At the equlibrium position, the spring is streched a distance b and the tension is mg, so kb^3 = mg.
If we let the PE = 0 at z = 0, the initial energy of the system (KE + PE) = 0.
At displacement z (positive upwards) the PE (gravity + spring) is mgz + kz^4/4
At the extremes of the oscillation the velocity is 0, so the KE = 0, and KE + PE = 0 (conservation of energy from the starting condition), so the PE = 0.
So at the extremes mgd + kd^4/4.
Sp d = 0 (duh - we knew that already) or d^3 = -4mg/k = -4b^3.
You can write down the equation of motion as total energy = constant, i.e. 1/2 m(z-dot)^2 = -mgz + kz^4/4. But how you solve that to get the period, I dunno. Maybe you can express the solution as an an elliptic function or something since there's something that looks like sqrt(a + bz^3) in there. I don't know if that's a sensible suggestion or not.