[Paper Review] Resource-rational Task Decomposition to Minimize Planning Costs.

Carlos G. Correa, Mark K. Ho+2 more|arXiv (Cornell University)|Jan 1, 2020

Decision-Making and Behavioral Economics2 citations

TL;DR

This paper proposes a resource-rational framework for task decomposition, modeling how people break down tasks into subtasks to optimize cognitive resource use. By formalizing decomposition as a representation problem that minimizes planning costs under limited resources, the model replicates key human planning behaviors and offers a normative account of subtask identification and hierarchical planning strategies.

ABSTRACT

People often plan hierarchically. That is, rather than planning over a monolithic representation of a task, they decompose the task into simpler subtasks and then plan to accomplish those. Although much work explores how people decompose tasks, there is less analysis of why people decompose tasks in the way they do. Here, we address this question by formalizing task decomposition as a resource-rational representation problem. Specifically, we propose that people decompose tasks in a manner that facilitates efficient use of limited cognitive resources given the structure of the environment and their own planning algorithms. Using this model, we replicate several existing findings. Our account provides a normative explanation for how people identify subtasks as well as a framework for studying how people reason, plan, and act using resource-rational representations.

Motivation & Objective

• To explain why people decompose tasks into subtasks rather than planning monolithically.
• To formalize task decomposition as a resource-rational representation problem that minimizes cognitive planning costs.
• To provide a normative account of how people identify and structure subtasks based on environmental structure and planning algorithms.
• To unify existing findings on hierarchical planning under a single cognitive efficiency principle.
• To offer a framework for studying human reasoning, planning, and action through resource-rational representations.

Proposed method

• Modeling task decomposition as a representation choice that minimizes planning costs under cognitive resource constraints.
• Formalizing the trade-off between representational complexity and planning efficiency using a normative optimization framework.
• Using the structure of the environment and known planning algorithms to guide subtask identification.
• Applying the model to replicate existing empirical findings on human task decomposition behavior.
• Deriving subtask structures that balance computational cost and representational clarity.
• Demonstrating that the proposed decomposition strategy leads to efficient planning under limited cognitive resources.

Experimental results

Research questions

• RQ1 Why do people decompose tasks into subtasks rather than planning over a monolithic task representation?
• RQ2 What principles guide people in selecting which subtasks to form during hierarchical planning?
• RQ3 How does the structure of the environment influence the way people decompose tasks?
• RQ4 To what extent can resource-rationality explain observed patterns in human task decomposition?
• RQ5 How do planning algorithms interact with representation choices to shape task decomposition?

Key findings

• The model successfully replicates established empirical findings on human task decomposition, validating its normative framework.
• People decompose tasks in ways that minimize planning costs by efficiently allocating limited cognitive resources.
• Subtask identification is guided by a balance between representational simplicity and computational efficiency.
• The framework provides a principled explanation for why certain decomposition structures are preferred over others.
• The model accounts for both the structure of the environment and the constraints of human planning algorithms in shaping decomposition choices.
• Resource-rational task decomposition emerges as a coherent strategy for optimizing cognitive efficiency in planning.