Diversity Evolutionary Policy Deep Reinforcement Learning

Table 3

The mean and standard deviation of the cumulative return per turn in different MuJoCo tasks.

Task	TD3	Multiactor TD3	CEM	CEM-TD3	DPERL
Hopper-v2	3025 ± 577	3241 ± 363	1054 ± 17	3652 ± 116	3732 ± 106
HalfCheetah-v2	10002 ± 930	10341 ± 578	2298 ± 690	10978 ± 758	11615 ± 464
Ant-v2	3618 ± 425	3881 ± 319	845 ± 52	4037 ± 466	4852 ± 317
Walker2d-v2	4399 ± 238	4470 ± 301	743 ± 225	4612 ± 357	5001 ± 562