Open sawcordwell opened 9 years ago
The _boundIter method causes the algorithm to stop before it has reached the epsilon-optimal value function. Steps to reproduce:
_boundIter
import mdptoolbox.example P, R = mdptoolbox.example.forest() vigs = mdptoolbox.mdp.ValueIterationGS(P, R, 0.96) vigs.setVerbose() vigs.run()
What is expected: _boundIter should compute something that the algorithm can finish before, or it should not be used.
The
_boundIter
method causes the algorithm to stop before it has reached the epsilon-optimal value function. Steps to reproduce:What is expected:
_boundIter
should compute something that the algorithm can finish before, or it should not be used.