Input-buffered asynchronous transfer mode (ATM) packet switches are simpler
than output-buffered switches. However, due to HOL blocking, their through
put is poor. Neural schedulers represent a promising solution for high thro
ughput input-buffered switching, but their response time variance is too hi
gh for realistic hard real-time constraints. To overcome this problem, in t
his paper we formulate and evaluate a new neural scheduler with bounded res
ponse time.