关键词: 管理科学与工程, 学习效应, 公平关切, 斯坦克伯格博弈, 强化学习模型

Abstract: Experimental studies of behavioral economics have provided convincing evidences that people are not perfectly rational but are affected by factors such as fairness concerns and learning. We conduct laboratory experiments based on a triadic supply chain to examine the horizontal fairness concerns of backup supplier and the learning curves of both manufacturer and backup supplier. Both individual learning and social learning settings are considered. The experiment results support the learning effect hypothesis in the supply chain: As the experiment proceeds, the decision time for a single period decreases; the overall rejection rate of backup supplier is gradually reduced; and the strategies of manufacturers are gradually concentrated. Furrhermore, we construct a reinforcement learning model with horizontal fairness concerns. Parameter estimation results show that the horizontal fairness concern of backup supplier is significant in both experiments with individual learning setting and social learning setting.

Key words: management science and engineering, learning effect, fairness concern, stackelberg game, reinforcement learning model