How to deal with KerasRL DDPG algorithm getting stuck in a local optima?

Question

I am using KerasRL DDPG to try to learn a policy on my own custom environment, but the agent is stuck in a local optima al