Writing pow(a,x) in cvx when x is variable and a

2

Does using
a^x = exp(x*log(a))
help you? CVX will accept the RHS if a > 0 and x is a variable or convex expression.

But your first task is to prove that your optimization problem is convex.