It depends on what type of scheduling policy has been
chosen. In general, the way to schedule the time quantum
(time slice)is either to take a fixed or variable length of
time unit. In an adaptive style of scheduling policy, the
scheduler will give a high priority process as much time as
it needs unless another higher priority process comes in the
ready queue to demand quantum of execution time. In this
case, a variable number of time units will be given to that
process. However, in a round-robin style of scheduling
policy, the time quantum is fiexed and predefined, and each
of processes that gets its turn to run through the time
quantum will complete the full length of the time unit
allocated to it before a context switch takes place. In
other words, the scheduler has to schedule the processes
according to its scheduling policy as it's been designed.