Resolve "Parallelize synchronous `apply_rule`" with pSTL

What does this MR do?

Provide parallel features for Utopia.

Build System

• Add TBB as optional dependency.
• Add CMake function enable_parallel() for enabling parallel features through a pre-processor definition on targets.
• Enable parallelism on models by default, which can be reverted using the new DISABLE_PARALLEL option of add_model().

Parallel Functions

• Add header parallel.hh, defining types and functions for deciding parallel execution at runtime.
• Add overloads for STL algorithms copy(), for_each(), and transform().
• Add tests for these overloads and parallel functionality in general, see CORE_test_parallel_stl.

Parallel apply_rule()

• Extend apply_rule() interface for manual state update by Utopia::ExecPolicy parameter for indicating execution policy to use when applying a rule.
• Use parallel features inside apply_rule() where obviously possible.
• Add apply_rule() overloads without execution policy parameter for backwards compatibility, which call the respective sequential version.

Documentation

• Document parallel features in Doxygen docs. Add submodule for STL algorithm overloads.

Is there something that needs to be double checked?

• I only added documentation to the Doxygen docs. Should we move/copy some of that into the Sphinx docs?
• Do we need a test for the addition of the base_cfg parameter?
• So far we only test if CORE_apply_test works with parallel features enabled. This only includes "internal" parallelization, as all function calls use sequential overloads. Do we need a dedicated test for calls including Utopia::ExecPolicy?

Can this MR be accepted?

• Add TBB to Docker base image
  • Build Docker base image eoan-v2
  • Protect CI Variables again
• Implement changes
  • Handle parallel features in build system
  • Implement runtime parallelism
  • Make apply_rule() parallel-aware
• Add tests
  • Check Utopia-internal parallel features
  • Confirm parallel STL algorithms yield expected results
  • Confirm that sequential apply_rule() works the same with parallel features enabled
• Reasonably up-to-date with current master
• Pipeline passing without warnings
• Checked test code coverage on new and adjusted code: 100% line coverage
• History cleaned-up
• Approved by @blsqr

Related issues

Closes #27 (closed)