首页 > 代码库 > (转)Fibonacci Tilings
(转)Fibonacci Tilings
Fibonacci numbers {Fn, n ≥ 0} satisfy the recurrence relation
along with the initial conditions F1 = 1 and F0 = 0.
The Fibonacci name has been attached to the sequence 0, 1, 1, 2, 3, 5, ... due to the inclusion in his 1202 book Liber Abaci of a rabbit reproduction puzzle: under certain constraints the rabbit population at discrete times is given exactly by that sequence. As naturally, the sequence is simulated by counting the tilings with dominoes of a 2×n board:
A tiling of a 2×n board may end with two horizontal dominoes or a single vertical domino:
In the former case, it‘s an extension of a tiling of a 2×(n-2) board; in the latter case, it‘s an extension of a tiling of a 2×(n-1). If Tn denotes the number of domino tilings of a 2×n board, then clearly
Tn = Tn-2 + Tn-1
which is the same recurrence relation that is satisfied by the Fibonacci sequence. By a direct verification, T1 = 1, T2 = 2, T3 = 3, T4 = 5, etc., which shows that {Tn} is nothing but a shifted Fibonacci sequence. If we define, T0 = 1, as there is only 1 way to do nothing; and T-1 = 0, because there are no boards with negative side lengths, then Fn = Tn-1, for n ≥ 0.
The domino tilings are extensively used in Graham, Knuth, Patashnik and by Zeitz. Benjamin & Quinn economize by considering only an upper 1×n portion of the board (and its tilings). This means tiling a 1×n board with 1×1 and 1×2 pieces.
I‘ll use Benjamin & Quinn‘s frugal tilings to prove Cassini‘s Identity
Fn+1·Fn+1 - Fn·Fn+2 = (-1)n
In terms of the tilings, I want to prove that Tn·Tn - Tn-1·Tn+1 = (-1)n.
The meaning of the term Tn·Tn is obvious: this is the number of ways to tile two 1×n boards where the tilings of the two boards are independent of each other. Similarly, Tn-1Tn+1 is the number of ways to tile two boards: one 1×(n-1) and one 1×(n+1). Now, the task is to retrieve the relation between the two numbers annunciated by Cassini‘s identity.
Our setup consists of two 1×n boards:
with the bottom board shifted one square to the right:
The tilings of the two boards may or may not have a fault line. A fault line is a line on the two boards at which the two tilings are breakable. For example, the tilings below have three fault lines:
The trick is now to swap tails: the pieces of the two tilings (along with the boards) after the last fault line:
Since the bottom board has been shifted just one square, the swap produces one tiling of a 1×(n+1) - the top board in the diagram - and one tiling of a 1×(n-1) board - the bottom board in the diagram. Note that the old faults have been preserved and no new faults have been introduced.
Thus, in the presence of faults, there is a 1-1 correspondence between two n-tilings (Tn) and a pair of (n-1)- and (n+1)-tilings. The time is to account for the faultless combinations, if any.
But there are. Any 1×1 square induces a fault. This leaves exactly two faultless tilings. If n is odd, both n-1 and n+1 are even, there is a unique pair of (n-1)- and (n+1)-tilings:
If n is even, there is a unique n-tiling that, when shifted, generates no fault lines:
References
- A. T. Benjamin, J. J. Quinn, Proofs That Really Count: The Art of Combinatorial Proof, MAA, 2003
- R. Graham, D. Knuth, O. Patashnik, Concrete Mathematics, 2nd edition, Addison-Wesley, 1994.
- P. Zeitz, The Art and Craft of Problem Solving, John Wiley & Sons, 1999
Related material | |
| |
| |
| |
| |
| |
| |
| |
| |
|Contact| |Front page| |Contents| |Algebra| |Store|
Copyright ? 1996-2011 Alexander Bogomolny
本文转自:
http://www.cut-the-knot.org/arithmetic/combinatorics/FibonacciTilings.shtml