P-complete

Two Player Game (GAME)

Category: Games

Input

A two player game $G = (P_1, P_2, W_0, s, M)$.

Problem

Is $s$ a winning position for the first player?

Definitions

In the game, $P_1 \cap P_2 = \varnothing$, $W_0 \subseteq P_1 \cup P_2$, $s \in P_1$, and $M \subseteq P_1 \times P_2 \cup P_2 \times P_1$. $P_i$ is the set of positions in which it is player $i$’s turn to move. $W_0$ is the set of immediate winning positions (defined below) for player one, and $s$ is the starting position. $M$ is the set of allowable moves; if $(p, q) \in M$ and $p \in P_1$ (or $P_2$) then player one (respectively, or two) may move from position $p$ to position $q$ in a single step. A position $x$ is winning for player one if and only if $x \in W_0$, or $x \in P_1$ and $(x, y) \in M$ for some winning position $y$, or $x \in P_2$ and $y$ is winning for every move $(x, y)$ in $M$.

Status

$\P$-complete (Jones and Laaser [1], Greenlaw, Hoover, and Ruzzo [3]).

Remarks

The reduction to show hardness is from GEN [1] or AM2CVP [3]. Since GAME is an instance of AND/OR Graph Solvability, it follows that determining whether an AND/OR graph has a solution is also $\P$-complete (Kasif [2]).

References

  1. N. D. Jones and W. T. Laaser (1976). Complete problems for deterministic polynomial time.
  2. S. Kasif (1986). On the parallel complexity of discrete relaxation in constraint satisfaction networks.
  3. Greenlaw, R., Hoover, H. J., and Ruzzo, W. L. (1995). Limits to Parallel Computation: P-Completeness Theory. Oxford University Press. Problem A.11.1.