Global search algorithms for finding optimal tensor network contraction sequences.
See our paper https://arxiv.org/abs/2001.08063 for more details on the algorithms and results for square and random tensor networks. And let me know if you have comments or questions!
-
Edit fn_tensors.py to add your own tensor networks. Implemented: square tensor networks, random graphs.
-
Edit and run rn_annealing.py for Simulated Annealing.
-
Edit and run rn_genetic.py for the Genetic Algorithm.
-
Edit and run rn_greedy.py for the Greedy Algorithm.
Evaluate algorithm performance with eval_sequences.py and eval_history.py.