• Corpus ID: 238259058

A Recursive Algorithm for Solving Simple Stochastic Games

@article{Montjoye2021ARA,
  title={A Recursive Algorithm for Solving Simple Stochastic Games},
  author={X. Badin de Montjoye},
  journal={ArXiv},
  year={2021},
  volume={abs/2110.01030}
}
We present two recursive strategy improvement algorithms for solving simple stochastic games. First we present an algorithm for solving SSGs of degree d that uses at most O (⌊ (d+ 1)/2 ⌋n/2) iterations, with n the number of MAX vertices. Then, we focus on binary SSG and propose an algorithm that has complexity O (φPoly(N)) where φ = (1 + √ 5)/2 is the golden ratio. To the best of our knowledge, this is the first deterministic strategy improvement algorithm that visits 2 strategies with c < 1… 

Figures from this paper

References

SHOWING 1-10 OF 14 REFERENCES
On Algorithms for Simple Stochastic Games
  • A. Condon
  • Computer Science, Mathematics
    Advances In Computational Complexity Theory
  • 1990
TLDR
It is shown that four natural approaches to solving the simple stochastic game problem are incorrect, and two new algorithms for the problem are presented, one of which extends a technique of Shapley called the successive approximation technique by using linear programming to maximize the improvement at each approximation step.
Solving simple stochastic games with few coin toss positions
TLDR
It is shown that an algorithm combining value iteration with retrograde analysis achieves a time bound of O(r 2r (r logr+n)), thus improving both time bounds.
Simple Stochastic Games with Few Random Vertices Are Easy to Solve
TLDR
This algorithm is based on an exhaustive search of a special kind of positional optimal strategies, the f-strategies, and performs well on SSGs with few random vertices, does not rely on linear or quadratic programming and applies to all SSGs.
Solving Simple Stochastic Games with few Random Nodes faster using Bland's Rule
TLDR
A simpler iterative variant of Ludwig's randomized algorithm, using Bland's rule from the simplex algorithm, which uses exponentially less random bits than Ludwig's version, and is adapted to the algorithm of Gimbert and Horn whose worst case complexity is O(k!).
On Strategy Improvement Algorithms for Simple Stochastic Games
TLDR
The Hoffman-Karp algorithm converges to optimal strategies of a given SSG, but no nontrivial bounds were previously known on its running time, and a bound of O(2n/n) on the convergence time of this algorithm is shown, and these are the first non-trivial upper bounds on the converge time of these strategy improvement algorithms.
On the Hardness of Simple Stochastic Games
We examine the problem of finding optimal strategies in simple stochastic games, and the equivalent problem of finding stable configurations of MIN/MAX/AVG circuits. The problem can be seen as a
Stochastic Games*
  • L. Shapley
  • Mathematics
    Proceedings of the National Academy of Sciences
  • 1953
TLDR
In a stochastic game the play proceeds by steps from position to position, according to transition probabilities controlled jointly by the two players, and the expected total gain or loss is bounded by M, which depends on N 2 + N matrices.
The Complexity of Stochastic Games
A Generic Strategy Improvement Method for Simple Stochastic Games
TLDR
A generic strategy improvement algorithm (GSIA) is presented to find an optimal strategy of simple stochastic games (SSG) and a tight bound on the denominator of the values associated to a strategy is proved to prove that all strategy improvement algorithms are in fact fixed parameter tractable in the number of random vertices.
...
...