Having problems in multiplying the inequalities

Join the discussion
Registration is free. Ask a follow-up in this thread, or start your own.
19 replies · 2K views
Adesh
Messages
735
Reaction score
191
Homework Statement
Can we multiply two inequalities element-wise?
Relevant Equations
## x \lt y##
Let’s say we are given two inequalities $$ x \lt y \\
a \lt b$$ then we can write (we can even prove it using logarithms) $$ ax \lt by$$ given that every number is positive.

In this article (Fact 4.1, point (ii) ) it is given that if ##x, y## are positive numbers and ##a, b## are negative numbers and if the following is the relation $$ x \lt y \\ a \lt b$$ then the multiplication would cause the inequality to flip over, I.e. $$ ax \gt b~y$$ .

But I have a counter example for that, $$ 2 \lt 3 \\ - 2 \lt -1 \\ -4 \lt -3$$ You see we got no flipping.

Please explain me.
 
  • Like
Likes   Reactions: PeroK
Physics news on Phys.org
Adesh said:
Homework Statement:: Can we multiply two inequalities element-wise?
Relevant Equations:: ## x \lt y##

Let’s say we are given two inequalities $$ x \lt y \\
a \lt b$$ then we can write (we can even prove it using logarithms) $$ ax \lt by$$ given that every number is positive.

In this article (Fact 4.1, point (ii) ) it is given that if ##x, y## are positive numbers and ##a, b## are negative numbers and if the following is the relation $$ x \lt y \\ a \lt b$$ then the multiplication would cause the inequality to flip over, I.e. $$ ax \gt b~y$$ .

But I have a counter example for that, $$ 2 \lt 3 \\ - 2 \lt -1 \\ -4 \lt -3$$ You see we got no flipping.

Please explain me.
The article is clearly wrong, as your counterexample shows. Can you fix it?
 
  • Sad
Likes   Reactions: Adesh
PeroK said:
The article is clearly wrong, as your counterexample shows. Can you fix it?
No! I don’t have any access to it. It just came in the google search when I typed “multiplying the opposite inequalities”.
 
  • Like
Likes   Reactions: PeroK
PeroK said:
I didn't mean fix the document. I meant fix the inequality!
Which inequality? I beg your pardon.
 
PeroK said:
The one you posted about. 4.1 (ii). What should it be?
Undetermined, as far as I can think ha?
 
PeroK said:
Maybe, maybe not.
Is there any rule which confirms that whether the inequality going to flip or not? Can we make sure whether it going to flip or not? Let’s consider $$ x\lt y $$ such that both of them are positive integers. And if ##a,b## are negative integers with the relation $$ a \lt b$$ then what conditions must be met for the following to be true $$ ax \lt b~y$$
 
Let's focus on the case:

Adesh said:
Let’s consider $$ x\lt y $$ such that both of them are positive integers. And if ##a,b## are negative integers with the relation $$ a \lt b$$

Although these need not be integers. What inequality must hold in that case?
 
PeroK said:
Let's focus on the case:
Although these need not be integers. What inequality must hold in that case?
$$-x \gt -y$$
$$- a \gt -b$$
 
Adesh said:
$$-x \gt -y$$
$$- a \gt -b$$

Okay, but we want something that involves all four numbers. I saw your attempted solution to that maths puzzle about the rationals. You were pretty close there. You're capable of working this out if you focus on it
 
  • Like
Likes   Reactions: Adesh
PeroK said:
Okay, but we want something that involves all four numbers. I saw your attempted solution to that maths puzzle about the rationals. You were pretty close there. You're capable of working this out if you focus on it
Okay sir, I will post my thinking (about the inequality) here after a strong contemplation and I know you’re always there to help and encourage students.
 
  • Like
Likes   Reactions: PeroK
All right, $$ x \lt y ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~x, y \in \mathbb {R^{+}} \\
a \lt b ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~a, b \in \mathbb{R^{-}} \\
\implies |b| \lt |a| $$
So, we got two inequalities (involving positive numbers) going in the same side $$ x \lt y ~~~~~~~~~~(1) \\
|b| \lt |a| ~~~~~~~~(2)$$ Add (1) and (2) component wise $$ x +|b| \lt y + |a| ~~~~~~~~~~~~~~(3) $$
Now, subtracting ##|a|## from ##x## and ##|b|## from ##y## will obviously make the inequality even stronger, therefore we have $$ x - |a| \lt y - |b|~~~~~~~~~~~~~~~~~~~(4) $$

Multiply inequality ##(3)## and ##(4)## component -wise, $$ ( x + |b|) ~(x - |a|) \lt (y+ |a|) ~(y-|b|) \\
x^2 - |a|x + |b|x - |a||b| \lt y^2 - |b|y + |a|y - |a||b| \\
x^2 -|a|x + |b|x \lt y^2 -|b|y +|a|y$$
Since, it is very obvious that if ##x \lt y## then ##x^2 \lt y^2##, therefore the above inequality may be written as $$ |b|x - |a|x \lt |a|y - |b|y ~~~~~~~~~~~~~~~~~~~~~~~~~~~(5)$$

From (1) and (2) we have ## |b|x \lt |a| y ## (as each and every number is positive) and therefore ## 0 \lt |a|y - |b| x ##. Writing out inequality (5) in this fashion
$$ |b| y - |a| x \lt |a|y - |b|x $$ Although, out RHS of the above inequality is positive but since LHS is less than our RHS, hence no judgement can be done about LHS. It may be positive or negative. That is we have two cases without any determination about which of them is true $$ |b|y - |a|x \lt 0~~~~~~~~~~~~~~~~~~~~~~(i) \\
\mathbf{OR} \\

|b|y - |a|x \gt 0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~(ii)$$

Doing some little algebra with the above inequalities $$ |b|y \lt |a|x \\
ax \lt b~y ~~~~~~~~~~~~~~~~~~~(\textrm{to remove the modulus we multiplied both sides by negative 1}) $$
$$ \mathbf{AND} \\

|b|y \gt |a|x \\
b~y \lt ax $$ That is, it is not possible to conclude whether the inequality going to flip or not.
 
PeroK said:
The pdf had a simple typo. It should have been:
0<A<B and C≤D<0 ⇒ BC<AD<00<A<B and C≤D<0 ⇒ BC<AD<0​
Yes, wow!

$$A \lt B \\ |D| \lt |C| \\ A|D| \lt B|C| \\ AD \gt BC $$
 
  • Like
Likes   Reactions: PeroK
PeroK said:
That shouldn't have been hard to work out.
I was just writing it :)
 
PeroK said:
That shouldn't have been hard to work out.
Have I understood the things the way you wanted me, sir?
 
Adesh said:
Have I understood the things the way you wanted me, sir?
You don't need to call me sir! Yes, it was just to work out that there is an inequality, but not the one in the article.
 
  • Like
Likes   Reactions: Adesh