College of Computer and Information Science, Southwest University, Chongqing, 400715, China.
Sci Rep. 2023 Jan 2;13(1):1. doi: 10.1038/s41598-022-26890-9.
In many dilemmas, decisions are determined not by a single factor, but by multiple ones, including memory, reputation, reward and punishment. In recent years, how to design a mechanism to promote cooperation has become a research hot-spot. However, most of the previous studies mainly consider the historical benefits of the game, and pay less attention to the stability of the strategy (the frequency of strategy changes in the length of memory) and the proportion of memory in decision-making. The decision-making process of group evolution involves the influence of memory information on cooperative evolution in multi round games. It makes up for the lack of stability factors and weights in previous studies. Based on the above factors, a new strategy update rule is proposed to study the influence of the stability of historical strategy information on the evolution of cooperation in prisoner's dilemma game, and the influence of memory weight on cooperation is considered. The stability of the current strategy is measured by the strategy in historical memory (the number of times the strategy in memory is continuous and consistent with the current strategy), which can determine the probability of an individual learning the neighbor strategy next time. Numerical simulation shows that an appropriate increase in the length of historical memory is more conducive to the emergence of cooperation, and the greater the weight of historical strategy information is, the more conducive to promoting cooperation, which shows that historical strategy information is still the main factor in decision-making. This study will help us understand the cooperative evolution of many real systems, such as nature, biology, society and so on, and effectively design reasonable mechanisms to promote cooperation.
在许多困境中，决策不是由单一因素决定的，而是由多个因素决定的，包括记忆、声誉、奖励和惩罚。近年来，如何设计一种促进合作的机制已成为研究热点。然而，之前的大多数研究主要考虑了游戏的历史收益，而对策略的稳定性（记忆长度内策略变化的频率）和记忆在决策中的比例关注较少。群体进化的决策过程涉及到记忆信息对多轮游戏中合作进化的影响。它弥补了之前研究中稳定性因素和权重的不足。基于上述因素，提出了一种新的策略更新规则，以研究历史策略信息的稳定性对囚徒困境游戏中合作进化的影响，以及记忆权重对合作的影响。当前策略的稳定性通过历史记忆中的策略（记忆中的策略与当前策略连续一致的次数）来衡量，这可以确定个体下次学习邻居策略的概率。数值模拟表明，适当增加历史记忆的长度更有利于合作的出现，而历史策略信息的权重越大，越有利于促进合作，这表明历史策略信息仍然是决策的主要因素。这项研究将有助于我们理解自然界、生物学、社会等许多真实系统的合作进化，并有效地设计合理的机制来促进合作。
