International Conference of Computational Methods in Sciences and Engineering (ICCMSE), Athens, Greece, 20 - 23 March 2015, vol.1702
In this work, Enhanced Multivariance Products Representation (EMPR) for bivariate functions is considered specifically. In order to approximate a function involving two independent variables, zeroth order EMPR approximant is taken into account. Due to this aim, univariate support functions are generated on a rectangular geometry using an optimization process. The area of the relevant rectangle is shrinked to zero value limit to increase the efficiency of the corresponding EMPR approximation with the help of a new technique, that is, geometric separation.