1: \begin{abstract}

2: In this paper, we provide a new and effective approach for studying

3: super-exponential solution of a retrial supermarket model with Poisson

4: arrivals, exponential service times and exponential retrial times and with two

5: different probing-server numbers. We describe the retrial supermarket model as

6: a system of differential equations by means of density-dependent jump Markov

7: processes, and obtain an iterative algorithm for computing the fixed point of

8: the system of differential equations. Based on the fixed point, we analyze the

9: expected sojourn time that a tagged arriving customer spends in this system,

10: and use numerical examples to indicate different influence of the two

11: probing-server numbers on system performance including the fixed point and the

12: expected sojourn time. Furthermore, we analyze exponential convergence of the

13: current location of the retrial supermarket model to the fixed point, and

14: apply the Kurtz Theorem to study density-dependent jump Markov process given

15: in the retrial supermarket model, which leads to a Lipschitz condition under

16: which the fraction measure of the retrial supermarket model weakly converges

17: to the system of differential equations. This paper arrives at a new

18: understanding of how the workload probing can help in load balancing jobs in

19: retrial supermarket models.

20: 

21: \vskip                                      0.5cm

22: 

23: \noindent\textbf{Keywords:} Randomized load balancing, retrial

24: queue; supermarket model, super-exponential solution,

25: density-dependent jump Markov process, fixed point, sojourn time,

26: exponential convergence, Lipschitz condition.

27: 

28: \end{abstract}