Puzzle link: Play on SudokuPad.
Rules: Normal Sudoku rules apply. Digits must not repeat within a cage.
Each cage labeled An represents a 2×2 matrix, with one of its eigenvectors contained in the cage labeled vn and the associated eigenvalue in the cage labeled λn (n = 1,...,4). All eigenvectors and eigenvalues fulfill the eigenvalue equation Anvn = λnvn, which is essentially written in matrix notation in the grid for each n, except that the equality sign and vn on the right-hand side of the equation were omitted.
The multiplication of the matrix and its eigenvector, Anvn, is carried out in the usual way, i.e., the result in the top row is given by the sum of the top left digit of An multiplied with the top digit of vn and the top right digit of An multiplied with the bottom digit of vn. Analogously, the result of Anvn in the bottom row is given by the sum of the bottom left digit of An multiplied with the top digit of vn and the bottom right digit of An multiplied with the bottom digit of vn. For the multiplication λnvn, the result in the top row is λn multiplied with the top digit of vn, and in the bottom row it is λn multiplied with the bottom digit of vn.
The cage labeled v1a contains an alternative eigenvector of A1 with associated eigenvalue λ1a, either of which may be identical to or different from v1 and λ1, respectively.
The cage labeled λ2+λ3+λ4 contains the sum of the three denoted eigenvalues in normal notation with the tens digit on the left and the ones digit on the right.
Example: The eigenvalue equation would be fulfilled for A = (6,2,3,1) with the elements listed as (top left, top right, bottom left, bottom right), v = (8,4) with the elements listed as (top, bottom), and λ = 7 because Av = (6×8+2×4,3×8+1×4) = (56,28), which equals λv = (7×8,7×4) = (56,28).
Your feedback, ratings and comments are highly appreciated. Have fun!
Background: The properties of eigenvalues and eigenvectors are treated in introductory courses of linear algebra, but the puzzle is solvable without knowing that theory. Generally, neither eigenvalues nor eigenvectors need to consist of natural numbers, but the examples here were chosen such that this is the case.
The Schrödinger equation H ψ = E ψ is an example of an eigenvalue equation for which the Hamilton operator H is generally Hermitian, i.e., when written as a complex-valued matrix, it is equal to its conjugate transpose. As a consequence, the eigenvalues E, i.e., the energies of the quantum system, are real-valued. The eigenvectors ψ are the associated wave functions of the system at that energy. My prior puzzle on the Schrödinger equation visualized the fundamental relation of quantum mechanics. Although looking pretty, it does not contain the correct mathematical treatment. The current puzzle remedies this with four examples of eigenvalue equations in matrix notation.
Solution code: All digits of row 5 (from left to right) without spaces.
on 1. August 2026, 21:15 by Arlo Lipof
Very cool, thank you!
----------
Many thanks for playing and commenting! - TB
on 25. July 2026, 03:45 by Gunkan
Thanks fpr a fascinating problem, impossible to rate, but I succeded despite not haveing solved eigenvalueproblems in 25 years.
----------
Many thanks for your feedback. Much appreciated! - TB
on 24. July 2026, 19:47 by EvaristeLogic
Absolutely delightful puzzle :) A bit difficult to estimate a difficulty due to differences in linear algebra background. As a mathematician this very much tickled my fancy. Thank you!
----------
Thanks a lot for playing and commenting. Glad you liked it! - TB