# Solve not satisfying simple boundary conditions

I am trying to solve the following equations for the coefficients c1 and c2.

w[x_] := c1*(1 - x)
a[x_] := c2*(1 - x)


Given the constraints:

Solve[(1 + w[x]*a[x]) == (1/x) && w[0] == wmax && a[1] == 0 , {c1, c2}]


Mathematica gives me:

{{c1->wmax,c2->1/((x-x^2) wmax)}}


However when plugging the formula for c2 into a, the condition a[1] == 0 is not satisfied. Instead of being zero, a[1] == 1/wmax.

How can I convince Solve to generate a c2 that truly makes a[1] == 0?

-
Welcome to Mathematica.SE! I have formatted the post for better readability. Please click the edit link above to see how to do this for your next post. –  Szabolcs Dec 14 '12 at 0:06
As you can see from the syntax colouring the the notebook interface, C is a defined system symbol. It is used for several purposes by the system. Make sure you never use a system symbol where you need an undefined one! Even if it hasn't caused trouble in this example, it will sooner or later. (Actually there was a post a couple of days ago where using C did cause problems.) In general, it is good practice never to use symbols starting with capitals. All system symbols start with a capital, so if your own ones start with a lowercase, you can be sure you won't run into this kind of trouble. –  Szabolcs Dec 14 '12 at 0:11
Thanks very much Szabolcs, I've replaced C with x in the post (the problem persists though). –  Miles Dec 14 '12 at 0:16
Given the definitions of w[x] and a[x] the equation a[1] == 0 is identically true, therefore no condition for c2 is needed. –  Artes Dec 14 '12 at 0:29
Someone will surely post a more mathematically sound answer, but the problem is the following: notice that a[1] == 0 holds for any constant c2 (i.e. independent of x). The remaining two conditions tell you that c1 and c2 must depend on x the way you posted in your answer. This c2 == 1/(wmax x - wmax x^2) is undefined for x==1 (1/0), but taking the limit x -> 1 gives a[x] -> 1/wmax (and not 0). It's a good example of a simple problem where you can't just blindly input equations into a CAS and look at what comes out. You need to be careful and look at what is happening. –  Szabolcs Dec 14 '12 at 0:31

(1) it works for me.

w[x_] := c1*(1 - x)
a[x_] := c2*(1 - x)

eqns = (1 + w[x]*a[x]) == (1/x) && w[0] == wmax && a[1] == 0;
sol = Solve[eqns, {c1, c2}]

(* {{c1 -> wmax, c2 -> -(1/(wmax (-x + x^2)))}} *)

Simplify[eqns /. sol]

(* {True} *)


(2) a[1] evaluates to 0 independently of solution values.

-
Thanks Daniel, I am still a little confused, if I substitue c2 into a I have: a[x_] := (1 - x)*-1/(wmax (-x + x^2)) a[1] = Indeterminate And the limit approaches 1/wmax as x->1. –  Miles Dec 14 '12 at 0:43