Is their any way to calculate/predict/estimate variation of YS/UTS with rolling direction? Is it possible to calculate using the texture data EBSD/XRD?
As far as I know, theory is not yet that developed to the point to allow this kind of prediction. I am aware of some works by the group of Leo Kestens in Gent which were successful in modeling texture development up to low levels of strain (so, the inverse problem). If you have the Lankford coefficients (r) it is pretty straightforward to predict the variation of YS using Hill's yielding law, this is done in his first work (http://dx.doi.org/10.1098/rspa.1948.0045).
You could start with the assumption of isotropic material, provide the stress-strain curve of the parent material and simulate the rolling process. By using the appropriate constitutive material models, you will be able to predict the rolling-induced material anisotropy, residual stresses and the change in the stress-strain curve. This will not give you though the variation of the UTS.
I am not referring to a specific publication. Use FEA to simulate the rolling process. Start with the initial geometry and stress-free material. Give as input the stress-strain curve of the parent material using a material model e.g Chaboche to aproximate it. Then simulate in a FEA package (e.g. Abaqus, Ansys etc.) the rolling process. The model should give you the effect of rolling on the plate by predicting the material yielding/plastic deformation. It is basically a cold-forming process if I get your question right.