“…Since (x, y) is a proper minimiser of (P n ), it has to be a minimiser of (P n ) with respect to some K £ K. Therefore, it follows from Lemma 3.2 and the imposed conditions that for such K G /C, we have It is clear that in Theorem 3.1 we have not imposed any differentiability assumption on the map H. Thus it would be of interest to obtain a variant of the well-known theorem of Lyusternik ( [10]), so that the cone M* contain information about some derivative of H. In fact, this is completely true if H is single-valued and sufficiently smooth ( [17]). Moreover, we can also define a variant of the generalised contingent epiderivative (see [3,8,11]) by taking the minimal points of Di(F + C)(x,y) with respect to the cone C. We mention that our results will remain valid for such an epiderivative.…”