a)
By definition, normal cone for a set X at point x is defined in such a way:
NX(x)={z∣∃xk⊂X,{zk} s.t. zk∈TX(xk)∗,xk→x,zk→z}
Here TX(x) is the tangent cone to the set X at point x. It is defined in such a way:
TX(x)={0}∪{y∣y=0,∃{xk}⊂X s.t. xk=x and xk→x,∥xk−x∥xk−x→∥y∥yk}
Without losing generality, let x be a point at the origin and circular region C lies in the upper half-plane. Since the tangent cone equals to
TX(x)={(x,y)∣y>0} the normal cone equals to
NX(x)={(x,y)∣y=0}
(b)
In case C is rectangular region in R2, x is point on the boundary, then as in (a)
TX(x)={(x,y)∣y>0}NX(x)={(x,y)∣y=0}