Back

Number Theory

1: Bounds for the p-adic Valuation of n!

Problem Statement: Prove that for every prime pp and any integer n1n \geq 1 we have:

np1<vp(n!)np1\frac{n}{p} - 1 < v_p(n!) \leq \frac{n}{p-1}

Proof: By Legendre's formula, the pp-adic valuation of n!n! (the exponent of the highest power of pp that divides n!n!) is given by:

vp(n!)=k=1npkv_p(n!) = \sum_{k=1}^{\infty} \left\lfloor \frac{n}{p^k} \right\rfloor

Upper Bound: We can bound the sum from above by dropping the floor function:

vp(n!)=k=1npkk=1npkv_p(n!) = \sum_{k=1}^{\infty} \left\lfloor \frac{n}{p^k} \right\rfloor \leq \sum_{k=1}^{\infty} \frac{n}{p^k}

Evaluating this infinite geometric series, we get:

k=1npk=nk=1(1p)k=n(1p11p)=np1\sum_{k=1}^{\infty} \frac{n}{p^k} = n \sum_{k=1}^{\infty} \left(\frac{1}{p}\right)^k = n \left( \frac{\frac{1}{p}}{1 - \frac{1}{p}} \right) = \frac{n}{p-1}

Thus, vp(n!)np1v_p(n!) \leq \frac{n}{p-1}.

Lower Bound: To find the lower bound, we observe that the sum is strictly greater than or equal to its first term (since all subsequent terms are non-negative):

vp(n!)=np+np2+npv_p(n!) = \left\lfloor \frac{n}{p} \right\rfloor + \left\lfloor \frac{n}{p^2} \right\rfloor + \dots \geq \left\lfloor \frac{n}{p} \right\rfloor

Knowing the property of the floor function that x>x1\lfloor x \rfloor > x - 1, we have:

vp(n!)np>np1v_p(n!) \geq \left\lfloor \frac{n}{p} \right\rfloor > \frac{n}{p} - 1

Conclusion: Combining both bounds, we obtain the required inequality:

np1<vp(n!)np1\frac{n}{p} - 1 < v_p(n!) \leq \frac{n}{p-1}

p-adic valuation bounds

p = 3, n = 1…50

02468101214161820222415101520253035404550nv₃(n!)
v₃(n!)bound region

\blacksquare

2: Bounds for Central Binomial Coefficients

Problem Statement: Prove that for every integer n1n \geq 1, we have:

4n>(2nn)4n2n+14^n > \binom{2n}{n} \geq \frac{4^n}{2n+1}

and

(2n+1n)<4n\binom{2n+1}{n} < 4^n

Proof:

Part 1: Bounding (2nn)\binom{2n}{n} We start by recalling the Binomial Theorem:

(a+b)m=k=0m(mk)akbmk(a+b)^m = \sum_{k=0}^{m} \binom{m}{k} a^k b^{m-k}

Substituting a=1a=1, b=1b=1, and m=2nm=2n, we get:

(1+1)2n=22n=4n(1+1)^{2n} = 2^{2n} = 4^n

Expanding this using the sum, we have exactly 2n+12n+1 terms:

4n=k=02n(2nk)=(2n0)+(2n1)++(2nn)++(2n2n)4^n = \sum_{k=0}^{2n} \binom{2n}{k} = \binom{2n}{0} + \binom{2n}{1} + \dots + \binom{2n}{n} + \dots + \binom{2n}{2n}

Since n1n \geq 1, there are multiple positive terms in this sum. Therefore, the entire sum is strictly greater than its largest single term, which is the central binomial coefficient (2nn)\binom{2n}{n}:

4n>(2nn)4^n > \binom{2n}{n}

Furthermore, since (2nn)\binom{2n}{n} is the maximum term among all 2n+12n+1 terms, replacing every term in the sum with (2nn)\binom{2n}{n} gives an upper bound on the total sum:

4n=k=02n(2nk)k=02n(2nn)=(2n+1)(2nn)4^n = \sum_{k=0}^{2n} \binom{2n}{k} \leq \sum_{k=0}^{2n} \binom{2n}{n} = (2n+1)\binom{2n}{n}

Dividing by 2n+12n+1, we obtain the lower bound:

4n2n+1(2nn)\frac{4^n}{2n+1} \leq \binom{2n}{n}

Combining these results, we get our first set of inequalities:

4n2n+1(2nn)<4n\frac{4^n}{2n+1} \leq \binom{2n}{n} < 4^n

Central binomial coefficient bounds

n = 1…15

10⁰10¹10²10³10⁴10⁵10⁶10⁷10⁸10^910^1013579111315nlog₁₀
C(2n, n)bound region

Part 2: Bounding (2n+1n)\binom{2n+1}{n} Now consider the expansion of (1+1)2n+1(1+1)^{2n+1}:

(1+1)2n+1=22n+1=222n=24n(1+1)^{2n+1} = 2^{2n+1} = 2 \cdot 2^{2n} = 2 \cdot 4^n

Expanding this as a sum:

24n=k=02n+1(2n+1k)2 \cdot 4^n = \sum_{k=0}^{2n+1} \binom{2n+1}{k}

From the symmetry of binomial coefficients, we know (mk)=(mmk)\binom{m}{k} = \binom{m}{m-k}. Thus, the two central terms are equal:

(2n+1n)=(2n+1n+1)\binom{2n+1}{n} = \binom{2n+1}{n+1}

Writing out the full sum, we get:

24n=(2n+10)++(2n+1n)+(2n+1n+1)++(2n+12n+1)2 \cdot 4^n = \binom{2n+1}{0} + \dots + \binom{2n+1}{n} + \binom{2n+1}{n+1} + \dots + \binom{2n+1}{2n+1}

Because n1n \geq 1, there are strictly positive terms on the outside of the two central terms. Therefore, the sum of all terms is strictly greater than the sum of just the two middle terms:

24n>(2n+1n)+(2n+1n+1)2 \cdot 4^n > \binom{2n+1}{n} + \binom{2n+1}{n+1}

Since these two middle terms are equal, we can substitute:

24n>2(2n+1n)2 \cdot 4^n > 2 \binom{2n+1}{n}

Dividing both sides by 2, we obtain the final inequality:

4n>(2n+1n)4^n > \binom{2n+1}{n}

Odd binomial coefficient bound

n = 1…15

10⁰10¹10²10³10⁴10⁵10⁶10⁷10⁸10^910^1013579111315nlog₁₀
C(2n+1, n)gap

\blacksquare

3: Bounding the Product of Primes

Problem Statement: Prove that for every integer n2n \geq 2, the product of all primes not exceeding nn is less than 4n14^{n-1}.

pnp<4n1\prod_{p \leq n} p < 4^{n-1}

Proof: We will prove this using strong induction on nn.

Base Case: For n=2n=2, the only prime not exceeding 22 is 22 itself.

2<421=42 < 4^{2-1} = 4

The base case holds.

Inductive Step: Assume the inequality holds for all integers kk such that 2k<n2 \leq k < n. We consider two cases for nn:

Case 1: nn is even. If nn is even and n>2n > 2, nn cannot be prime. Therefore, the product of primes up to nn is exactly the same as the product of primes up to n1n-1:

pnp=pn1p\prod_{p \leq n} p = \prod_{p \leq n-1} p

By our strong induction hypothesis, since n1<nn-1 < n, we have:

pn1p<4(n1)1=4n2<4n1\prod_{p \leq n-1} p < 4^{(n-1)-1} = 4^{n-2} < 4^{n-1}

Thus, the inequality holds when nn is even.

Case 2: nn is odd. Let n=2m+1n = 2m+1 for some integer m1m \geq 1. We can split the product of primes into two parts:

p2m+1p=(pm+1p)×(m+1<p2m+1p)\prod_{p \leq 2m+1} p = \left( \prod_{p \leq m+1} p \right) \times \left( \prod_{m+1 < p \leq 2m+1} p \right)

By the induction hypothesis, the first part is bounded by:

pm+1p<4(m+1)1=4m\prod_{p \leq m+1} p < 4^{(m+1)-1} = 4^m

For the second part, consider the binomial coefficient from Exercise 2:

(2m+1m)=(2m+1)!m!(m+1)!\binom{2m+1}{m} = \frac{(2m+1)!}{m!(m+1)!}

Any prime pp strictly greater than m+1m+1 and less than or equal to 2m+12m+1 will divide the numerator (2m+1)!(2m+1)!, but it is too large to divide any term in the denominator m!(m+1)!m!(m+1)!. Therefore, the product of all such primes must divide (2m+1m)\binom{2m+1}{m}. This gives us the inequality:

m+1<p2m+1p(2m+1m)\prod_{m+1 < p \leq 2m+1} p \leq \binom{2m+1}{m}

From the previous exercise, we already established that (2m+1m)<4m\binom{2m+1}{m} < 4^m. Therefore:

m+1<p2m+1p<4m\prod_{m+1 < p \leq 2m+1} p < 4^m

Combining the bounds for both parts of our product:

p2m+1p<4m4m=42m\prod_{p \leq 2m+1} p < 4^m \cdot 4^m = 4^{2m}

Since n=2m+1n = 2m+1, we know that 2m=n12m = n-1. Substituting this back in yields:

pnp<4n1\prod_{p \leq n} p < 4^{n-1}

Conclusion: By the principle of strong induction, the inequality holds for all integers n2n \geq 2.

Product of primes bound

n = 2…30

10⁰10¹10²10³10⁴10⁵10⁶10⁷10⁸10⁹10¹⁰10¹¹10¹²10¹³10¹⁴10¹⁵10¹⁶10¹⁷10¹⁸251015202530nlog₁₀
prime (product jumps)gap

\blacksquare