A perturbative QCD based jet tomographic Monte Carlo model, CUJET2.0, is presented to predict jet quenching observables in relativistic heavy ion collisions at RHIC/BNL and LHC/CERN energies. This model generalizes the DGLV theory of flavor dependent
radiative energy loss by including multi-scale running strong coupling effects. It generalizes CUJET1.0 by computing jet path integrations though more realistic 2+1D transverse and longitudinally expanding viscous hydrodynamical fields contrained by fits to low $p_T$ flow data. The CUJET2.0 output depends on three control parameters, $(alpha_{max},f_E,f_M)$, corresponding to an assumed upper bound on the vacuum running coupling in the infrared and two chromo-electric and magnetic QGP screening mass scales $(f_E mu(T), f_M mu(T))$ where $mu(T)$ is the 1-loop Debye mass. We compare numerical results as a function of $alpha_{max}$ for pure and deformed HTL dynamically enhanced scattering cases corresponding to $(f_E=1,2, f_M=0)$ to data of the nuclear modification factor, $R^f_{AA}(p_T,phi; sqrt{s}, b)$ for jet fragment flavors $f=pi,D, B, e$ at $sqrt{s}=0.2-2.76$ ATeV c.m. energies per nucleon pair and with impact parameter $b=2.4, 7.5$ fm. A $chi^2$ analysis is presented and shows that $R^pi_{AA}$ data from RHIC and LHC are consistent with CUJET2.0 at the $chi^2/d.o.f< 2$ level for $alpha_{max}=0.23-0.30$. The corresponding $hat{q}(E_{jet}, T)/T^3$ effective jet transport coefficient field of this model is computed to facilitate comparison to other jet tomographic models in the literature. The predicted elliptic asymmetry, $v_2(p_T;sqrt{s},b)$ is, however, found to significantly underestimated relative to RHIC and LHC data. We find the $chi^2_{v_2}$ analysis shows that $v_2$ is very sensitive to allowing even as little as 10% variations of the path averaged $alpha_{max}$ along in and out of reaction plane paths.
