P-complete

Relation Coarsest Set Partition (RCSP)

Input

A finite set $S$, an initial partition $\pi_0$ of $S$, a designated set $S’ \subseteq S$, and a set of binary relations $R_1, \ldots, R_k$ on $S$.

Problem

Is $S’$ a member of the coarsest refinement $\pi$ of $\pi_0$ such that for each pair of blocks $B_1$ and $B_2$ of $\pi$, and for each integer $i = 1, \ldots, k$ either $$B_1 \subseteq PI_i(B_2) \text{ or } B_1 \cap PI_i(B_2) = \varnothing,$$ where $PI_i(B_2)$ is the pre-image set $PI_i(B_2) = \{x \mid \text{there is a } y \in B_2 \text{ with } x R_i y\}$.

Status

$\P$-complete (Sang Cho and Huynh [1]).

Remarks

In [1], they observe that Kanellakis and Smolka [2] showed the $\NC^1$ equivalence of RCSP to a $\P$-complete variant of SBTS.

Even if there is only a single relation, $k$ equals one, the problem is $\P$-complete [1]. Sang Cho and Huynh give a $\log n$ time, $O(n^2)$ processor CREW-PRAM algorithm for the single function Coarsest Set Partition Problem as well as a $\log n$ time, $O(n^3)$ processor EREW-PRAM algorithm for the same problem [1]. See also the remarks for SBTS and SBDTS.

References

  1. Sang Cho and D. T. Huynh (1992). The parallel complexity of coarsest set partition problems.
  2. P. C. Kanellakis and S. A. Smolka (1990). CCS expressions, finite state processes, and three problems of equivalence.
  3. Greenlaw, R., Hoover, H. J., and Ruzzo, W. L. (1995). Limits to Parallel Computation: P-Completeness Theory. Oxford University Press. Problem A.7.14.