[RLlib] Bug fix: DQN goes into negative epsilon values after reaching explora…

[sven1977]

…tion percentage.

DQN goes into negative epsilon values after reaching exploration_percentage. This effectively disables all exploration for the rest of the run (usually, a small epsilon for the rest of the run is desirable, e.g. 0.01/0.02).

Related issue number

Checks

• I've run scripts/format.sh to lint the changes in this PR.
• I've included any doc changes needed for https://ray.readthedocs.io/en/latest/.
• I've made sure the tests are passing. Note that there might be a few flaky tests, see the recent failure rates at https://ray-travis-tracker.herokuapp.com/.

[AmplabJenkins]

Can one of the admins verify this patch?

[AmplabJenkins]

Test FAILed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/Ray-PRB/21164/
Test FAILed.

[ericl]

LGTM pending tests.

[AmplabJenkins]

Test FAILed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/Ray-PRB/21192/
Test FAILed.

[AmplabJenkins]

Test FAILed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/Ray-PRB/21196/
Test FAILed.

[sven1977]

@ericl RLlib tests are all good.

[AmplabJenkins]

Test PASSed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/Ray-PRB/21201/
Test PASSed.

[sven1977]

Thanks!