Scheduling assessment gates

Score each gate 0 missing/incorrect, 1 partial or prompted, 2 independently correct with inspectable evidence, 3 correct transfer to a fresh case with limitations. Every gate must reach 2; scores cannot compensate for a failed gate.

• PS1 — Event contract: trace simultaneous arrivals, quantum-boundary arrivals and idle intervals, showing queue order before each dispatch.
• PS2 — Metrics: derive per-job response, turnaround and waiting independently; reconcile run/switch/idle totals with the horizon and explain the utilization denominator.
• PS3 — Implementation: own starter passes supplied checks and at least three independently designed cases; explain state transitions without reading the reference.
• PS4 — Fault discrimination: reject front requeue and explain why conservation alone misses it; supply an exact trace and repair.
• PS5 — Experiment and scope: compare six quantum/cost settings, defend a workload-specific choice, and distinguish the simulator from Linux, I/O and multicore scheduling.
• PS6 — Retained transfer: after 48 hours, solve a fresh cost/arrival case and explain how process state, context-switch mechanism and scheduling policy differ. Preserve help and source records.

Allocate 5–7 hours including repair and delayed recall; elapsed waiting is not study time. Record initial score, evidence path, feedback and fresh retest for each gate. Missing evidence means incomplete. Automated checks do not grade independence or a policy recommendation. Use calibration after your attempt. Original module completion requirements remain in force.