Consider the set of process with arrival time (in milliseonds), CPU burst time…

Key idea: use preemptive priority scheduling (lower priority number means higher priority). A running process is preempted when a newly arrived process has a smaller priority value.

• Timeline of execution:

• 0–2: P1 runs (P1 remaining = 9).

• 2–5: P4 runs (P4 remaining = 7) after arriving at 2 with higher priority than P1.

• 5–33: P2 runs (arrives at 5 with highest priority 0) and completes at 33.

• 33–40: P4 resumes and completes at 40.

• 40–49: P1 resumes and completes at 49.

• 49–51: P3 runs (arrived at 12) and completes at 51.

• 51–67: P5 runs (arrived at 9) and completes at 67.

Now compute waiting time = completion time − arrival time − burst time for each process.

• P1: completion 49, arrival 0, burst 11 → waiting = 49 − 0 − 11 = 38

• P2: completion 33, arrival 5, burst 28 → waiting = 33 − 5 − 28 = 0

• P3: completion 51, arrival 12, burst 2 → waiting = 51 − 12 − 2 = 37

• P4: completion 40, arrival 2, burst 10 → waiting = 40 − 2 − 10 = 28

• P5: completion 67, arrival 9, burst 16 → waiting = 67 − 9 − 16 = 42

Average waiting time = (38 + 0 + 37 + 28 + 42) / 5 = 145 / 5 = 29 milliseconds.