swastik pramanik

Forum Replies Created

Viewing 5 posts - 1 through 5 (of 15 total)
  • Author
    Posts
  • in reply to: Number Theory #25090
    swastik pramanik
    Participant

    In part (c) in line 3 the inequality should be \(1-\frac{1}{p}\le 1-\frac{1}{\sqrt{n}}\) not  \(1-\frac{1}{p}\le 1-\sqrt{n}\).....

    in reply to: Number Theory #25089
    swastik pramanik
    Participant

    (a) Let \(n=2^{k_0}p_1^{k_1}p_2^{k_2}\cdots p_r^{k_r}\) . So, \(\phi(n)=2^{k_0-1}p_1^{k_1-1}p_2^{k_2-1}\cdots p_r^{k_r-1}(2-1)(p_1-1)(p_2-1)\cdots (p_r-1)\) . Now, we use the inequalities \(k-\frac{1}{2}\ge \frac{k}{2}\)  and \(p-1>\sqrt{p}\) for \(p>2\). So, we have \(\phi(n)\le 2^{k_0-1}p_1^{k_1/2}p_2^{k_2/2}\cdots p_r^{k_r/2}\ge \frac{1}{2}\sqrt{n}\)  .

    Now, also \(p-1<p\)  . So, \(\phi(n)\le 2^{k_0}p_1^{k_1}p_2^{k_2}\cdots p_r^{k_r}=n\)  .

    Combining them up we get $$\frac{1}{2}\sqrt{n}\le \phi(n)\le n$$ as required.

     

    (c) Let \(p\) be the smallest prime divisor of \(n\) so that, \(p\le \sqrt{n}\).  Then \(\phi(n) \le n\left(1-\frac{1}{p}\right)\) . We know that, \(p\le \sqrt{n}\) which implies \(1-\frac{1}{p}\le 1-\sqrt{n}\). So, \(\phi(n) \le n\left(1-\frac{1}{p}\right)\le n\left(1-\frac{1}{\sqrt{n}}\right)=n-\sqrt{n}\).

    Hence, we are done!  🙂

    in reply to: Number Theory #25059
    swastik pramanik
    Participant

    (b) Let \(n=\prod_{k=1}^r p_k^{a_k}=p_1^{a_1}p_2^{a_2}p_3^{a_3}\cdots p_r^{a_r}\) . So, \(\phi(n) =n\left(1-\frac{1}{p_1}\right) \left(1-\frac{1}{p_2}\right)\cdots \left(1-\frac{1}{p_r}\right)\) .

    For any prime \(p\)  the inequality $$2p-2\geq p$$ or, $$\frac{p-1}{p}\geq \frac{1}{2}$$ So, we have \(r\) prime number for which the inequality is true. Multiplying them gives $$\left(1-\frac{1}{p_1}\right) \left(1-\frac{1}{p_2}\right)\cdots \left(1-\frac{1}{p_r}\right)\geq \frac{1}{2^r}$$ multiplying both sides with \(n\) gives So, we want to show that $$n\left(1-\frac{1}{p_1}\right) \left(1-\frac{1}{p_2}\right)\cdots \left(1-\frac{1}{p_r}\right)\geq \frac{n}{2^r}$$ or concisely $$\phi(n)\geq \frac{n}{2^r}$$ as required.

    in reply to: counting in two ways... find bijection #24710
    swastik pramanik
    Participant

    Notice that \(|A\times B|=mn\). Also, \(|C_i|=n\) for all \(i=1,2,3,\cdots ,m\). hence \(\sum_{i=1}^m |C_i|=\sum_{i=1}^m n=mn\). Similarly, we can prove that \(\sum_{j=1}^n |D_j|=mn\).

    And hence, we get our desired result.

    in reply to: Is a circumcircle real? #24709
    swastik pramanik
    Participant

    This is just proving that:

    Theorem: 

    Given any three non-collinear points \(A, B, C\)  there exists a unique circle passing through \(A, B, C\).

    Proof: 

    Let us consider three vertices of \(\Delta ABC\) i.e. \(A, B, C\). Suppose the perpendicular bisectors of \(BC\) and \(CA\) meet at \(S\). Then \(S\) lies on the perpendicular bisector of \(BC\) implies that \(SB=SC\). Again, \(S\) is also a perpendicular bisector of \(CA\)  which implies that \(SC=SA\). Hence, we have \(SA=SB=SC\). Further more, any point equidistant from \(A, B, C\)  should lie on the perpendicular bisectors of \(BC, CA\). Therefore, \(S\) is the only point equidistant from \(A, B, C\) and so the circle with centre \(S\) and radius \(SA\) is the unique circle passing through \(A, B, C\).

Viewing 5 posts - 1 through 5 (of 15 total)