Multilevel feedback queue

In <a href="/facts/Computer_science/lN3pEyPz">computer science</a>, a multilevel feedback queue is a <a href="/facts/Scheduling_(computing)/XptmcM1a">scheduling</a> algorithm. Scheduling algorithms are designed to have some process running at all times to keep the <a href="/facts/Central_processing_unit/Jxp1bqMV">central processing unit</a> (CPU) busy. The multilevel feedback queue extends standard algorithms with the following design requirements:

<ol><li>Separate processes into multiple ready queues based on their need for the processor.</li>
<li>Give preference to processes with short CPU bursts.</li>
<li>Give preference to processes with high <a href="/facts/I%2fO_bound/RMhfJBsn">I/O</a> bursts. (I/O bound processes will sleep in the wait queue to give other processes CPU time.)</li></ol>
The multilevel feedback queue was first developed by <a href="/facts/Fernando_J._Corbat%25C3%25B3/2kXkQk3I">Fernando J. Corbató</a> (1962). For this accomplishment, the <a href="/facts/Association_for_Computing_Machinery/abI2atlP">Association for Computing Machinery</a> awarded Corbató the <a href="/facts/Turing_Award/HbEEp0GF">Turing Award</a>.

Multilevel feedback queue open-in-new

Multilevel feedback queue