শুক্রবার, ৭ মার্চ, ২০১৪

LCM (Least Common Multiple)

The least common multiple of two numbers a and b, variously denoted LCM(a,b) (this work; Zwillinger 1996, p. 91; Råde and Westergren 2004, p. 54), lcm(a,b) (Gellert et al. 1989, p. 25; Graham et al. 1990, p. 103; Bressoud and Wagon 2000, p. 7; D'Angelo and West 2000, p. 135; Yan 2002, p. 31; Bronshtein et al. 2007, pp. 324-325; Mathematica), l.c.m.(a,b) (Andrews 1994, p. 22; Guy 2004, pp. 312-313), or [a,b], is the smallest positive number m for which there exist positive integers n_a and n_b such that
 n_aa=n_bb=m.
(1)
The least common multiple LCM(a,b,c,...) of more than two numbers is similarly defined.
The least common multiple of ab, ... is implemented in Mathematica as LCM[ab, ...].
The least common multiple of two numbers a and b can be obtained by finding the prime factorization of each
a=p_1^(a_1)...p_n^(a_n)
(2)
b=p_1^(b_1)...p_n^(b_n),
(3)
where the p_is are all prime factors of a and b, and if p_i does not occur in one factorization, then the corresponding exponent is taken as 0. The least common multiple is then given by
 LCM(a,b)=product_(i=1)^np_i^(max(a_i,b_i)).
(4)
For example, consider LCM(12,30).
12=2^2·3^1·5^0
(5)
30=2^1·3^1·5^1,
(6)
so
 LCM(12,30)=2^2·3^1·5^1=60.
(7)
LCM
The plot above shows LCM(1,r) for rational r=m/n, which is equivalent to the numerator of the reduced form of m/n.
LCMArray
The above plots show a number of visualizations of LCM(i,j) in the (i,j)-plane. The figure on the left is simply LCM(i,j), the figure in the middle is the absolute values of the two-dimensional discrete Fourier transform of LCM(i,j) (Trott 2004, pp. 25-26), and the figure at right is the absolute value of the transform of 1/LCM(i,j).
LeastCommonMultipleDensity
The least common multiples of the first n positive integers for n=1, 2, ... are 1, 2, 6, 12, 60, 60, 420, 840, ... (Sloane's A003418; Selmer 1976), which is related to the Chebyshev function psi(n). For n>=7LCM(1,2,...,n)>2^n (Nair 1982ab, Tenenbaum 1990). The prime number theorem implies that
 LCM(1,2,...,n)=e^(n(1+o(1)))
(8)
as n->infty, in other words,
 (ln(LCM(1,2,...,n)))/n->1
(9)
as n->infty.
Let m be a common multiple of a and b so that
 m=ha=kb.
(10)
Write a=a_1GCD(a,b) and b=b_1GCD(a,b), where a_1 and b_1 are relatively prime by definition of the greatest common divisor GCD(a_1,b_1)=1. Then ha_1=kb_1, and from the division lemma (given that ha_1 is divisible by b_1 and GCD(b_1,a_1)=1), we have h is divisible by b_1, so
 h=nb_1
(11)
 m=ha=nb_1a=n(ab)/(GCD(a,b)).
(12)
The smallest m is given by n=1,
 LCM(a,b)=(ab)/(GCD(a,b)),
(13)
so
 GCD(a,b)LCM(a,b)=ab
(14)
The LCM is idempotent
 LCM(a,a)=a,
(15)
commutative
 LCM(a,b)=LCM(b,a),
(16)
associative
LCM(a,b,c)=LCM(LCM(a,b),c)
(17)
=LCM(a,LCM(b,c)),
(18)
distributive
 LCM(ma,mb,mc)=mLCM(a,b,c),
(19)
and satisfies the absorption law
 GCD(a,LCM(a,b))=a.
(20)
It is also true that
LCM(ma,mb)=(GCD(ma)GCD(mb))/(GCD(ma,mb))
(21)
=m(ab)/(GCD(a,b))
(22)
=mLCM(a,b).
(23)


Ref: mathworld.wolfram.com