| Article ID: | iaor20041105 |
| Country: | Germany |
| Volume: | 95 |
| Issue: | 3 |
| Start Page Number: | 512 |
| End Page Number: | 532 |
| Publication Date: | Jan 2003 |
| Journal: | Mathematical Programming |
| Authors: | Raghavan T.E.S., Syed Z. |
We give a policy-improvement type algorithm to locate an optimal pure stationary strategy for discounted stochastic games with perfect information. A graph theoretic motivation for our algorithm is presented as well.