Refining calculations of electric dipole moments in the supersymmetry and multi-Higgs models

I calculate fermion electric dipole moments generated by T- and P-odd WW{dollar}gamma{dollar} interactions in the multi-Higgs and supersymmetry models without using an approximation first introduced by Marciano and Queijeiro. In essence, this approximation consists of ignoring the details of the high energy physics responsible for the W electric dipole moment. For the multi-Higgs model, my more exact results are roughly 3 times those obtained by way of the above mentioned approximation. In the case of supersymmetry, I find a similar enhancement for gaugino masses larger than {dollar}msb{lcub}w{rcub}.{dollar} However, if the gaugino masses are {dollar}sbsp{lcub}sim{rcub}{lcub}<{rcub}msb{lcub}w{rcub},{dollar} my results are less than would be expected from Marciano and Queijeiro's estimate. Due, in part, to this suppression, I discover that the experimental bounds on {dollar}dsb{lcub}n{rcub}{dollar} place no restrictions, whatsoever, on the allowed masses of the minimal supersymmetry model. This contradicts the findings of Vendramin who used Marciano and Queijeiro's results to deduce such prohibited regions of parameter space and mildly improves the prospects for observing CP-violating W boson interactions in future accelerator experiments.; I also perform the complete 2-loop calculation of the W boson electric dipole moment in the multi-Higgs model. My result is approximately 1 order of magnitude larger than that obtained by He and McKellar in a previous calculation. The difference between these two results is due in part to contributions from diagrams neglected by He and McKellar and in part to discrepancies between our results for the diagram that they do consider. I thus find that W-induced fermion electric dipole moments are roughly 30 times those expected from their previous multi-Higgs calculation.; Finally I present two Mathematica programs, FeynmanParameter and trace, that automate the evaluation of Feynman diagrams with loops. FeynmanParameter converts integrals over loop momenta into integrals over Feynman parameters. If the input for FeynmanParameter contains any traces these are first evaluated by the program trace. Both trace and FeynmanParameter can work in an arbitrary number of dimensions, if desired--even if the traces to be evaluated contain {dollar}gammasb5{dollar}'s.