That is impossible, since your equations are not written in unambiguous mathematics.
Is e^15b intended to mean exp(15*b)?
Is e^15b intended to mean exp(15)*b?
Is e the base of the natural logarithem? Thus
format long
exp(1)
Even if as I suspect, you actually have these equations:
syms R b
E1 = R*exp(15*b) == 35.75;
E2 = R*exp(25*b) == 29.125;
E3 = R*exp(35*b) == 22.875;
then no unique solution can possibly exist. That is, divide equation 2 by equation 1. That leaves us with
E2/E1
As you can see, R goes away, and we could compute b from that relation.
vpa(solve(E2/E1))
However, suppose we chose to do this instead?
E3/E1
vpa(solve(E3/E1))
And finally, we could do this:
vpa(solve(E3/E2))
Which as you see, returns a completely different result for b. And when we do so, depending on the value of b we chose from how we might derive it, then R is also completely different.
