Econ Papers Daily / 期刊论文
序贯定价与随机成本下的算法合谋
Algorithmic collusion under sequential pricing and stochastic costs
摘要
The use of pricing algorithms raises concerns about algorithmic collusion. This paper considers a sequential pricing model where marginal cost fluctuates over time. I find that Q-learning algorithms autonomously collude even under cost uncertainty. Collusion is sustained by strategies that involve reward-punishment schemes. It suggests that cost uncertainty is not an obstacle to autonomous algorithmic collusion.
中文摘要
定价算法的使用引发了对算法合谋的担忧。本文考虑了一个边际成本随时间波动的序贯定价模型。我发现，即使在成本不确定的情况下，Q学习算法也会自主合谋。合谋通过涉及奖励-惩罚机制的策略得以维持。这表明，成本不确定性并非自主算法合谋的障碍。