Uploaded June 2020 | Updated November 2026, 3 hours ago
CSP (constraint satisfaction problem) optimization (trying to find good solutions) and certification (trying to find upper bounds on the value of the best solution). Also discussed: some of the best known algorithmic approximability and NP-hardness results, including the 2-to-2 Theorem. Lecture 20c of "CS Theory Toolkit": a semester-long graduate course on math and CS fundamentals for research in theoretical computer science, taught at Carnegie Mellon University.
Resources for this lecture:
Approximability of CSPs: http://www.cs.cmu.edu/~odonnell/slides/csp-approx.pdf
"Polymorphisms, and how to use them" by Barto, Krokhin, and Willard (for the Dichotomy Conjecture/Theorem. This survey was written *just* before the Algebraic Dichotomy Conjecture was proven.)
Taught by Ryan O'Donnell (https://www.cs.cmu.edu/~odonnell)
Course homepage on CMU's Diderot system: https://www.diderot.one/course/28/
Thumbnail photo by Rebecca Kiger (rebeccakphoto.com/)
CSP (constraint satisfaction problem) optimization (trying to find good solutions) and certification (trying to find upper bounds on the value of the best solution). Also discussed: some of the best known algorithmic approximability and NP-hardness results, including the 2-to-2 Theorem. Lecture 20c of "CS Theory Toolkit": a semester-long graduate course on math and CS fundamentals for research in theoretical computer science, taught at Carnegie Mellon University.
Resources for this lecture:
Approximability of CSPs: http://www.cs.cmu.edu/~odonnell/slides/csp-approx.pdf
"Polymorphisms, and how to use them" by Barto, Krokhin, and Willard (for the Dichotomy Conjecture/Theorem. This survey was written *just* before the Algebraic Dichotomy Conjecture was proven.)
Taught by Ryan O'Donnell (https://www.cs.cmu.edu/~odonnell)
Course homepage on CMU's Diderot system: https://www.diderot.one/course/28/
Thumbnail photo by Rebecca Kiger (rebeccakphoto.com/)










