Log In Sign Up

Strategy Complexity of Reachability in Countable Stochastic 2-Player Games

by   Stefan Kiefer, et al.

We study countably infinite stochastic 2-player games with reachability objectives. Our results provide a complete picture of the memory requirements of ε-optimal (resp. optimal) strategies. These results depend on whether the game graph is infinitely branching and on whether one requires strategies that are uniform (i.e., independent of the start state). Our main result is that ε-optimal (resp. optimal) Maximizer strategies in infinitely branching turn-based reachability games require infinite memory. This holds even under very strong restrictions. Even if all states have an almost surely winning Maximizer strategy, strategies with a step counter plus finite private memory are still useless. Regarding uniformity, we show that for Maximizer there need not exist positional uniformly ε-optimal strategies even in finitely branching turn-based games, whereas there always exists one that uses one bit of public memory, even in concurrent games with finite action sets.


Reachability for Branching Concurrent Stochastic Games

We give polynomial time algorithms for deciding almost-sure and limit-su...

Stochastic Games with Lexicographic Reachability-Safety Objectives

We study turn-based stochastic zero-sum games with lexicographic prefere...

Playing Safe, Ten Years Later

We consider two-player games over graphs and give tight bounds on the me...

Beyond admissibility: Dominance between chains of strategies

Admissible strategies, i.e. those that are not dominated by any other st...

Optimal strategies in concurrent reachability games

We study two-player reachability games on finite graphs. At each state t...

Synthesis of Deceptive Strategies in Reachability Games with Action Misperception

We consider a class of two-player turn-based zero-sum games on graphs wi...

Infinite Separation between General and Chromatic Memory

In this note, we answer a question from [Alexander Kozachinskiy. State C...