The world of strategies with memory

TL;DR

This paper investigates how memory depth influences the evolution of strategies in a generalized prisoner's dilemma, showing that longer memory generally leads to more successful, cooperative strategies with higher complexity.

Contribution

It introduces the concept of strategy complexity and demonstrates that strategies with maximum or near-maximum memory and complexity tend to prevail in natural selection.

Findings

Increased memory depth benefits evolutionary success.

Winning strategies are highly complex and tend to be cooperative.

Maximum memory strategies outperform less complex ones.

Abstract

As part of a generalized "prisoners' dilemma", is considered that the evolution of a population with a full set of behavioral strategies limited only by the depth of memory. Each subsequent generation of the population successively loses the most disadvantageous strategies of behavior of the previous generation. It is shown that an increase in memory in a population is evolutionarily beneficial. The winners of evolutionary selection invariably refer to agents with maximum memory. The concept of strategy complexity is introduced. It is shown that strategies that win in natural selection have maximum or near maximum complexity. Despite the fact that at a separate stage of evolution, according to the payout matrix, the individual gain, while refusing to cooperate, exceeded the gain obtained while cooperating. The winning strategies always belonged to the so-called respectable strategies that are clearly prone to cooperation.