Implement branching ratio information

[redeboer]

Is your feature request related to a problem? Please describe. The QuickStart mentions that it's not yet possible to select solutions based on branching ratio and that you have to decide about this yourself by setting the allowed intermediate states.

Describe the solution you'd like It would be cool if PDG info on branching ratios is somehow integrated in the graphs so that you can can select on that as well. May be related to ComPWA/expertsystem#23.

[spflueger]

For clarification: I implemented this "allowed intermediate states" feature mainly to reduce the amount of possible solutions. Because some decay channels are very general and many known particles would match. The J/Psi -> gamma pi0 pi0 is a good example here. Since in the gamma pi0 Subsystem a rho0 state is actually more probable than a omega state according to pure interaction strength logic. The thing is that branching ratios are missing in the expert system. So a rho0 has only a small branching ratio to gamma pi0 (super dominantly decays into pipi) as compared to the omega which has 10% for gamma pi0.

Having branching fractions would be cool, but note that its quite difficult... Either you can estimate them via the expert system. Check all possible final states and group them according to interaction type and phase space volume and estimate the branching ratio. Alternatively for well known states the branching ratios can be extracted from the pdg.

[redeboer]

In addition, I think this is fundamentally a problem with the helicity formalism (using BW as resonance descriptions): this works only if poles are separated, therefore you can only have a limited number of poles. Taking it a bit further: the expertsystem is in a way an overkill if it's just about predicting resonances in the helicity formalism.

[spflueger]

In addition, I think this is fundamentally a problem with the helicity formalism (using BW as resonance descriptions): this works only if poles are separated, therefore you can only have a limited number of poles. Taking it a bit further: the expertsystem is in a way an overkill if it's just about predicting resonances in the helicity formalism.

Hmm, that strongly connects to the State vs ComplexEnergyState issue (#127). Generally the expert system is not bound to a specific formalism for finding solutions. Let me clarify this width some "hands on" example again:

So if you look at an amplitude, for example a helicity amplitude, the only relevant quantities are the complex energy positions based on the dynamics functions, for example Breit Wigners with mass and width. Imagine if you analyze a dataset from an experiment. What you do is you define waves in your model with a fixed angular parts (spin info) and for example a BW function. The fit will optimize the mass and width according to the data. You would perform many fits with many models (different spin waves) to figure out which is fitting best. Now lets say you found a new particle state! But what do you know about this state? From the helicity formalism amplitude model alone (the same goes for any other formalism) you only extract the mass and width (or complex energy pole). Based on the various waves you use you also get information about the spin of the particle state. But how about other quantum numbers such as isospin, parity, c-parity, g-parity, charge, baryon number etc....? They are all inferred from the solutions of the expert system. So in case of the J/psi -> gamma pi0 pi0 example, the states found in the pi0pi0 subsystem all have isospin 0, g-parity +1, parity +1, c-parity +1, charge 0 and so on.

Therefore the expert system is some useful extension, quite orthogonal to the formalisms used in PWA. I'm not sure if I mentioned this also somewhere, but maybe its best to do this state solution finding independent of the formalism. And only the amplitude generators translate this information into the appropriate formalism models.