CEMPlanner

class torchrl.modules.CEMPlanner(*args, **kwargs)

CEMPlanner 模組。

參考：用於最佳化的交叉熵方法，Botev 等人，2013 年

當給定包含初始狀態的 TensorDict 時，此模組將執行 CEM 規劃步驟。CEM 規劃步驟透過從均值為零、方差為一的高斯分佈中取樣動作來執行。然後，使用取樣到的動作在環境中執行滾動。然後對滾動獲得的累積獎勵進行排名。我們選擇前 k 個回合，並使用它們的動作來更新動作分佈的均值和標準差。CEM 規劃步驟會重複指定的次數。

呼叫該模組會返回在給定規劃地平線的情況下經驗性最大化回報的動作

引數:

• env (EnvBase) – 用於執行規劃步驟的環境（可以是 ModelBasedEnv 或 EnvBase）。

• planning_horizon (int) – 模擬軌跡的長度

• optim_steps (int) – MPC 規劃器使用的最佳化步數

• num_candidates (int) – 從高斯分佈中取樣的候選數量。

• top_k (int) – 用於更新高斯分佈均值和標準差的前 k 個候選數量。

• reward_key (str, optional) – TensorDict 中用於檢索獎勵的鍵。預設為“reward”。

• action_key (str, optional) – TensorDict 中用於儲存動作的鍵。預設為“action”。

示例

>>> from tensordict import TensorDict
>>> from torchrl.data import Composite, Unbounded
>>> from torchrl.envs.model_based import ModelBasedEnvBase
>>> from torchrl.modules import SafeModule
>>> class MyMBEnv(ModelBasedEnvBase):
...     def __init__(self, world_model, device="cpu", dtype=None, batch_size=None):
...         super().__init__(world_model, device=device, dtype=dtype, batch_size=batch_size)
...         self.state_spec = Composite(
...             hidden_observation=Unbounded((4,))
...         )
...         self.observation_spec = Composite(
...             hidden_observation=Unbounded((4,))
...         )
...         self.action_spec = Unbounded((1,))
...         self.reward_spec = Unbounded((1,))
...
...     def _reset(self, tensordict: TensorDict) -> TensorDict:
...         tensordict = TensorDict(
...             {},
...             batch_size=self.batch_size,
...             device=self.device,
...         )
...         tensordict = tensordict.update(
...             self.full_state_spec.rand())
...         tensordict = tensordict.update(
...             self.full_action_spec.rand())
...         tensordict = tensordict.update(
...             self.full_observation_spec.rand())
...         return tensordict
...
>>> from torchrl.modules import MLP, WorldModelWrapper
>>> import torch.nn as nn
>>> world_model = WorldModelWrapper(
...     SafeModule(
...         MLP(out_features=4, activation_class=nn.ReLU, activate_last_layer=True, depth=0),
...         in_keys=["hidden_observation", "action"],
...         out_keys=["hidden_observation"],
...     ),
...     SafeModule(
...         nn.Linear(4, 1),
...         in_keys=["hidden_observation"],
...         out_keys=["reward"],
...     ),
... )
>>> env = MyMBEnv(world_model)
>>> # Build a planner and use it as actor
>>> planner = CEMPlanner(env, 10, 11, 7, 3)
>>> env.rollout(5, planner)
TensorDict(
    fields={
        action: Tensor(shape=torch.Size([5, 1]), device=cpu, dtype=torch.float32, is_shared=False),
        done: Tensor(shape=torch.Size([5, 1]), device=cpu, dtype=torch.bool, is_shared=False),
        hidden_observation: Tensor(shape=torch.Size([5, 4]), device=cpu, dtype=torch.float32, is_shared=False),
        next: TensorDict(
            fields={
                done: Tensor(shape=torch.Size([5, 1]), device=cpu, dtype=torch.bool, is_shared=False),
                hidden_observation: Tensor(shape=torch.Size([5, 4]), device=cpu, dtype=torch.float32, is_shared=False),
                reward: Tensor(shape=torch.Size([5, 1]), device=cpu, dtype=torch.float32, is_shared=False),
                terminated: Tensor(shape=torch.Size([5, 1]), device=cpu, dtype=torch.bool, is_shared=False)},
            batch_size=torch.Size([5]),
            device=cpu,
            is_shared=False),
        terminated: Tensor(shape=torch.Size([5, 1]), device=cpu, dtype=torch.bool, is_shared=False)},
    batch_size=torch.Size([5]),
    device=cpu,
    is_shared=False)

planning(tensordict: TensorDictBase) → Tensor

執行 MPC 規劃步驟。

引數:

td (TensorDict) – 要在其上執行規劃步驟的 TensorDict。