# Problem with equations like ax^2+bx+c = 0

### #1

Posted 28 June 2004 - 03:20 PM

I put some equantion in solve, and some results are correct others are wrong !!

Could anybody explain to me what is going wrong ?

### #2

Posted 28 June 2004 - 03:42 PM

check also that your variables haven't wrong values in memory..

### #3

Posted 28 June 2004 - 05:37 PM

**Clear_a_z**. If this doesn't work give some examples...

### #4

Posted 28 June 2004 - 07:07 PM

When we put in SOLVE MENU the equation x^2-4 = 0 , the result should be x= -2 and x=2 , but in result oly appears x=2 , WHY ??

### #5

Posted 28 June 2004 - 07:56 PM

I get {x=-2,x=2}When we put in SOLVE MENU the equation x^2-4 = 0 , the result should be x= -2 and x=2 , but in result oly appears x=2 , WHY ??

try

**solve(x^2-4=0)**

### #6

Posted 28 June 2004 - 08:06 PM

But why don't appers the same in the solve menu ?

### #7

Posted 28 June 2004 - 08:24 PM

What do you mean by "solve menu"? Do you mean the NumSolve application? If so, the NumSolve application solves numerically, not symbolically. You won't get all roots. There is't even a guarantee you'll get an exact solution.In Main Menu i appers {x=-2, x=2} !!

But why don't appers the same in the solve menu ?

In the NumSolve application you have to give an initial guess (x=) and it will attempt to find the closest solution. Enter x=1 and it will find the solution x=2. Enter x=-1 and it will find the solution x=-2.

If you want a complete, symbolic solution you should try using "solve(" in MAIN first.

### #8

Posted 28 June 2004 - 08:43 PM

=>

1. it can found only 1 value at one time

2. you can't be sure it will find a solution even if there are solutions (for a success with n-r you must satisfy a mathematical condition ; most time it will be satisfied if your guess is relatively "close" to the solution or if the function is "regular" enough

3. sometimes it can give a erratic solution... don't forget to check the solution

but, it can estimate solutions that the calc or even we can't find by symbolic methods

(and it can also be useful for teachers to test the students at an numeric methods exam :-/ grrrrr...)

### #9

Posted 02 July 2004 - 11:24 AM

### #10

Posted 02 July 2004 - 04:51 PM

I just wanted to make the point that a numeric solver isn't guaranteed to get a correct solution. People often expect that a calculator must always give the correct answer or it is some type of bug. A numeric root solver is just riding the curve trying to find the answer.

#### 0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users