[RLlib] Address possible memory inefficiency in WorkerSet.sync_weights (b/c not using ray.put()).

[sven1977]

• Address possible memory inefficiency in WorkerSet.sync_weights (b/c not using ray.put()). NOTE that this is already implemented correctly for the new API stack (see Algorithm.sync_env_runner_states()), but we missed retro-fixing this for the old API stack.
• Address instances of RLlib referencing larger objects (via self) from within a remote function, causing this large object to be pickle-copied num_worker times (instead of just passing the necessary property value from that self).
• Minor Algorithm method deprecation and cleanup.

Why are these changes needed?

Related issue number

Checks

• I've signed off every commit(by using the -s flag, i.e., git commit -s) in this PR.
• I've run scripts/format.sh to lint the changes in this PR.
• I've included any doc changes needed for https://docs.ray.io/en/master/.
  • I've added any new APIs to the API Reference. For example, if I added a
    method in Tune, I've added it in doc/source/tune/api/ under the
    corresponding .rst file.
• I've made sure the tests are passing. Note that there might be a few flaky tests, see the recent failures at https://flakey-tests.ray.io/
• Testing Strategy
  • Unit tests
  • Release tests
  • This PR is not tested :(

[simonsays1980]

LGTM: Great catches by the customer. I am curious, if this will give some significant performance gains in the base algorithms.