Optimize Pricing Policy in Evolutionary Market with Multiple Proactive Competing Cloud Providers

In the cloud computing market with multiple competing cloud providers, each provider needs to run an optimize pricing policy to determine the price for cloud resources in order to beat other providers and maximize the long-term revenue. In this paper, we intend to design an optimal pricing policy for the cloud provider by taking into account the following two facts: (1) the market is evolutionary in terms of the increased amount of cloud users and the decreased marginal cost of providers; (2) the market is fully competitive since all providers expect to receive more profits while users expect to pay less for their requirements. By considering these realistic situations, we analyze how the cloud provider designs an optimal pricing policy while competing against the other provider. Specifically, we model the pricing issue in the competing environment as a Markov games, and then adopt the Minimax-Q and Q-learning algorithm to find the optimal pricing policy. Finally, we evaluate our algorithms and compare the pricing policy with other existing polices. Numerical simulations demonstrate the effectiveness of our proposed approach.