Some recent analytical techniques, including energy balance method, max-min method, and so on, work well for single degree of freedom conservative oscillators. Is it possible to extend them to the more complicated problem of multi- degrees of freedom oscillators (for example double pendulum for small oscillations)?