Multi-Criteria Optimization of Refinery

Vector optimization of refinery is discussed. As the target function, a multi-level hierarchy criterion convolution is used. This convolution is formed following the recurrent procedure based on a scalar invariant representing a linear combination of Hölder norms: the first order norm and the sup-norm. The issue of matrix representation of the suggested invariant has been analyzed; the procedure of building the respective matrix structure has been developed. The approach has been implemented in a respective package of application software programs