I have a question regarding the Arrow-Enthoven-Theorem which is used to find sufficient conditions for a global maximum of a quasi-concave function under quasi-convex restrictions also see: http://michaelcarteronline.com/FOME/pdf/ConstraintQualification.pdf, P. 4). In "Fundamental Methods of Mathematical Economics" from Chiang and Wainwright page 426 it is stated that for one condition: "(d) any one of the following is satisfied:

(d-i) $f_j (x^*) <0 $ for at least one variable $x_j$

(d-ii) $f_j (x^*)>0$ for some variable $x_j$ that can take on a positive value without violating the constraints

(d-iii) the $n$ derivatives $f_j(x^*)$ are not all zero, and the function $f(x)$ is twice differentiable in the neighbordoof of $x^*$ (i. e. all the second-order partial derivatives of $f(x)$ exist at $x^*$)"

In this case $f(x_j) $ is the objective function with $x_j =1, 2, 3, ..., n$ and $f_j$ is the partial derivative of $\partial f/ \partial x_j $. The value $x^*$ is the optimal value which maximizes $f(x)$.

My problem is the following: Which conditions of the three above can I use if I only have a function with one variable, namely $x_1$? The first condition can't be satisfied because the derivative of $f$ in $x^*$ should be zero and the same goes for the second condition. In the third condition I don't have any other derivatives for $f$ because I only have one $x$ value. Although the function is twice differentiable in the neighborhood of $x^*$. Maybe I can only use this last statement to meet the necessary conditions for the Arrow-Enthoven Theorem?

Thank you for your comments.



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.