[WIP] Add isostress and Voigt-Reuss-Hill averaging to CompositeViscoPlastic

[bobmyhill]

Yesterday evening, @naliboff and I discussed the possibility of implementing different averaging schemes with the CompositeViscoPlastic rheology model. The model as implemented in #3884 assumes isostrain (Voigt) averaging of material viscosities, which can result in very high viscosities.

As CompositeViscoPlastic decomposes the strains for each mechanism, isostress or Voigt-Reuss-Hill* averaging can both be implemented. The isostress averaging scheme is likely to be more efficient than isostrain, because isostrain averaging requires one stress inversion per material, whereas isostress averaging requires only one stress inversion in total.

This PR adds the option to choose between isostrain, isostress and Voigt-Reuss-Hill viscosity averaging schemes. Opened for discussion. PR #3884 should be merged before this is reviewed.

* an arithmetic average of isostress and isostrain

• I have followed the instructions for indenting my code.
• I have tested my new feature locally to ensure it is correct.
• I have created a testcase for the new feature/benchmark in the tests/ directory.
• I have added a changelog entry in the doc/modules/changes directory that will inform other users of my change.

[bobmyhill]

Superceded by #5990