support multi-dim reward for AC and PPO (#952)

alf/algorithms/actor_critic_algorithm.py

alf/algorithms/actor_critic_algorithm_test.py

alf/algorithms/actor_critic_loss.py

alf/algorithms/agent.py

alf/algorithms/lagrangian_reward_weight_algorithm.py

alf/algorithms/lagrangian_reward_weight_algorithm_test.py

alf/algorithms/muzero_algorithm.py

alf/algorithms/ppo_algorithm.py

alf/algorithms/ppo_algorithm_test.py

alf/algorithms/ppo_loss.py

alf/algorithms/rl_algorithm_test.py

alf/algorithms/sac_algorithm.py

alf/examples/ppo_conf.py

alf/networks/network.py

alf/utils/tensor_utils.py