Learning Agents With Prioritization and Parameter Noise in Continuous State and Action Space