IIRC alcohol (EtOH) agonizes GABAa primarily, AMPARs and KARs at extremely (near OD) levels, plus induces a switch from inhibitory MOR activity to excitatory, presumably fucking with G-protein coupling but that last is a total guess.
I do know that it causes glutamatergic hyperactivity plus opioidergic nastiness of some description, alcohol withdrawal at least sensitizes (murine model) animals to the effects of opioid WD, and makes the WD worse.
I'd suggest barbiturates for the alcohol addiction, or a combination of very low dose chlormethiazole (barb site ligand , no glutamatergic action, pure clean barb-site agonist) plus an intermediate acting strong barb like amobarbital or better still, pentobarb, because barbiturates are both GABAa agonists that can direcly open the chloride channel of GABAaRs in the absence of GABA, and also they are AMPAr antagonists, that undergo a type of ligand-trapping within the AMPAr receptor 'clamshell' region, that leaves the barbs stuck there, and leaving the bound AMPARs in a closed, agonist-bound, ligand trapped state.
Adding memantine would be a good idea to leave NMDA activity as it should be relatively alone, ignoring the feedforward impulse propagation from AMPAr stimulation, and addition of perampanel, telampanel, tezampamel etc would unlikely be a bad idea.