Roots of equations texts by Patrizio Gravano
Index 0. Introduction 1. Quadratic equations 2. Algebraic equations in x of degree n > 2
Autumn 2014
A new little note on equations in x of degree n ≼ 2, studied in R. Possible location of positive roots. by Patrizio Gravano
0. Introduction In this paper, I would like to consider the possible location of positive roots of algebraic equations of degree n ≼ 2, studied in R. Firstly, I would like to consider the case n = 2 (quadratic equations). Secondly, I would like to discuss the general case n > 2.
1. Quadratic equations To solve a quadratic equation as ađ?‘Ľ 2 + bx + c = 0 you can use the following formula x=
−đ?‘?Âąâˆšđ?‘? 2 −4đ?‘Žđ?‘? 2đ?‘Ž
.
It is a very important formula. However, using elementary algebraic theory you can say if exists real number x, called unknown, for example đ?‘Ľđ?‘œ that ađ?‘Ľ0 2 + bđ?‘Ľ0 + c = 0 is true. In general, you can say when x exists and when it is an real number! In fact, in this speech I would observe that if you have the following equation (x ∓ a) ( x ∓b) = 0 ⇒ x = Âą a and x = Âą đ?‘?. You can observe that in general AB = 0 ⇒ A = 0 or B = 0 or A = B = 0. More, you can use this symbol ⇔ (necessary and sufficient condition). In general, you have j algebraic products of sums such as (x ∓đ?‘&#x;đ?‘ ) for any r real number and for every integer s ≤ j you can obviusly extend this procedure at the case j > 2.
When you consider the case that you have j =2 products of sums you can note that it is not necessary use formulas as x =
−đ?‘?Âąâˆšđ?‘? 2 −4đ?‘Žđ?‘? 2đ?‘Ž
.
Now, I would like to explain a particular case to demonstrate no real solutions in a specific case. In fact, in this one I admit a = b = c â&#x2030; 0. We can write a â&#x2C6;&#x2C6; ( -â&#x2C6;&#x17E;, 0) â&#x2C6;Ş ( 0 , -â&#x2C6;&#x17E;), or a â&#x2C6;&#x2C6; R {0}. In this case there is no solution in R. R denote the set which contains all real numbers (continuous real set). If you admit a = b = c â&#x2030; 0 you can write đ?&#x2018;Ľ 2 + x + 1 = 0. x = 0 is not a solution. In fact for x = 0 we have 1 = 0. I consider x â&#x2030; 0. I suppose x > 0 and observe đ?&#x2018;Ľ 2 + x = - 1. When x > 0 this relation is never true. In fact, we obtain that a positive real (as đ?&#x2018;Ľ 2 + x) equal to a negative integer, as - 1. Now we can consider the subcase following. For x < 0 we can write đ?&#x2018;Ľ 2 = - 1 - x = - 1 + |x|, formally correct (but only formally!) when x < 0. But when we consider ( -â&#x2C6;&#x17E;, 0) we can admit (-â&#x2C6;&#x17E; , -1] â&#x2C6;Ş (-1 , 0). We have to consider two different subcases. The first one is when x â&#x2C6;&#x2C6; (-â&#x2C6;&#x17E;, -1]. Using elementary algebraic rules you can have the following formal equality: |x | - đ?&#x2018;Ľ 2 = 1. We can observe no real solution in this one when x â&#x2C6;&#x2C6; (-â&#x2C6;&#x17E; , 0). In fact in this case the real number - |x | - đ?&#x2018;Ľ 2 = 1. What I have just written is right when we consider the second possible case, when x â&#x2C6;&#x2C6; ( 1, 0). However, we can unify these two different subcases. In fact we can write đ?&#x2018;Ľ 2 = |x||x|. We have |x| - |x||x| = 1, and then |x| (1 - |x|) = 1. This equality is not true. In fact when |x| > 1, the number |x| (1 - |x|) < 0, then, a fortiori, |x| (1 - |x|) < 1. But you can note that from |x| (1 - |x|) = 1 you have (1/s) + s where s = |x| â&#x2C6;&#x2C6; đ?&#x2018;&#x2026; + with đ?&#x2018;&#x2026; + to formalize the set of the positive real numbers. But (1/s) + s = 1+s /s > 1, for each positive real s, as definite.
Then we can observe that it is impossible to have (1+s/s) = 1 when we put s = |x| I would like to say how I can write đ?&#x2018;Ľ 2 + x + 1 = 0 from ađ?&#x2018;Ľ 2 + bx + c = 0 when you suppose a = b = c â&#x2030; 0. You can take ađ?&#x2018;Ľ 2 + bx + c = 0 and divide by a. You must remember that 0/a = 0. We obtain đ?&#x2018;Ľ 2 + x + 1 = 0 immediatly. Finally, I would to remember that đ?&#x2018;Ľ 2 = - 1 - x = - 1 + |x|, formally admissible (but only formally!) when x < 0. But there are not some x real numbers that verify this equality. Next days I will send a new speech to demonstrate there are no real solutions to quadratic equation when you consider (a, b, c) â&#x2030; (0, 0, 0) in case a>0, b>0 and c > 0 and in different case too, when you consider a < 0, b < 0 and c < 0, as Cartesius said. It is easier to solve two different cases, when we have b = 0 or c = 0. But in general we must consider a â&#x2030; 0. But the general case is the following: ađ?&#x2018;Ľ 2 + bx + c = 0 â&#x2020;&#x2019; đ?&#x2018;Ľ 2 + Îśx + Ď&#x192;c = 0 with Îś = b/a and Ď&#x192; = c/a. I will demonstrate the following statement: There are no positive real numbers, such as đ?&#x2018;Ľ0 > 0, that đ?&#x2018;Ľ0 2 +Îś đ?&#x2018;Ľ0 + Ď&#x192; c = 0 is true when Îś > 0 and Ď&#x192; > 0. I continue to discuss about inexistence of the real number đ?&#x2018;Ľ0 > 0 as solution of the second order algebraic equations. I will demonstrate the following statement: There are not positive real numbers, such as đ?&#x2018;Ľ0 > 0, that đ?&#x2018;Ľ0 2 +Îś đ?&#x2018;Ľ0 + Ď&#x192; c = 0 is true when Îś > 0 and Ď&#x192; > 0. Now I wish consider the general case, when aâ&#x2030; b â&#x2030; c â&#x2030; 0. In this case we can note there are no real solutions to quadratic equation when you consider (a, b, c) â&#x2030; (0, 0, 0) in case a>0, b>0 and c > 0 and in different case too, when you consider a < 0, b < 0 and c < 0, as Cartesius said. From aâ&#x2030; b â&#x2030; c â&#x2030; 0 we have Îś â&#x2030; Ď&#x192;. đ?&#x2018;Ľ0 â&#x2C6;&#x2C6; (0, + â&#x2C6;&#x17E;) is not a solution of đ?&#x2018;Ľ0 2 +Îś đ?&#x2018;Ľ0 + Ď&#x192; = 0 because from đ?&#x2018;Ľ0 2 +Îś đ?&#x2018;Ľ0 + Ď&#x192; = 0 we can write đ?&#x2018;Ľ0 2 = - Îś đ?&#x2018;Ľ0 - Ď&#x192;. But this equality is not true because đ?&#x2018;Ľ0 2 > 0 but - Îś đ?&#x2018;Ľ0 - Ď&#x192; < 0 when đ?&#x2018;Ľ0 â&#x2C6;&#x2C6; (0, + â&#x2C6;&#x17E;) and Îś > 0 and Ď&#x192; > 0 too. đ?&#x2018;Ľ0 = 0 is not a solution. It is very easy to demonstrate it by substitution.
Now we can consider the case đ?&#x2018;Ľ0 â&#x2C6;&#x2C6; ( - â&#x2C6;&#x17E;, 0). In fact, you can consider the following algebraic steps. đ?&#x2018;Ľ0 2 = - Îś đ?&#x2018;Ľ0 - Ď&#x192; â&#x2020;&#x2019; đ?&#x2018;Ľ0 2 + Îś đ?&#x2018;Ľ0 = - Ď&#x192; â&#x2020;&#x2019;
(đ?&#x2018;Ľ0 )<0
đ?&#x2018;Ľ0 2 - Îś |đ?&#x2018;Ľ0 |= - Ď&#x192; â&#x2020;&#x2019; |đ?&#x2018;Ľ0 ||đ?&#x2018;Ľ0 |- |đ?&#x2018;Ľ0 | Îś =
- Ď&#x192; â&#x2020;&#x2019; |đ?&#x2018;Ľ0 |(|đ?&#x2018;Ľ0 |- Îś )= - Ď&#x192;. Former equality is not true when we consider the case |đ?&#x2018;Ľ0 |> Îś. In fact in this case |đ?&#x2018;Ľ0 |(|đ?&#x2018;Ľ0 |Îś ) > 0 but â&#x20AC;&#x201C;Ď&#x192; < 0 when Ď&#x192; > 0. When we consider |đ?&#x2018;Ľ0 |= Îś = b/a we have |đ?&#x2018;Ľđ?&#x2018;&#x153; |0 = - Ď&#x192;. Hence 0 = - Ď&#x192; (impossible when Ď&#x192; â&#x2030; 0). This statement is not possible. Then đ?&#x2018;Ľ0 = - Îś = - (b/a) is not a solution. I remember I have just written đ?&#x2018;Ľ0 2 - Îś |đ?&#x2018;Ľ0 |= - Ď&#x192;. I obtained it from đ?&#x2018;Ľ0 2 = - Îś đ?&#x2018;Ľ0 - Ď&#x192; when đ?&#x2018;Ľ0 < 0. For Îś < |đ?&#x2018;Ľ0 | we note đ?&#x2018;Ľ0 2 - Îś |đ?&#x2018;Ľ0 |= - Ď&#x192; is a consistent equality when Ď&#x192; > 0. From this formal consistency we can write the following equality Ď&#x192; = Îś| đ?&#x2018;Ľ0 |- (đ?&#x2018;Ľ0 )2 . From these one we have c/a = (b/a)r - đ?&#x2018;&#x; 2 when r = |đ?&#x2018;Ľ0 |. From last equality we obtain c = br đ?&#x2018;&#x; 2 and then ađ?&#x2018;&#x; 2 - br + c = 0. For every (a, b, c) â&#x2030; (0, 0, 0) we can solve ađ?&#x2018;&#x; 2 - br +c = 0 respect to unknown r. However, we have obtain this quadratic equation considering the case r = |đ?&#x2018;Ľ0 | when đ?&#x2018;Ľ0 < 0. This polynomial f( r) = đ?&#x2018;&#x17D;đ?&#x2018;&#x; 2 - br + c is formally different (Abelâ&#x20AC;&#x2122;s identity polynomial criterion) from polynomial f(x) = đ?&#x2018;Ľ 2 + (b/a)x +(c/a). In fact if we consider the first one for x < 0 we have r = |x| â&#x2021;&#x2019; x = - r or identically r = - x. Then from f(r) =ađ?&#x2018;&#x; 2 - br + c, consistent when r > 0, we have f(-x) = ađ?&#x2018;Ľ 2 + bx + c. In general f(â&#x2C6;&#x201C;x) = f(Âą r) and f(x) = f(-r). Now I would like to consider ađ?&#x2018;&#x; 2 - br + c = 0 noting that đ?&#x2018;&#x; 2 (b/a) r + (c/a) = 0 â&#x2020;&#x2019; đ?&#x2018;&#x; 2 = (b/a) r - (c/a). I remember I have made it with a > 0, b > 0 , c > 0. I have admit the condition r > 0. Considering former equality, I observe đ?&#x2018;&#x; 2 > 0, always. We consider the case r > 0. r = b/a is not a possible solution, in fact đ?&#x2018;&#x; 2 = đ?&#x2018;&#x; 2 - (c/a) is true only and only if you have (c/a) = 0. It is impossible. r < o is not a solution. In fact for any r < o we have a positive number such as đ?&#x2018;&#x; 2 equal to a negative real number as (b/a) r - (c/a). We have a negative number because the sum of two negative numbers is a negative number (closure). From đ?&#x2018;&#x; 2 = (b/a) r - (c/a) we can say r > b/a can be a solution of this equation. But from (b/a)r - (c/a) > 0 we can also write (b/a)r > c/a â&#x2020;&#x2019; r > c/b.
If you consider these conditions you can admit r is a solution for r > max ( c/b ; b/a). Why x = - r we have just demonstrate there are no positive real solutions for quadratic equation such as ađ?&#x2018;Ľ 2 + bx + c = 0 when we admit a > 0, b > 0 and c > 0 or we admit a < 0, b < 0, c < 0. When (a, b, c) â&#x2030; (0, 0, 0) the number đ?&#x2018;Ľ0 = 0 is not a solution for this equation. We remember in this case f(x) = f(-x) and for every (a, b, c) we have lim đ?&#x2018;&#x201C;(đ?&#x2018;Ľ) = Âą â&#x2C6;&#x17E; in đ?&#x2018;Ľâ&#x2020;&#x2019; Âąâ&#x2C6;&#x17E;
according to a > 0 or a < 0. In this paragraph about quadratic equations and its possible real roots, I have just demonstrated the following statement: There are no positive real numbers, such as đ?&#x2018;Ľ0 > 0, that đ?&#x2018;Ľ0 2 +Îś đ?&#x2018;Ľ0 + Ď&#x192; c = 0 is true when Îś > 0 and Ď&#x192; > 0. It is very easy to demonstrate it in more simple way. In fact from ađ?&#x2018;Ľ0 2 + bđ?&#x2018;Ľ0 + c = 0 we obtain ađ?&#x2018;Ľ0 2 + bđ?&#x2018;Ľ0 = - c. However - c < 0 when c > 0. Moreover, when a > 0 and b > 0 for đ?&#x2018;Ľ0 < 0 last equality is not true. Similarly, we can consider the subcase c < 0. For c < 0 we have - c > 0. However, when a < 0 and b < 0 for every đ?&#x2018;Ľ0 > 0 the number ađ?&#x2018;Ľ0 2 + bđ?&#x2018;Ľ0 is a negative number but - c is, for c < 0, a positive number. Then we have just obtained an equality impossible. In fact is not possible a positive number equal to a negative number.
2. Algebraic equations in x of degree n > 2 Now, I would like to generalize this method considering a polynomial function such as P(x) = â&#x2C6;&#x2018;đ?&#x2018;&#x203A;0 đ?&#x2018;&#x17D;đ?&#x2018;&#x203A; đ?&#x2018;Ľ đ?&#x2018;&#x203A; in x of degree n. We can write P(x) = â&#x2C6;&#x2018;đ?&#x2018;&#x203A;0 đ?&#x2018;&#x17D;đ?&#x2018;&#x203A; đ?&#x2018;Ľ đ?&#x2018;&#x203A; = 0. From this one we can obtain the following: â&#x2C6;&#x2018;đ?&#x2018;&#x203A;đ?&#x153;&#x2122;â&#x2C6;&#x2C6;đ?&#x2018;&#x2039; đ?&#x2018;&#x17D;đ?&#x153;&#x2122; đ?&#x2018;Ľ đ?&#x153;&#x2122; đ?&#x153;&#x2014; = - (â&#x2C6;&#x2018;đ?&#x2018;&#x203A;â&#x2C6;&#x2019;1 đ?&#x153;&#x2014;â&#x2C6;&#x2C6;đ?&#x2018;&#x152; đ?&#x2018;&#x17D;đ?&#x153;&#x2014; đ?&#x2018;Ľ ) when n is an even positive integer. In this case we have X = (0 , 2, 4, 6, â&#x20AC;Ś.
, 2u=n) and Y = (1, 3, 5, â&#x20AC;Ś. , 2u - 1). But when you consider the case n = 2u + 1 for every positive integer u we have the đ?&#x153;&#x2014; following relation: â&#x2C6;&#x2018;đ?&#x2018;&#x203A;đ?&#x153;&#x2122;â&#x2C6;&#x2C6;đ?&#x2018;&#x2021; đ?&#x2018;&#x17D;đ?&#x153;&#x2122; đ?&#x2018;Ľ đ?&#x153;&#x2122; = - (â&#x2C6;&#x2018;đ?&#x2018;&#x203A;â&#x2C6;&#x2019;1 đ?&#x153;&#x2014;â&#x2C6;&#x2C6;đ?&#x2018;? đ?&#x2018;&#x17D;đ?&#x153;&#x2014; đ?&#x2018;Ľ ) . In this case we have T = (1, 3, 5, â&#x20AC;Ś. , n)
and Z = (0, 2, 4, â&#x20AC;Ś.. , n - 1).
When you consider the first one you can observe that if đ?&#x2018;&#x17D;đ?&#x153;&#x2122; > 0 and đ?&#x2018;&#x17D;đ?&#x153;&#x2014; > 0 last equality can be
right
if
x
<
0.
For
x
>
0
this
one
cannot
be
true
because
in
general a positive number such as â&#x2C6;&#x2018;nĎ&#x2022;â&#x2C6;&#x2C6;T aĎ&#x2022; x Ď&#x2022; cannot be equal to a negative number such as Ď&#x2018; - (â&#x2C6;&#x2018;nâ&#x2C6;&#x2019;1 Ď&#x2018;â&#x2C6;&#x2C6;Z a Ď&#x2018; x ). In a similar way we can demonstrate that x > 0 cannot be a solution (root)
when you consider the special case for n = 2u + 1 with u an integer positive number. If you consider these cases you can admit it exists one or more numbers, e.g. Ď&#x2022; or Ď&#x2018; we have đ?&#x2018;&#x17D;đ?&#x153;&#x2014; = 0 or đ?&#x2018;&#x17D;đ?&#x153;&#x2122; = 0. In fact if you consider polynomial function such as P(x) = đ?&#x2018;&#x17D;đ?&#x2018;&#x203A; đ?&#x2018;Ľ đ?&#x2018;&#x203A; + đ?&#x2018;&#x17D;đ?&#x2018;&#x203A;â&#x2C6;&#x2019;1 đ?&#x2018;Ľ đ?&#x2018;&#x203A;â&#x2C6;&#x2019;1 + đ?&#x2018;&#x17D;đ?&#x2018;&#x203A;â&#x2C6;&#x2019;2 đ?&#x2018;Ľ đ?&#x2018;&#x203A;â&#x2C6;&#x2019;2 + â&#x20AC;Śâ&#x20AC;Ś + đ?&#x2018;&#x17D;0 đ?&#x2018;Ľ 0 . We consider P(x) = 0 with đ?&#x2018;&#x17D;đ?&#x2018;&#x203A; â&#x2030; 0 to have an algebraic equation in x of degree n. From P(x) = đ?&#x2018;&#x17D;đ?&#x2018;&#x203A; đ?&#x2018;Ľ đ?&#x2018;&#x203A; + đ?&#x2018;&#x17D;đ?&#x2018;&#x203A;â&#x2C6;&#x2019;1 đ?&#x2018;Ľ đ?&#x2018;&#x203A;â&#x2C6;&#x2019;1 + đ?&#x2018;&#x17D;đ?&#x2018;&#x203A;â&#x2C6;&#x2019;2 đ?&#x2018;Ľ đ?&#x2018;&#x203A;â&#x2C6;&#x2019;2 + â&#x20AC;Śâ&#x20AC;Ś +đ?&#x2018;&#x17D;0 đ?&#x2018;Ľ 0 we can write (dividing by đ?&#x2018;&#x17D;đ?&#x2018;&#x203A; â&#x2030; 0 ) P(x) = đ?&#x2018;Ľ đ?&#x2018;&#x203A; + đ?&#x153;&#x152;đ?&#x2018;&#x203A;â&#x2C6;&#x2019;1 đ?&#x2018;Ľ đ?&#x2018;&#x203A;â&#x2C6;&#x2019;1 + đ?&#x153;&#x152;đ?&#x2018;&#x203A;â&#x2C6;&#x2019;2 đ?&#x2018;Ľ đ?&#x2018;&#x203A;â&#x2C6;&#x2019;2 + â&#x20AC;Ś.. + đ?&#x153;&#x152;0 = 0, when đ?&#x153;&#x152;đ?&#x2018;&#x2013; = đ?&#x2018;&#x17D;đ?&#x2018;&#x2013; / đ?&#x2018;&#x17D;đ?&#x2018;&#x2013;â&#x2C6;&#x2019;1 for i positive integer and i â&#x2030;¤ n-1. Thus đ??ź0 = { n-1, n-2, â&#x20AC;Śâ&#x20AC;Ś. , 0 }. I can consider the case đ?&#x153;&#x152;đ?&#x2018;&#x2013; < 0 for one and only one i < n. I consider the case đ?&#x153;&#x152;0 â&#x2030; 0. Hence đ?&#x2018;Ľ đ?&#x2018;&#x203A; + đ?&#x153;&#x152;đ?&#x2018;&#x203A;â&#x2C6;&#x2019;1 đ?&#x2018;Ľ đ?&#x2018;&#x203A;â&#x2C6;&#x2019;1 + đ?&#x153;&#x152;đ?&#x2018;&#x203A;â&#x2C6;&#x2019;2 đ?&#x2018;Ľ đ?&#x2018;&#x203A;â&#x2C6;&#x2019;2 +đ?&#x153;&#x152;ĂŹ+1 đ?&#x2018;Ľ ĂŹ+1 + đ?&#x153;&#x152;ĂŹâ&#x2C6;&#x2019; 1 đ?&#x2018;Ľ ĂŹâ&#x2C6;&#x2019;1 +â&#x20AC;Ś.. + đ?&#x153;&#x152;0 = |đ?&#x2018;&#x17D;đ?&#x2018;&#x2013; |đ?&#x2018;Ľ đ?&#x2018;&#x2013; for 0 < i < n. I observe đ?&#x153;&#x152;đ?&#x2018;&#x203A;â&#x2C6;&#x2019;1 đ?&#x2018;Ľ đ?&#x2018;&#x203A;â&#x2C6;&#x2019;1 + đ?&#x153;&#x152;đ?&#x2018;&#x203A;â&#x2C6;&#x2019;2 đ?&#x2018;Ľ đ?&#x2018;&#x203A;â&#x2C6;&#x2019;2 + â&#x20AC;Ś.. + đ?&#x153;&#x152;0 > 0 when we consider x > 0. But from it I can write đ?&#x2018;Ľ đ?&#x2018;&#x203A; < |đ?&#x153;&#x152;đ?&#x2018;&#x2013; |đ?&#x2018;Ľ đ?&#x2018;&#x2013; â&#x2020;&#x2019; đ?&#x2018;Ľ đ?&#x2018;&#x203A;â&#x2C6;&#x2019;đ?&#x2018;&#x2013; < |đ?&#x153;&#x152;đ?&#x2018;&#x2013; | â&#x2020;&#x2019; x < đ?&#x2018;&#x203A;â&#x2C6;&#x2019;đ?&#x2018;&#x2013;â&#x2C6;&#x161;|đ?&#x153;&#x152;đ?&#x2018;&#x2013; |. Thus if đ?&#x2018;Ľ0 > 0 is a postitive real number to have P(đ?&#x2018;Ľ0 ) = 0, called a zero of P(x), I have just obtained đ?&#x2018;Ľ0 â&#x2C6;&#x2C6; (0 ,
đ?&#x2018;&#x203A;â&#x2C6;&#x2019;đ?&#x2018;&#x2013;
â&#x2C6;&#x161;|đ?&#x153;&#x152;đ?&#x2018;&#x2013; |). When the solution exists this one is a number conteined in
former open interval. But we can generalize. In fact we can consider the case there are two or more đ?&#x153;&#x152;đ?&#x2018;&#x2DC; < 0. We can consider the case đ?&#x153;&#x152;đ?&#x2018;&#x2013; < 0 and đ?&#x153;&#x152;đ?&#x2018;&#x2014; < 0. For the same reason generally we have đ?&#x2018;Ľ đ?&#x2018;&#x203A; < |đ?&#x153;&#x152;đ?&#x2018;&#x2013; |đ?&#x2018;Ľ đ?&#x2018;&#x2013; + |đ?&#x153;&#x152;đ?&#x2018;&#x2014; |đ?&#x2018;Ľ đ?&#x2018;&#x2014; . A fortiori đ?&#x2018;Ľ đ?&#x2018;&#x203A; < min ( |đ?&#x153;&#x152;đ?&#x2018;&#x2013; |đ?&#x2018;Ľ đ?&#x2018;&#x2013; , |đ?&#x153;&#x152;đ?&#x2018;&#x2014; |đ?&#x2018;Ľ đ?&#x2018;&#x2014; ). Obviously, we can study the case when there are more than two real parameters đ?&#x153;&#x152;đ?&#x2018;&#x2DC; < 0.
publication by Pascal McLee
2014 Š mclee consulting | web solutions. all rights reserved. web www.pascalmclee.com - mail pas.meli@gmail.com - mob. +39 335 6856486