Optimal scheduling of data centers based on multiple games

Abstract

With the increasing dominance of electricity retailers in the electricity market, it has become a new trend for the data center (DC) to participate in sales-side transactions. However, data center electricity retailers (DCERs) and DCs that purchase electricity by DCERs, as different stakeholders, will inevitably face conflicts of interest. To promote the benefit distribution of DCs and DCERs to achieve a win–win situation, our study proposes an optimal scheduling method based on multiple games and establishes a mixed game model by integrating the master–slave game method and the cooperative game method, in which DCERs take profit maximization as the optimization goal, while Internet DCs take the lowest total cost as the optimization goal. The master–slave game is adopted between the DCER and the DC, and the cooperative game is adopted among the members of the DC. The benefits are distributed through Nash bargaining. The model is solved by using the particle swarm optimization algorithm combined with the alternating direction method of multipliers. To demonstrate the effectiveness of our proposed method, we provide an illustrative example that showcases its ability to not only increase DCER revenue by 136.04% but also decrease total DC costs by 9.39%. As a result, our method facilitates a more equitable distribution of cooperation revenues.