P-complete

Two-person Pebble Game Fixed Rank (PEBBLE)

Category: Games

Input

A fixed rank pebble game $G = (X, R, S, t)$.

Problem

Does player $P_1$ have a winning strategy?

Definitions

A pebble game $G = (X, R, S, t)$ has the following properties:

  1. $X$ is a finite set of vertices.
  2. $R \subseteq \{(x, y, z) \mid x, y, z \in X, x \neq y, y \neq z, z \neq x\}$ is called the set of rules. For $A, B \subseteq X$, we write $A \vdash B$ if $(x, y, z) \in R$, $x, y \in A$, $z \notin A$, and $B = (A \setminus \{x\}) \cup \{z\}$. The move $A \vdash B$ is made by the rule $(x, y, z)$. The symbol $\vdash^\ast$ denotes the reflexive and transitive closure of $\vdash$.
  3. $S$ is a subset of $X$.
  4. $t \in X$ is the terminal vertex.

A pebble game has fixed rank if the number of vertices in $S$ is fixed. In a two-person pebble game two players $P_1$ and $P_2$ alternate moving pebbles with $P_1$ moving first. The winner is the first player who can pebble $t$ or who can force the other player into a no move situation.

Status

$\P$-complete (Kasai, Adachi, and Iwata [1]).

Remarks

The reduction to show hardness is from a logarithmic space bounded alternating Turing machine.

A pebble game is solvable if there exists $A \subseteq X$ such that $S \vdash^\ast A$ and $t \in A$. The one-person pebble game, which is to determine if the game is solvable, of fixed ranked is $\mathsf{NLOG}$-complete [1]. For questions with a similar flavor see GAME and CM. For additional remarks about pebbling see [2, Section 5.4].

References

  1. T. Kasai, A. Adachi, and S. Iwata (1990). Classes of pebble games and complete problems.
  2. Greenlaw, R., Hoover, H. J., and Ruzzo, W. L. (1995). Limits to Parallel Computation: P-Completeness Theory. Oxford University Press. Problem A.11.7.