Subject and Keywords:
This paper assesses the performance of the queuing system with several fixed and additional multitask servers. The additional server serves customers in the case of necessity to avoid congestion and decrease the number of impatient customers. An additional server will be activated whenever the length of queue becomes greater than or equal to user defined size i.e. threshold value. Also, this server will be switched off when the length of queue falls below user defined size, but ongoing service has to be finished. The main issue is to decide when and how long the additional server would work with customers. This paper proposes a methodology to choose the best activation rule for the work of an additional server, based on the performance evaluation of queuing system. The Petri nets (PN) simulation is used to calculate the performance measures. Relative efficiency evaluation of different activation rules is carried out by data envelopment analysis (DEA). The methodology is applied on real-life data collected from a postal office.