Fix policy iteration for minimal reach reward properties

moves-rwth / storm

A Modern Probabilistic Model Checker

https://www.stormchecker.org

GNU General Public License v3.0

135 stars 74 forks source link

Closed sjunges closed 4 months ago

sjunges commented 4 months ago

~This PR tracks bughunting for a bug~

Fixes the initial policy computation. The previous implementation did not correctly handle choices to states with infinite reward.

tquatmann commented 4 months ago

I hijacked this PR to push a fix to #555 (I wanted to keep the assertions introduced by @sjunges ).

sjunges commented 4 months ago

Great