×

Minimal expansions in redundant number systems: Fibonacci bases and greedy algorithms. (English) Zbl 1072.11007

Let \(\{G_j\}\) denote a sequence of integers. The cost of the (redundant) digit expansion \(n= \sum_{j\geq 0}\varepsilon_j G_j, \varepsilon_j\in {\mathbb Z}\) is defined as \[ c({\mathbf \varepsilon}(n))= \sum_{j\geq 0}| \varepsilon_j|. \] In the present paper minimal expansions with respect to two base sequences are examined.
In part one \(G_j=F_j\), where \(F_j\) denotes the \(j\)th term of the Fibonacci sequence. A sequence \({\mathbf \varepsilon}=(\dots,\varepsilon_2,\varepsilon_1)\in \{0,\pm 1\}^{\mathbb N}\) with finitely many nonzero elements is called admissible, if \(\varepsilon_1=0\) and none of the patterns \((-1,1),(1,1),(-1,0,1),(1,0,1),(1,0,0,1)\) or their negatives appear as subsequences of \(\mathbf \varepsilon\). An admissible sequence \({\mathbf\varepsilon}\) is called an admissible expansion of an integer \(n\), if \(n=\sum_{j\geq 1}\varepsilon_j F_j\). Then there is a unique admissible expansion of every integer \(n\). Moreover it is minimal with respect to the cost function and can be calculated by a greedy algorithm and it cannot be calculated by a deterministic transducer from right to left. However, it is possible to calculate some minimal expansion from right to left.
In part two some result about the Fibonacci expansion is ported back to the \(q\)-ary case, i.e. when \(G_j=q^j\) for a fixed integer \(q\geq 2\). It is proved that the symmetric signed digit expansion can be computed by a greedy algorithm.

MSC:

11A63 Radix representation; digital problems
11B39 Fibonacci and Lucas numbers and polynomials and generalizations
Full Text: DOI

References:

[1] A. D. Booth, A signed binary multiplication technique, Quart. J. Mech. Appl. Math. 4 (1951), 236-240, MR 12,860a. · Zbl 0043.12902 · doi:10.1093/qjmam/4.2.236
[2] P. Flajolet and L. Ramshaw, A note on Gray code and odd-even merge, SIAM J. Comput. 9 (1980), 142-158, MR 81h:68025. · Zbl 0447.68083 · doi:10.1137/0209014
[3] P. Grabner, C. Heuberger and H. Prodinger, Subblock occurrences in signed digit representations, Glasg. Math. J. 45 (2003), 427-440. · Zbl 1049.11084 · doi:10.1017/S0017089503001368
[4] R. L. Graham, D. E. Knuth and O. Patashnik, Concrete mathematics ? A foundation for computer science, (2nd edn.), Addison-Wesley, 1994, MR 97d:68003. · Zbl 0836.00001
[5] C. Heuberger, Minimal redundant digit expansions in the Gaussian integers, J. Th?or. Nombres Bordeaux 14 (2002), 517-528. · Zbl 1076.11005
[6] C. Heuberger and H. Prodinger, On minimal expansions in redundant number systems: Algorithms and quantitative analysis, Computing 66 (2001), 377-393. · Zbl 1030.11003 · doi:10.1007/s006070170021
[7] C. Heuberger and H. Prodinger, Carry propagation in signed digit representations, European J. Combin. 24 (2003), 293-320. · Zbl 1026.11015 · doi:10.1016/S0195-6698(03)00008-8
[8] M. Joye and S.-M. Yen, Optimal left-to-right binary signed digit recoding, IEEE Transactions on Computers 49, No.7 (2000), 740-748. · Zbl 1300.94069 · doi:10.1109/12.863044
[9] J. L. Massey and O. N. Garcia, Error-correcting codes in computer arithmetic, Advances Inform. Syst. Sci. 4 (1972), 273-326. · Zbl 0256.94019
[10] F. Morain and J. Olivos, Speeding up the computations on an elliptic curve using addition-subtraction chains, RAIRO Inform. Th?or. Appl. 24 (1990), 531-543, MR 91i:11189. · Zbl 0724.11068
[11] G. W. Reitwiesner, Binary arithmetic, Advances in computers, Vol. 1, Academic Press, New York, 1960, 231-308.
[12] E. Zeckendorf, Repr?sentation des nombres naturels par une somme de nombres de Fibonacci ou de nombres de Lucas, Bull. Soc. Roy. Sci. Li?ge 41 (1972), 179-182, MR 46 #7147. · Zbl 0252.10011
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.