Exponential Discounting

Exponential discounting, a common method for evaluating future rewards in decision-making models, is being challenged by research showing its limitations in capturing human behavior and complex real-world scenarios. Current research focuses on developing alternative discounting models, such as quasi-hyperbolic and time-varying geometric discounting, and incorporating these into reinforcement learning algorithms like model-free methods and those based on Markov Perfect Equilibrium. This work aims to improve the accuracy and efficiency of decision-making models in various fields, from economics and psychology to artificial intelligence, by better reflecting the nuanced ways humans and agents value rewards over time.