I'm studying a statistical mechanics problem and I have two conserved quantities:
$$ E = \sum_{k=0}^{M} \left[ a_1^2(k) + a_2^2(k) + b_1^2(k) + b_2^2(k)\right] $$
$$ H = \sum_{k=0}^{M} 2 k \left[ a_1^2(k) - a_2^2(k) + b_1^2(k) - b_2^2(k)\right] $$

Is there a way to know analytically the probability density of the system of $ \{ a_1(k) ; a_2(k) ; b_1(k) ; b_2(k) \} $ to have $a_1(k) = A$?