That problem is a bit more involved than it might otherwise appear. I used the patch function because I can make it do what I want:
x=linspace(0, 18);
y=x.^2;
y2=x.*(-1./4)+(9./2);
y3=x.*0;
isct_y_y2 = @(x) x.^2 - (x.*(-1./4)+(9./2));
isct_y2_y3 = @(x) x.*(-1./4)+(9./2);
isct_y_y3 = @(x) x.^2;
x_y_y2 = fzero(isct_y_y2, 1);
x_y2_y3 = fzero(isct_y2_y3, 1);
x_y_y2_ix = find(x <= x_y_y2, 1, 'last');
x_y2_y3_ix = find(x <= x_y2_y3, 1, 'last');
%Fill in the Shape
%Figure
plot(x,y, x,y2, x,y3); grid on;
axis([-1,20,-1,5])
patch([x(1:x_y_y2_ix) x(x_y_y2_ix:x_y2_y3_ix) flip(x(1:x_y2_y3_ix))], [y(1:x_y_y2_ix) y2(x_y_y2_ix:x_y2_y3_ix) flip(y3(1:x_y2_y3_ix))], 'y')
