Fluid Invasion (FLUID)
Input
A graph $G = (V, E)$, a source $s$ and a sink $s’$, a distinguished vertex $u$, a time $t$, nonnegative real conductances $k_{ij}$ for each edge $ij$, and nonnegative real capacities $\phi_l$ for each vertex $l \in V$ .
Problem
Is vertex $u$ filled by the invading fluid at time $t$ according to the fluid invasion algorithm?
Definitions
Informally, the fluid invasion algorithm proceeds by adding vertices to the cluster one at a time depending on when they fill up with the fluid being inserted at the source $s$. The cluster consists of those vertices that have been filled by the invading fluid. The evolution of the fluid configuration is governed by Darcy’s law. Refer to Machta [1] for additional details of the fluid model.
Status
$\P$-complete (Machta [1]).
Remarks
The reduction to show hardness is from NORCVP.
Fluid Invasion remains $\P$-complete when $G$ is restricted to being a $l \times l$ two dimensional lattice in which $s$ ($s’$) is connected to all sites on one face (respectively, the opposite face) of the lattice and with all of the conductances equal to one [1]. The proof that this restricted version is $\P$-complete is complicated and does not seem to follow easily from the result for FLUID. Additional schemes designed to model pattern formation processes are studied from a computational complexity point of view in Greenlaw and Machta [2][3]. There they provide $\NC$ algorithms to compute percolation clusters for three different models — invasion percolation, invasion percolation with trapping, and ordinary percolation.
References
- J. Machta (1993). The computational complexity of pattern formation.
- R. Greenlaw and J. Machta (1994). The parallel complexity of Eden growth, solid-on-solid growth and ballistic deposition.
- R. Greenlaw and J. Machta (1994). The parallel complexity of algorithms for pattern formation models.
- Greenlaw, R., Hoover, H. J., and Ruzzo, W. L. (1995). Limits to Parallel Computation: P-Completeness Theory. Oxford University Press. Problem A.12.4.