[Feature Request]: Implement Policy Learning based on APO Model

### Describe the feature you want to propose or implement

Implement a general policy learning method based on multiple possible treatments.

### Propose a possible solution or implementation

Based on [Policy Learning with Confidence](https://arxiv.org/abs/2502.10653).
Using the APOs and several possible weighted options (and e.g. tree search) one could implement the proposed algorithms as the solely rely on value estimates and standard errors.