Ask A Scientist

Mathematics Archive


Solving equations with fractional exponents Newton's Method -Exponents



Question: I have been asked how to solve X^(.6)-X=Y by a teacher at a 
nearby school. I could not figure out how to solve for X with a fractional 
exponent.  Can anyone help?
------------------------------------------------
Except for very special exponents and equations, there are no 
analytic solutions for this type of equation.  You have to solve this kind of 
equation numerically.  However, by doing some plotting, you can find out lots 
of things about the solutions.  If you plot x^(0.6)-x as a function of x, you 
will find a maximum at about x=0.8154 and the value for the maximum is about 
0.0694.  If we restrict our search to positive x, (we do not want complex 
solutions) we find that there is no solution if y>0.0694.  There are two 
solutions if 0 < y < 0.0694 and one solution if y<0.  There is an old formula for 
calculating the square root of a number y.  It is an iterative formula, 
meaning that you plug the value of x into it over and over to get the answer.  
So, to get the square root of y, do the following over and over until you get 
bored.  Guess a value of x to start with.  Take half of x and add to it half 
of y divided by x.  Take the resulting value of x and apply it to the formula 
again.  Keep repeating this, and in about 5 tries for any x, you will have the 
square root to many significant figures.  For this formula and for negative y, 
try the following iterative formula.  The new x = (0.4*x^(0.6)-y)/(1-0.6*x^(-
0.4)), keep iterating the x value and it should be very convergent.  Your 
first guess should be bigger than .8154.  This comes from calculus and is 
related to a method called Newton's method.  I hope this is  helpful.
Samuel B. Bowen
=========================================================



Back to Mathematics Ask A Scientist Index
NEWTON Homepage Ask A Question

NEWTON is an electronic community for Science, Math, and Computer Science K-12 Educators.
Argonne National Laboratory, Division of Educational Programs, Harold Myron, Ph.D., Division Director.