Given a parabola y = ax^2 + bx + c find the slope of the parabola at the point (x, y) without using derivatives or any limits? is it possible to generalize for any function f(x)?
3 Answers
Yes, we can find it using infinitesimals...
Explanation:
We can find the slope using infinitesimals.
Let
Let
Then the slope at
(f(x+epsilon) - f(x))/((x+epsilon) - x)
= ((a(x+epsilon)^2+b(x+epsilon)+c)-(ax^2+bx+c))/epsilon
= ((a(x^2+2epsilonx+epsilon^2)+b(x+epsilon)+c)-(ax^2+bx+c))/epsilon
= ((ax^2+(b+2epsilona)x+(c+bepsilon+aepsilon^2))-(ax^2+bx+c))/epsilon
= (2epsilonax+(bepsilon+aepsilon^2))/epsilon
= 2ax+b+aepsilon
The standard part is obtained by discarding the infinitesimal part
The same method works for any differentiable function
See https://socratic.org/s/aLLn6ic7 for more discussion.
See below.
Explanation:
Giving two points pertaining to the parabola, we have the slope between the points
now making
Here's another method based on geometrical considerations...
Explanation:
Here's another method, without limits (explicit or not) or infinitesimals...
Given:
f(x) = ax^2+bx+c
Suppose
So:
y_1 = ax_1^2+bx_1+c
The equation of a non-vertical line through
y-y_1 = m(x-x_1)
i.e.:
y = mx+(y_1-mx_1)
This will be a tangent to the parabola if and only if the only intersection with the parabola is at
To find the points of intersection, we want to solve the system of equations:
{ (y = ax^2+bx+c), (y = mx+(y_1-mx_1)) :}
So:
ax^2+bx+c = y = mx+ax_1^2+bx_1+c-mx_1
That is:
a(x^2-x_1^2)+b(x-x_1)-m(x-x_1) = 0
i.e.:
(x-x_1)(a(x+x_1)+b-m) = 0
So if
a(x+x_1)+b-m = 0
If the only solution is
2ax_1+b-m = 0
So:
m = 2ax_1+b
We can potentially use a similar method with other functions, but the tangent to a parabola is guaranteed to touch the curve just once.