In this paper, we consider time-varying multiserver queues with abandonment and retrials. For their performance analysis, fluid and diffusion limits utilizing strong approximations have been widely used in the literature. Although those limits are asymptotically exact, they may not accurately approximate performance of multiserver queues even if the number of servers is large. To address that concern, this paper focuses on developing a methodology by taking fluid and diffusion limits in a nontraditional fashion. We show that our approximation is significantly more accurate and also asymptotically true. We illustrate the effectiveness of our methodology by performing several numerical experiments.
