Game-Theoretic Analysis of a Fluid Queue with Markovian Vacations and an N-Policy
摘要
We present a game-theoretic analysis of a fluid queue with Markovian vacations and an N-policy. Once the buffer becomes empty, it closes and takes a vacation. When this vacation ends, if the fluid level exceeds the threshold N, the buffer immediately resumes service. Otherwise, if the fluid level after the vacation is still below N, the buffer takes another vacation, and this pattern continues until the fluid level in the buffer exceeds N after one vacation ends. For such a fluid model, based on a natural reward/cost function, we study the equilibrium strategies of the fluids regarding the joining/balking dilemma under the fully observable case, almost observable case, and fully unobservable case. Furthermore, we obtain the equilibrium social benefits for different cases. Finally, through numerical examples, we illustrate the influence of system parameters on the equilibrium social benefit and explore the optimal social benefit strategy. The price of anarchy (PoA) is used to measure the inefficiency of equilibrium strategies and give some qualitative insights.