Back to the question, this has got to be a very difficult thing to calculate because you have both pressure and temperature loads, plus localized loads from the head bolts. I can't think of any way other than a finite elephant analysis, and that will only be as good as how accurately you model the loading. Your model of the head will have to be limited in detail to save on computer time. After all this effort, a head that passes the analysis might end up failing at a thin wall, an insufficient fillet, or a hot spot.
You could do the finite element and then follow it up with testing, or you could use the old "rules of thumb" and then test it.
You can also optimize the design and then have the foundry tell you they can't cast it.