This paper intents to show the ties between the results of solving lin
ear programs and the methods of allocation of costs in cost accounting
. It refers to the method of Kaplan and Thomson and shows its weakness
. It extents this method by using pseudo-dual variables defined for an
y production instead of dual variables which, in the method of Kaplan
and Thomson, are just defined for an optimal production in the sense o
f profit maximization, rarely the case for an actual production.