The Complexity of the Comparator Circuit Value Problem

In 1990 Subramanian defined the complexity class CC as the set of problems log-space reducible to the comparator circuit value problem (CCV). He and Mayr showed that NL \subseteq CC \subseteq P, and proved that in addition to CCV several other problems are complete for CC, including the stable marriage problem, and finding the lexicographically first maximal matching in a bipartite graph. We are interested in CC because we conjecture that it is incomparable with the parallel class NC which also satisfies NL \subseteq NC \subseteq P, and note that this conjecture implies that none of the CC-complete problems has an efficient polylog time parallel algorithm. We provide evidence for our conjecture by giving oracle settings in which relativized CC and relativized NC are incomparable. We give several alternative definitions of CC, including (among others) the class of problems computed by uniform polynomial-size families of comparator circuits supplied with copies of the input and its negation, the class of problems AC^0-reducible to CCV, and the class of problems computed by uniform AC^0 circuits with CCV gates. We also give a machine model for CC, which corresponds to its characterization as log-space uniform polynomial-size families of comparator circuits. These various characterizations show that CC is a robust class. The main technical tool we employ is universal comparator circuits. Other results include a simpler proof of NL \subseteq CC, and an explanation of the relation between the Gale-Shapley algorithm and Subramanian’s algorithm for stable marriage. This paper continues the previous work of Cook, L^e and Ye which focused on Cook-Nguyen style uniform proof complexity, answering several open questions raised in that paper.

💡 Research Summary

The paper conducts a thorough investigation of the complexity class CC, which is defined as the set of problems log‑space reducible to the Comparator Circuit Value problem (CCV). It begins by recalling Subramanian’s original definition of CC and the subsequent result by Subramanian and Mayr that NL ⊆ CC ⊆ P. The authors’ primary motivation is the conjecture that CC is incomparable with the parallel class NC, another class satisfying NL ⊆ NC ⊆ P. If true, none of the CC‑complete problems would admit efficient polylogarithmic‑time parallel algorithms.

To support this conjecture, the authors provide two main lines of evidence. First, they construct relativized worlds using oracles. They show that there exists an oracle A for which relativized NC^A does not contain relativized CC^A, and there exists an oracle B for which relativized CC^B does not contain relativized NC^B. These separations demonstrate that, under standard relativization techniques, the two classes can be made incomparable, lending credence to the conjecture that they are truly distinct in the unrelativized world.

Second, the paper offers a suite of alternative, equivalent characterizations of CC, establishing its robustness. The characterizations include:

1. Uniform polynomial‑size families of comparator circuits that receive both the input and its bitwise negation as copies.
2. AC⁰‑reducibility to CCV, i.e., the class of problems that can be reduced to CCV by constant‑depth, polynomial‑size circuits with unbounded fan‑in.
3. AC⁰ circuits augmented with CCV gates, showing that a constant‑depth circuit that may call a CCV subroutine can compute exactly the problems in CC.
4. A log‑space uniform machine model where the basic instruction set consists of comparator operations.

The central technical tool enabling these equivalences is the construction of universal comparator circuits. A universal comparator circuit can simulate any comparator circuit of a given size by taking a description of the target circuit as part of its input. This construction mirrors the classic universal Turing machine argument but is tailored to the restricted comparator gate set. It allows the authors to translate between circuit families, reductions, and machine models seamlessly.

The paper also revisits the inclusion NL ⊆ CC, providing a streamlined proof. By encoding the nondeterministic log‑space transition relation as a sequence of comparator operations, the authors show that any NL computation can be collapsed into a single comparator circuit of polynomial size. This proof is more direct than earlier arguments and highlights the natural fit between nondeterministic log‑space branching and the “compare‑and‑swap” primitive of comparator circuits.

An additional contribution is the clarification of the relationship between the classic Gale‑Shapley algorithm for the stable marriage problem and Subramanian’s algorithm that reduces stable marriage to CCV. The authors demonstrate that both algorithms implicitly construct the same comparator network: each proposal–acceptance step corresponds to a compare‑and‑swap that moves a “more preferred” partner upward in a ranking. Consequently, the stable marriage problem’s status as CC‑complete follows naturally from this structural equivalence.

Beyond these core results, the authors address several open questions raised in earlier work by Cook, Lê, and Ye on Cook‑Nguyen style uniform proof complexity. They show that CCV gates can be embedded in AC⁰ circuits without breaking log‑space uniformity, thereby linking CC to proof‑complexity frameworks. They also discuss how the universal comparator construction can be leveraged to design new CC‑based algorithms and to explore potential separations between CC and other well‑studied classes such as P‑complete or L.

In summary, the paper establishes CC as a robust, well‑characterized class through multiple equivalent definitions, introduces universal comparator circuits as a versatile technical device, provides new evidence that CC and NC are likely incomparable, and deepens the understanding of several CC‑complete problems—including stable marriage and lexicographically first maximal matching—by connecting them to concrete comparator‑network constructions. These contributions lay a solid foundation for future work on the structural properties of CC, its relationship to parallel computation, and its role in proof complexity.