A weighted constraint satisfaction approach to human goal-directed decision making

Area of Science:

• Cognitive Psychology
• Decision Science
• Computational Neuroscience

Background:

• Effective long-range planning requires considering future information without myopic focus.
• Subgoal resolution necessitates prioritizing proximal, relevant information and ignoring irrelevant future data.

Purpose of the Study:

• To investigate if decision-making relies on flexible information weighting rather than strict problem decomposition.
• To model human path choices in mazes as a weighted information integration process.

Main Methods:

• Participants navigated mazes to find shortest goal-reaching paths.
• Path choices were modeled as a noisy, weighted information integration process.
• Tasks involved a base scenario and a hierarchical scenario with a subgoal.

Main Results:

• In a base task, participants balanced starting-point and goal-proximal factors.
• In a hierarchical task, participants prioritized subgoal constraints but were influenced by the final goal.
• More accurate participants showed better weighting of relevant information and less influence from irrelevant information.

Conclusions:

• Human decision-making employs graded, task-sensitive weighting of multiple constraints for approximate optimality.
• This flexible weighting mechanism supports both hierarchical and non-hierarchical goal-directed tasks.
• Findings challenge rigid, step-by-step problem-solving models in favor of dynamic information integration.