Non-Equilibrium Relations for Bounded Rational Decision-Making in Changing Environments.

Area of Science:

• Decision-making and behavioral adaptation
• Non-equilibrium thermodynamics
• Information theory

Background:

• Organisms continuously adapt to environmental changes for survival.
• Behavioral adaptation involves optimizing decision-making under resource constraints.
• Thermodynamic systems balance energy and entropy; rapid environmental changes lead to non-equilibrium states.

Purpose of the Study:

• To apply non-equilibrium thermodynamics to model decision-making in adaptive organisms.
• To quantify performance loss due to imperfect adaptation in changing environments.
• To derive new relations for decision-making analogous to fluctuation theorems.

Main Methods:

• Developed an abstract model of decision-makers with limited information-processing resources.
• Utilized concepts from non-equilibrium thermodynamics to analyze adaptation dynamics.
• Simulated decision and inference problems in discrete and continuous domains.

Main Results:

• Characterized decision-makers adapting to externally driven, changing utility functions.
• Quantified performance loss from imperfect adaptation.
• Derived decision-making relations analogous to Crooks' fluctuation theorem and Jarzynski's equality.

Conclusions:

• Non-equilibrium thermodynamics provides a framework for understanding adaptive decision-making.
• The study offers generalizable methods for quantifying adaptation efficiency and performance.
• New theoretical relations advance the understanding of decision processes in dynamic environments.