Efficiency and the Euclidean Algorithm

Click For Summary
The discussion focuses on proving that every two steps in the Euclidean algorithm reduce the remainder by at least half, specifically showing that r_(i+2) < (1/2)r_i for all i. An initial proof attempt is presented, but concerns arise regarding the conditions of a and b, particularly if a is negative. The proof relies on the relationship between successive remainders and the quotients, asserting that if b > 0, the quotients will be greater than or equal to 1. The conversation emphasizes the importance of using absolute values for a and b to maintain valid conditions throughout the algorithm. Overall, the participants seek clarity on the proof while addressing potential issues with negative values.
Buri
Messages
271
Reaction score
0
1. Question

Let b = r_0, r_1, r_2,... be the successive remainders in the Euclidean algorithm applied to a and b. Show that every two steps reduces the remainder by at least one half. In other words, verify that r_(i+2) < (1/2)r_i for every i = 0,1,2,...

2. Attempt at a solution

I take an arbitrary 'line' in the euclidean algorithm which is of the following form:

r_i = q_(i+2) x r_(i+1) + r_(i + 2)

So clearly I have the following:

1 > r_(i+2) / r_(i+1), because r_0 > r_1 > r_2 ...

Now q_(i+2) >= 1. So we have:

q_(i+2) > r_(i+2) / r_(i+1)

Which implies:

q_(i+2) x r_(i+1) > r_(i+2)
q_(i+2) x r_(i+1) + r_(i+2) > 2r_(i+2)

But q_(i+2) x r_(i+1) + r_(i+2) = r_i. Thus, r_i > 2r_(i+2) which is what I wanted to prove.

3. Comments

However, I see possible problems with this proof. To begin with, the Euclidean Algorithm I am working with doesn't explicitly state that a > 0 or b > 0, but in this question it seems to be implicitly be saying that b > 0 (and says nothing of a) because b = r_0 which must be greater than 0. But if a were negative and b positive I could have a q which is not necessarily positive (more specifically, not greater or equal to 1) which would mess up my proof. But since it seems to be that b > 0 then q's will always be strictly greater than 1, even though it could be possible that the first q could turn out to be negative (for example in, a = -9 and b = 4 in which case q = -3 and r = 3). See what I mean?

I would prefer if people could help me sort this proof out rather than give me ideas of how to go about it differently.I'd appreciate the help! :)
 
Last edited:
Physics news on Phys.org
Anyone?
 
I know you asked for people not to give you a different method. However, it seems to me you're doing a lot more work than you need to.

Since ri=qi+2ri+1+ri+2 and qi+2≥1, you have ri≥ri+1+ri+2. Then since ri+1>ri+2, you have ri>ri+2+ri+2⇔ri>2ri+2⇔ri+2<1/2ri.

Further, since gcd(a,b)=gcd(-a,b)=gcd(a,-b)=gcd(-a,-b), you may use the absolute values of a and b while using the algorithm, ensuring that every q is always greater or equal to 1.
 
Last edited:
Thread 'How to define a vector field?'
Hello! In one book I saw that function ##V## of 3 variables ##V_x, V_y, V_z## (vector field in 3D) can be decomposed in a Taylor series without higher-order terms (partial derivative of second power and higher) at point ##(0,0,0)## such way: I think so: higher-order terms can be neglected because partial derivative of second power and higher are equal to 0. Is this true? And how to define vector field correctly for this case? (In the book I found nothing and my attempt was wrong...

Similar threads

  • · Replies 4 ·
Replies
4
Views
4K
  • · Replies 21 ·
Replies
21
Views
1K
  • · Replies 4 ·
Replies
4
Views
879
Replies
5
Views
2K
Replies
15
Views
2K
Replies
1
Views
2K
  • · Replies 5 ·
Replies
5
Views
3K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 11 ·
Replies
11
Views
3K
  • · Replies 9 ·
Replies
9
Views
1K