Hessian procedure

The syntax of this function is:

 
Hess=INTERVAL_MATRIX IntervalHessian (int l1,int l2,INTERVAL_VECTOR & in)

This procedure should return an interval matrix of size m $\times$ n, m in which the Hessian of function numbered l1 to l2 has been updated (function number start at 1). The Hessian matrix of function $i$ (which is of size n $\times$ m) is stored at location Hess(($i$-1)m+1$\ldots$$i$m,1$\ldots$m). Remember that for each function the Hessian matrix is symmetric: this fact should be used in order to speed up the evaluation of this matrix. If a function in the system is not $C^2$ you set all the elements of its hessian matrix to the interval [-1e30,1e30]. Remember also here to verify that each element of the Hessian should be interval-valuable (see section 2.1.1.3).