V+ASAP w/o MAMLPolicy=INViT-1V+ASAP w/o MAML, Fine-tuning split (FT)=TSP-1002025.01 | 10 | 8.11 | 1.89 | 5.3 | |
| INViT-1V+MAMLPolicy=INViT-1V+MAML, Fine-tuning split (FT)=TSP-1002025.01 | 11.27 | 9.75 | 1.52 | 4.5 | |
| INViT-1VPolicy=INViT-1V, Fine-tuning split (FT)=TSP-1002025.01 | 13.18 | 11.98 | 1.2 | 4.5 | |
| POMO+ASAP w/ MAMLPolicy=POMO+ASAP w/ MAML, Fine-tuning split (FT)=TSP-1002025.01 | 56.73 | 52.11 | 4.62 | 2.3 | |
| POMO+ASAP w/ MAML(Tune All)Policy=POMO+ASAP w/ MAML(Tune All), Fine-tuning split (FT)=TSP-1002025.01 | 56.73 | 53.32 | 3.41 | 2.3 | |
| POMO+ASAP w/o MAMLPolicy=POMO+ASAP w/o MAML, Fine-tuning split (FT)=TSP-1002025.01 | 60.11 | 57.96 | 2.15 | 2.3 | |
| POMO+MAMLPolicy=POMO+MAML, Fine-tuning split (FT)=TSP-1002025.01 | 65.74 | 64.13 | 1.61 | 2.3 | |
| POMOPolicy=POMO, Fine-tuning split (FT)=TSP-1002025.01 | 69.93 | 74.23 | -4.3 | 2.3 |