Actually, to first order Isp is independent of molecular weight, for a given mass flow rate and combustion power. You can show this with the equations of momentum and energy: F = mdot*v_exh, eta*P = 0.5*mdot*v_exh^2. Notice how molecular weight doesn't show up?
The stoichiometric mixture ratio for hydrolox is O/F = 8. Hydrolox engines always run fuel-rich.
Its my understanding that the exhaust velocity is inversely proportional to the average molecular wt of the gases. Perhaps that is where the MW enters your equations (hidden in the Ve term).
In any case, low mol wt is desirable. My query was regarding variation in the mixture ratio. Is there any benefit in continuous variation?I expect its complex to implement, but would there be any theoretical basis for such a practice?