Limit theorems for priority queues with patch processing in heavy traffic. I (Q756869)

An M/G/1 queue with r types of customers is considered. Customers of type i-1 have priority over those of type i, and when the server becomes idle, he immediately serves the whole batch of customers having the lowest priority number. A heavy-traffic limit theorem is proved for the virtual waiting time \(w_ i\) of a batch that contains an i-request arriving at time t.