Transient analysis of the Erlang A model

Area of Science:

• Operations Research
• Applied Probability
• Queueing Theory

Background:

• The Erlang A model, a queueing system with Poisson arrivals, exponential service times, and m parallel servers, incorporates customer abandonment.
• Transient characteristics of such queueing systems, including time-dependent distributions and first passage times, have been historically challenging to analyze.
• The queue length process in the Erlang A model can be represented as a birth-death process.

Purpose of the Study:

• To derive explicit expressions for the Laplace transforms of the time-dependent distribution and first passage time for the Erlang A queue.
• To demonstrate that these previously intractable transient characteristics can be solved.
• To extend the findings to related queueing models and diffusion approximations.

Main Methods:

• Utilized Green's functions and contour integrals involving hypergeometric functions to solve for the Laplace transforms.
• Analyzed the queue length process as a birth-death process.
• Applied the derived methods to specialize results for the M/M/m queue and the M/M/m/m loss model.

Main Results:

• Obtained explicit expressions for the Laplace transforms of the time-dependent distribution and first passage time for the Erlang A queue.
• Demonstrated the tractability of these transient characteristics, which were previously considered intractable.
• Derived corresponding results for diffusion approximations of these queueing models.

Conclusions:

• The study successfully provides novel analytical solutions for the transient behavior of the Erlang A queueing model.
• The methods developed offer a significant advancement in the analysis of queueing systems with customer abandonment.
• The findings are applicable to various queueing scenarios, including the M/M/m and M/M/m/m models, and their diffusion approximations.