considers the question of capacity implicitly by including homogenous group members and
excluding the heterogeneous provinces from the computation of targets.
However, the exclusion of those provinces which are at a lower level of development from the
computation of targets may be arguable. Generally it may be argued that the disparities between
provinces are of two kinds, external and internal. By external we mean that a province may
have been developed disproportionately at the expense of another province being left behind,
and by internal we mean that a province may have been developed disproportionately in a few
aspects at the expense of being left behind in other aspects. Therefore one can suggest that
computing the targets on the basis of the actual values belonging to better off provinces may be
arguable. However, as homogeneity is based on all selected indicators of social and economic
aspects one hopes that the extent of this bias in the computed targets would be limited.
A project selection model for reducing regional disparities
If the government would wish to pursue the policy of reducing regional disparities, the targets
computed by the above procedure would be helpful in formulating appropriate policies in a
variety of ways. One approach consists of including these targets in a mixed integer
programming zero one project selection model in the form of a set of constraints. For example a
cost minimisation model of this form for a single province may look as follow:
J nj
Minimise Z =∑∑cijXij
j=1 i=1
Subject to: ∑ajXj ≥ Tj for j=1,2,...,J.
i =1
nj
∑ljxj ≤ Lj for j=1,2,∙∙∙,j (5)
i =1
1, if the ith project related to the jth indicator is selected,
Xij =
0, if the ith project related to the jth indicator is not selected.
Where:
cij=the cost of implementing project i related to indicator j,
Xij=the ith project related to the jth indicator,
12