We consider the polynomial approximation behavior of the problem of finding
, in a graph with weighted vertices, a maximal independent set minimizing t
he sum of the weights. In the spirit of a work of Halldorson dealing with t
he unweighted case, we extend it and perform approximation hardness results
by using a reduction from the minimum coloring problem. In particular, a c
onsequence of our main result is that there does not exist any polynomial t
ime algorithm approximating this problem within a ratio independent of the
weights, unless P = NP. We bring also to the fore a very simple ratio rho g
uaranteed by every algorithm while no polynomial time algorithm can guarant
ee the ratio (1 - epsilon)rho. The known hardness results for the unweighte
d case can be deduced. We finally discuss approximation results for both we
ighted and unweighted cases: we perform an approximation ratio that is vali
d for any algorithm for the former and propose an analysis of a greedy algo
rithm for the latter.