A stochastic Pairwise Fermi rule modified by utilizing the average in payoff differences of neighbors leads to increased network reciprocity in spatial prisoner's dilemma games

In a 2 × 2 prisoner's dilemma (PD) game, network reciprocity is one of the mechanisms for increasing social viscosity, which leads to a cooperative equilibrium. The Pairwise Fermi (PW-Fermi) rule has been accepted as an updating protocol, as its stochasticity is similar to the real-world human decision-making process. In this paper, we elucidated a modification to the PW-Fermi rule by utilizing the averaged payoff difference instead of the simple payoff difference between a focal agent and their neighbors. This led to a significantly enhanced level of network reciprocity. The mechanism of this enhancement is clarified by discussing the concepts of the enduring period (END) and the expanding period (EXP).