Partial differentials?

Are you sure that the answer they gave you was not u(x,t)=cos(a*x)exp(-k*(a^2)*t)? That would actually work.