Optimal use of simplified social information in sequential decision-making

Area of Science:

• Behavioral Ecology
• Cognitive Science
• Evolutionary Biology

Background:

• Social animals often leverage the decisions of others to improve their own choices.
• The benefits of using social information are counterbalanced by the costs of acquiring and processing it.
• Prior models assumed optimal responses to full decision sequences, which may be cognitively demanding.

Purpose of the Study:

• To derive optimal decision-making policies for agents using simplified social information.
• To compare behaviors based on aggregate choices versus the most recent decision.
• To analyze the evolutionary stability of different social information processing strategies.

Main Methods:

• Theoretical derivation of optimal policies for simplified social information use.
• Modeling agent behavior under different information-gathering strategies (aggregate, recent, hybrid).
• Analysis of evolutionary stability under varying cognitive costs.

Main Results:

• Agent behavior differs significantly when attending to aggregate choices versus only the most recent decision.
• A hybrid strategy accurately approximates the optimal policy derived from full decision sequences.
• The hybrid strategy is evolutionarily dominant with low cognitive costs; recent-decision strategy dominates with high costs.

Conclusions:

• Effective social decision-making is achievable with simplified rules, not requiring excessive cognitive capacity.
• Cognitive costs influence the evolutionarily stable strategy for social information use.
• Ecological variations likely drive differences in the social information strategies employed by various animal species.