As with in the first priority there are deviational variables
measured in different units we have to use some normalization
technique to overcome in commensurability .We will use percentage normalization,i.e. we will transform the goal function so as to
divide deviation variables set in the same priority level by their
target values . Here the decision maker wants to make minimization
of exceeding costs more important than minimization of protein
excess. Weight coefficients(within the same goal)are therefore
introduced into the goal function so that costs are three times more
important. In this way the model becomes a combination of
weighted and lexicographic goal programming(see Tamiz et al.
1998). Finally the objective function of the goal programming
model for the scenario D has the following form: