Austin Buchanan
austinlbuchanan.bsky.social
Austin Buchanan
@austinlbuchanan.bsky.social
Associate Professor of Industrial Engineering & Management. Interests in #orms #networks #districting. Senior Member of INFORMS. Blogs at https://farkasdilemma.wordpress.com/ Website: https://austinlbuchanan.github.io/
Congrats to my former student Maral for winning the Best Paper Award from the @informs.bsky.social Section on Telecommunications and Network Analytics! #orms
October 26, 2025 at 11:47 PM
I have arrived at the @informs.bsky.social Annual Meeting. #orms Believe it or not, I was here for the Olympics back in 1996 too. I took a photo of Charles Barkley on a police horse using my kid camera
October 25, 2025 at 12:19 AM
Finally, we conduct a case study for Tennessee's state house districts and find that the plaintiffs' case in Wygant v. Lee could have been even stronger if our optimization methods had assisted in the drawing of demonstration plans.
October 13, 2025 at 2:36 PM
Our approach provides easy-to-understand optimality proofs suitable for courts and laypeople. Specifically, it produces a set family with the property that at least one county from each set must be split. This is depicted as a county-level map in which a curve encircles each set.
October 13, 2025 at 2:35 PM
The main problem decides which counties to keep whole, and the subproblem coarsens the selected counties and then seeks a feasible plan. We apply the approach to all congressional and legislative instances across the USA, generating plans that are provably optimal.
October 13, 2025 at 2:32 PM
To answer this question, we propose integer programming techniques based on combinatorial Benders decomposition.
October 13, 2025 at 2:30 PM
[new paper] Political districting to maximize whole counties, coauthored with Maral Shahmizad

github.com/maralshahmiz...
October 13, 2025 at 2:23 PM
For posterity—I later realized that a lower case v works better 😅
October 13, 2025 at 11:31 AM
Now that the kid’s asleep, dad gets to play with the birthday gifts #orms
October 13, 2025 at 12:42 AM
For those with LaTeX allergies, try Typst?
October 10, 2025 at 3:13 PM
Don’t know why the link isn’t working but this is the episode
August 21, 2025 at 2:02 AM
Wondering if they got their inspiration from this Bob’s Burgers episode.. bobs-burgers.fandom.com/wiki/P.T.A._...
August 21, 2025 at 1:58 AM
This part stuck out too
July 27, 2025 at 3:35 PM
Texas was already arguably the biggest Republican gerrymander in the nation, according to this PNAS paper by @chriskenny.bsky.social @corymccartan.com @simko.bsky.social @shirokuriwaki.bsky.social www.pnas.org/doi/abs/10.1...
July 20, 2025 at 12:22 AM
This summer’s beach read from @longformmath.bsky.social
May 24, 2025 at 6:37 PM
The complaint mentions prioritizing least change over “Wisconsin’s traditional redistricting criteria of minimizing county splits”
May 12, 2025 at 1:19 PM
The deadline for the George Nicholson Student Paper Competition is June 2.
#orms @informs.bsky.social

www.informs.org/Recognizing-...
May 6, 2025 at 6:26 PM
Our methods find legislative plans for several southern states that fare better than the enacted maps w.r.t. # majority-minority districts, while splitting counties less (sometimes half as much) and being much more compact (sometimes twice the average Polsby-Popper score).
March 24, 2025 at 12:46 PM
The fires in Oklahoma today led to several detours while getting home #okwx
March 14, 2025 at 10:04 PM
The latest newsletter from the INFORMS Computing Society is out. It includes a message from new chair @thserra.bsky.social. And, a spotlight of districting work by Hamid, Eugene, and me higherlogicdownload.s3.amazonaws.com/INFORMS/a4f8...
January 31, 2025 at 12:30 PM
Took a break from the holidays to do research 🤓
December 27, 2024 at 10:23 PM
The paper ends with some open questions relating to parameterized complexity, the polynomial hierarchy, and average case analysis. Any #tcs #cstheory folks wanna jump in?
November 27, 2024 at 2:51 PM
For smaller TSPLIB instances, our algorithm finds the minimum number of SECs in a second or two. We also find that HardTSPLIB instances, which are hard for state-of-the-art approaches like Concorde, require substantially more SECs
November 27, 2024 at 2:11 PM
The TSP formulation of Dantzig, Fulkerson, and Johnson works so well despite having exponentially many subtour elimination constraints (SEC). Why? Researchers have observed that only a handful of SECs are needed in practice. Can this be exploited?
November 27, 2024 at 2:03 PM
So far, I’m getting more activity on BlueSky than Twitter, despite a much smaller follower count. BlueSky looks to be the way forward.
November 17, 2024 at 1:25 PM