# A Guiding Principle for Causal Decision Problems

## TL;DR

This paper introduces a decision-making framework based on causal graphical models, combining Pearl's Do-Calculus with expected utility to enable online decisions with theoretical guarantees and causal environment learning.

## Contribution

It proposes a novel decision criterion integrating causal inference with utility maximization, providing a new approach for causal decision problems.

## Key findings

- Performance comparable to classic Reinforcement Learning algorithms
- Allows learning of causal models of the environment
- Provides theoretical guarantees of decision optimality

## Abstract

We define a Causal Decision Problem as a Decision Problem where the available actions, the family of uncertain events and the set of outcomes are related through the variables of a Causal Graphical Model $\mathcal{G}$. A solution criteria based on Pearl's Do-Calculus and the Expected Utility criteria for rational preferences is proposed. The implementation of this criteria leads to an on-line decision making procedure that has been shown to have similar performance to classic Reinforcement Learning algorithms while allowing for a causal model of an environment to be learned. Thus, we aim to provide the theoretical guarantees of the usefulness and optimality of a decision making procedure based on causal information.