An Answer to the Bose-Nelson Sorting Problem for 11 and 12 Channels
Jannis Harder

TL;DR
This paper establishes the minimal size of optimal sorting networks for 11 and 12 channels, confirming the best known networks are indeed optimal and closing open problems in the Bose-Nelson sorting problem.
Contribution
It generalizes Van Voorhis's result to a broader class of comparator networks and introduces a verified dynamic programming method to compute optimal sorting network sizes.
Findings
11-channel sorting networks require at least 35 comparators
12-channel sorting networks require at least 39 comparators
The bounds confirm the optimality of known networks
Abstract
We show that 11-channel sorting networks have at least 35 comparators and that 12-channel sorting networks have at least 39 comparators. This positively settles the optimality of the corresponding sorting networks given in The Art of Computer Programming vol. 3 and closes the two smallest open instances of the Bose-Nelson sorting problem. We obtain these bounds by generalizing a result of Van Voorhis from sorting networks to a more general class of comparator networks. From this we derive a dynamic programming algorithm that computes the optimal size for a sorting network with a given number of channels. From an execution of this algorithm we construct a certificate containing a derivation of the corresponding lower size bound, which we check using a program formally verified using the Isabelle/HOL proof assistant.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Code & Models
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
Topicssemigroups and automata theory · graph theory and CDMA systems · Algorithms and Data Compression
