Berlekamp–Massey algorithm: Difference between revisions

Revision as of 20:29, 12 September 2016

Latest revision as of 06:47, 3 May 2025

(42 intermediate revisions by 28 users not shown)

Line 1:

{{Short description|Algorithm on linear-feedback shift registers}}

{{distinguish|Berlekamp's algorithm}}

[[File:Berlekamp–Massey algorithm.png|thumb|right|Berlekamp–Massey algorithm]]

The '''Berlekamp–Massey algorithm''' is an [[algorithm]] that will find the shortest [[linear -feedback shift register]] (LFSR) for a given binary output sequence. The algorithm will also find the [[Minimal polynomial (field theory)|minimal polynomial]] of a linearly [[Recurrence relation|recurrent sequence]] in an arbitrary [[field (mathematics)|field]]. The field requirement means that the Berlekamp–Massey algorithm requires all non-zero elements to have a multiplicative inverse.<ref>{{Harvnb|Reeds|Sloane|1985|p=2}}</ref> Reeds and Sloane offer an extension to handle a [[ring (mathematics)|ring]].<ref>{{Citation |lastlast1=Reeds |firstfirst1=J. A. |last2=Sloane |first2=N. J. A. |authorlink2author-link2=N. J. A. Sloane |journal=SIAM Journal on Computing |volume=14 |issue=3 |pages=505–513 |year=1985 |title=Shift-Register Synthesis (Modulo n) |url=http://www2.research.attneilsloane.com/~njas/doc/1218shiftMe111.pdf |doi=10.1137/0214038 |citeseerx=10.1.1.48.4652 }}</ref>

[[Elwyn Berlekamp]] invented an algorithm for decoding [[BCH code|Bose–Chaudhuri–Hocquenghem (BCH) codes]].<ref>{{Citation

|last= Berlekamp

|first= Elwyn R.

|authorlinkauthor-link= Elwyn Berlekamp

|title= Nonbinary BCH decoding

|year= 1967

Line 12 ⟶ 14:

|last= Berlekamp

|first= Elwyn R.

|authorlinkauthor-link= Elwyn Berlekamp

|title=Algebraic Coding Theory

|place=Laguna Hills, CA

|origyearorig-year=1968

|year=1984

|edition= Revised

|publisher=Aegean Park Press

|isbn= 978-0-89412-063-83}}. Previous publisher McGraw-Hill, New York, NY.</ref> [[James Massey]] recognized its application to linear feedback shift registers and simplified the algorithm.<ref>{{Citation

|first= J. L.

|last= Massey

|authorlinkauthor-link= James Massey

|title= Shift-register synthesis and BCH decoding

|journal= IEEE Trans.Transactions on Information Theory

|volume= IT-15

|issue= 1

|yeardate= January 1969

|pages= 122–127

|url= http://crypto.stanford.edu/~mironov/cs359/massey.pdf}}</ref><ref>{{Citation

|doi= 10.1109/TIT.1969.1054260

|last= Ben Atti

|firsts2cid= Nadia9003708

}}</ref><ref>{{Citation

|lastlast1= Ben Atti

|first1= Nadia

|last2= Diaz-Toca

|first2= Gema M.

Line 37 ⟶ 42:

|first3= Henri

|title= The Berlekamp–Massey Algorithm revisited

|journal= Applicable Algebra in Engineering, Communication and Computing

|id= {{citeseerx|10.1.1.96.2743}}

|date=April 2006 |volume=17 |issue=1 |pages=75–82

|doi= }}</ref> Massey termed the algorithm the LFSR Synthesis Algorithm (Berlekamp Iterative Algorithm),<ref>{{Harvnb|Massey|1969|p=124}}</ref> but it is now known as the Berlekamp–Massey algorithm.

|id= {{citeseerx|=10.1.1.96.2743}}

|url= http://hlombardi.free.fr/publis/ABMAvar.html

|doi= 10.1007/s00200-005-0190-z

|arxiv= 2211.11721

|s2cid= 14944277

|doi= }}</ref> Massey termed the algorithm the LFSR Synthesis Algorithm (Berlekamp Iterative Algorithm),<ref>{{Harvnb|Massey|1969|p=124}}</ref> but it is now known as the Berlekamp–Massey algorithm.

==Description of algorithm==

The Berlekamp–Massey algorithm is an alternate methodalternative to solve the set of linear equations described in [[Reed–Solomon error correction#PetersonPeterson–Gorenstein–Zierler decoder|Reed–Solomon Peterson decoder]], whichfor solving the set of linear equations. It can be summarized as finding the coefficients Λ_''j'' of a polynomial Λ(''x'') so that for all positions ''i'' in an input stream ''S'':

:<math> S_{i + \nu} + \Lambda_1 S_{i+\nu-1} + \cdots + \Lambda_{\nu-1} S_{i+1} + \Lambda_{\nu} S_i = 0. </math>

In the code examples below, ''C''(''x'') is a potential instance of ''Λ''(''x''). The error locator polynomial ''C''(''x'') for ''L'' errors is defined as:

:<math> C(x) = C_{L} \C_L x^{L} + C_{L-1} \ x^{L-1} + \cdots + C_2 \ x^2 + C_1 \ x + 1 </math>

or reversed:

:<math> C(x) = 1 + C_1 \ x + C_2 \ x^2 + \cdots + C_{L-1} \ x^{L-1} + C_{L} \C_L x^{L}. </math>

The goal of the algorithm is to determine the minimal degree ''L'' and ''C''(''x'') which results in all [[Decoding methods#Syndrome decoding|syndromes]]

:<math> S_{n}S_n + C_1 \ S_{n-1} + \cdots + C_L \ S_{n-L}</math>

being equal to 0:

:<math> S_{n}S_n + C_1 \ S_{n-1} + \cdots + C_L \ S_{n-L} = 0,\qquad L\le n\le N-1.</math>

Algorithm:

''C''(''x'') is initialized to 1, ''L'' is the current number of assumed errors, and initialized to zero. ''N'' is the total number of syndromes. ''n'' is used as the main iterator and to index the syndromes from 0 to (''N''-1)−1. ''B''(''x'') is a copy of the last ''C''(''x'') since ''L'' was updated and initialized to 1. ''b'' is a copy of the last discrepancy ''d'' (explained below) since ''L'' was updated and initialized to 1. ''m'' is the number of iterations since ''L'', ''B''(''x''), and ''b'' were updated and initialized to 1.

Each iteration of the algorithm calculates a discrepancy ''d''. At iteration ''k'' this would be:

:<math> d =\gets S_{k}S_k + C_1 \ S_{k-1} + \cdots + C_L \ S_{k-L}.</math>

If ''d'' is zero, the algorithm assumes that ''C''(''x'') and ''L'' are correct for the moment, increments ''m'', and continues.

If ''d'' is not zero, the algorithm adjusts ''C''(''x'') so that a recalculation of ''d'' would be zero:

:<math>C(x) =\gets C(x) \ - \ (d / b) \ x^m \ B(x).</math>

The ''x^m'' term ''shifts'' B(x) so it follows the syndromes corresponding to ''b''. If the previous update of ''L'' occurred on iteration ''j'', then ''m'' = ''k'' -− ''j'', and a recalculated discrepancy would be:

:<math> d =\gets S_{k}S_k + C_1 \ S_{k-1} + \cdots - (d/b) (S_{j}S_j + B_1 \ S_{j-1} + \cdots ).</math>

This would change a recalculated discrepancy to:

:<math> d = d - (d/b)b = d - d = 0. \ </math>

The algorithm also needs to increase ''L'' (number of errors) as needed. If ''L'' equals the actual number of errors, then during the iteration process, the discrepancies will become zero before ''n'' becomes greater than or equal to (2 ''L''). Otherwise ''L'' is updated and the algorithm will update ''B''(''x''), ''b'', increase ''L'', and reset ''m'' = 1. The formula ''L'' = (''n'' + 1 -− ''L'') limits ''L'' to the number of available syndromes used to calculate discrepancies, and also handles the case where ''L'' increases by more than 1.

==Code samplePseudocode ==

The algorithm from {{Harvtxt|Massey|1969|p=124}} for an arbitrary field:

<!-- Notes: notation changes from Massey:

Massey Here

Line 94 ⟶ 104:

T(D) T(x) polynomial

-->

<div class="mw-highlight mw-highlight-lang-c mw-content-ltr">

<source lang=C>

polynomial(field ''K'') s(x) = ... <span class="cm">/* coeffs are s_js_j; output sequence as N-1 degree polynomial) */</span>

<span class="cm">/* connection polynomial */</span>

polynomial(field K) C(x) = 1; <span class="cm">/* coeffs are c_jc_j */</span>

polynomial(field K) B(x) = 1;

int L = 0;

int m = 1;

field K b = 1;

int n;

<span class="cm">/* steps 2. and 6. */</span>

<span class="k">for</span> (n = 0; n < N; n++) {

/* steps 2. and 6. */

<span class="cm">/* step 2. calculate discrepancy */</span>

for (n = 0; n < N; n++)

field K d = s_n + {{math|&sum;{{su|p=L|b=i=1}} c_i s_{n - i}}} <!--Σi=1Lci⋅sn−i;-->

{

/* step 2. calculate discrepancy */

field K d<span class="k">if</span> s_n(d + \Sigma_{i=1}^L c_i= *0) s_{n-i};

<span class="cm">/* step 3. discrepancy is zero; annihilation continues */</span>

if (d m == 0)m + 1;

} <span class="k">else</span> <span class="k">if</span> (2 * L <= n) {

{

<span class="cm">/* step 35. discrepancy is zero; annihilation continues */</span>

m <span class="cm">/* temporary copy mof +C(x) 1;*/</span>

polynomial(field K) T(x) = C(x);

{

C(x) = C(x) - d b^{-1}⁻¹ x^^m B(x);

L = n + 1 - L;

m B(x) = 1T(x);

m = mb += 1d;

/* step 5.m */= 1;

} <span class="k">else</span> {

<span class="cm">/* step 4. */</span>

C(x) = C(x) - d b^{-1}⁻¹ x^^m B(x);

else if (2 * L < m = n)m + 1;

}

}

else if (2 * L <= n)

<span class="k">return</span> L;

{

</div>

/* step 5. */

/* temporary copy of C(x) */

polynomial(field K) T(x) = C(x);

In the case of binary GF(2) BCH code, the discrepancy d will be zero on all odd steps, so a check can be added to avoid calculating it.

C(x) = C(x) - d b^{-1} x^m B(x);

{{sxhl|2=c|1=<nowiki/>

L = n + 1 - L;

/* ... B(x) = T(x);*/

for (n = 0; n < bN; =n++) d;{

/* if odd step number, discrepancy == 0, no need to calculate it */

m = 1;

if ((n&1) != 0) {

m = m + 1;

continue;

}

/* ... else*/

}}

{

/* step 4. */

C(x) = C(x) - d b^{-1} x^m B(x);

m = m + 1;

}

}

return L;

</source>

==The algorithm for the binary field==

The following is the Berlekamp–Massey algorithm specialized for the typical binary [[finite field]] F₂ (also written GF(2)). The field elements are '0' and '1'. The field operations '+' and '−' are identical and are equivalent to the 'exclusive or' operation, XOR. The multiplication operator '*' becomes the logical AND operation. The division operator reduces to the identity operation (i.e., field division is only defined for dividing by 1, and x/1 = x).

#Let <math>s_0, s_1, s_2 \cdots s_{n-1}</math> be the [[bit]]s of the stream.

#Initialise two arrays <math>b</math> and <math>c</math> each of length <math>n</math> to be zeroes, except <math>b_0 \leftarrow 1, c_0 \leftarrow 1</math>

#[[Assignment (computer science)|assign]] <math>L \leftarrow 0, m \leftarrow -1</math>.

#'''For''' <math>N = 0</math> '''step''' 1 '''while''' <math>N < n </math>:<!-- should be N <= n ??? -->

#*Let <math>d</math> be <math>s_N + c_1s_{N-1} + c_2s_{N-2} + \cdots + c_Ls_{N-L}</math>.<!-- These are operations in the FIELD -->

#*'''if''' <math>d = 0</math>, '''then''' <math>c</math> is already a polynomial which annihilates the portion of the stream from <math>N-L</math> to <math>N</math>.

#*'''else''':

#** Let <math>t</math> be a copy of <math>c</math>.

#** Set <math>c_{N-m} \leftarrow c_{N-m} \oplus b_0, c_{N-m+1} \leftarrow c_{N-m+1} \oplus b_1, \dots </math> up to <math>c_{n-1} \leftarrow c_{n-1} \oplus b_{n-N+m-1}</math> (where <math>\oplus</math> is the [[Exclusive or]] operator).

#** If <math>L \le \frac{N}{2}</math>, set <math>L \leftarrow N+1-L</math>, set <math>m \leftarrow N</math>, and let <math>b \leftarrow t</math>; otherwise leave <math>L</math>, <math>m</math> and <math>b</math> alone.

At the end of the algorithm, <math>L</math> is the length of the minimal LFSR for the stream, and we have <math>c_Ls_a + c_{L-1}s_{a+1} + c_{L-2}s_{a+2} + \cdots = 0</math> for all <math>a</math>.<!-- this expression is in the FIELD -->

==See also==

* [[Reed–Solomon error correction]]

* [[Reeds–Sloane algorithm]], an extension for sequences over integers mod&nbsp;''n''

* [[Nonlinear-feedback shift register]] (NLFSR)

* [[Berlekamp–Welch algorithm]]

* [[NLFSR]], Non-Linear Feedback Shift Register

==References==

Line 164 ⟶ 162:

==External links==

* {{springer|title=Berlekamp-Massey algorithm|id=p/b120140}}

* [https://web.archive.org/web/20120716181541/http://planetmath.org/encyclopedia/{{PlanetMath|BerlekampMasseyAlgorithm.html |Berlekamp–Massey algorithm] at [[PlanetMath]].}}

* {{MathWorld|urlname=Berlekamp-MasseyAlgorithm|title=Berlekamp–Massey Algorithm}}

* [https://code.google.com/p/lfsr/ GF(2) implementation in Mathematica]

* {{dein iconlang|de}} [http://www.informationsuebertragung.ch/indexAlgorithmen.html Applet Berlekamp–Massey algorithm]

* [httphttps://berlekamp-massey-algorithm.appspot.com/ Online GF(2) Berlekamp-Massey calculator]

{{DEFAULTSORT:Berlekamp-Massey Algorithm}}

[[Category:Error detection and correction]]

[[Category:Cryptanalytic algorithms]]

[[Category:Articles with example code]]