12.2 Scheduling and PAC Methods

Production Activity Control (PAC) is the execution layer of manufacturing planning and control. After detailed schedules exist, PAC releases orders to the floor, sequences work at resources, tracks progress, and triggers corrective action when actual performance diverges from plan. On CPIM, PAC is not a software brand—it is a management system: authorize work, prioritize work, measure work, and close the loop.

The PAC Loop

A practical PAC cycle has four linked activities:

1. Order release — Convert planned or firm planned orders into shop packets (or electronic travelers) when materials, tools, and capacity are ready enough to start.
2. Dispatching — At each work center, decide the next job from the queue using priority rules and constraints.
3. Data collection and feedback — Capture completions, scrap, labor hours, and delays so status is visible.
4. Corrective action — Expedite, split, reschedule, add overtime, or stop releasing more work into a choked resource.

If release is undisciplined, queues explode. If dispatching is arbitrary, due dates slip even when capacity exists. If feedback is late, corrective action arrives after the customer already knows. Domain VII items often hide the failure in one of these four links.

Push Versus Pull

Push systems authorize production based on an upstream plan or schedule. MRP-generated planned orders, shop packets released on a weekly schedule, and forecast-driven batch launches are classic push. Push is powerful when lead times are long, product variety is high, and you must coordinate many components. The weakness is that push can flood the floor if the plan is wrong or release ignores current WIP.

Pull systems authorize production based on downstream consumption. Kanban cards, CONWIP (constant work-in-process) limits, and supermarket replenishment are pull mechanisms. Pull limits WIP and makes overproduction harder, which shortens queue time and exposes problems. Pull works best with relatively stable mix, reliable processes, and quick changeovers. Many plants hybridize: MRP plans materials (push planning), while final assembly or a bottleneck cell runs on pull signals.

Exam tip: “We release everything MRP says on Monday” is push. “We machine the next lot only when the assembly kanban is empty” is pull. Neither word alone proves lean maturity—execution discipline does.

Input/Output Control

Input/output (I/O) control is a PAC technique that manages work-center queues by comparing planned and actual input hours (work arriving) and output hours (work completed) over a period—often a week. The cumulative difference between input and output changes the queue (backlog) at the center.

Key relationships:

• If input persistently exceeds output, the queue grows and lead time lengthens.
• If output exceeds input, the queue shrinks—good if intentional, dangerous if it starves the next process.
• Planned input should be gated by demonstrated output capacity, not wishful release.

A simple weekly I/O report for a grinding cell might show planned input 160 hours, actual input 190, planned output 160, actual output 150. Queue grew by 40 hours. PAC response is not only “work overtime”—it is also to throttle input (delay releasing more grind jobs) until output catches up. I/O control is how finite thinking survives on a floor that still receives MRP pushes.

In week 2, overload arrived as excess input and weak output. In week 3, PAC cut planned input and raised planned output (overtime or extra shift), recovering queue. That storyline is classic Domain VII reasoning.

Dispatching

Dispatching is the short-horizon decision of which queued job a resource runs next. The dispatcher (human or system) uses a dispatch list ranked by priority rule, due date, setup family, or management override. Dispatching does not replace the detailed schedule; it executes and fine-tunes it under real-time conditions—machine down, operator absent, hot customer call.

Good dispatching practices:

• Keep the dispatch horizon short (today / this shift) so priorities stay current
• Respect routing sequence and material readiness before priority cosmetics
• Limit expedite overrides; every hot job creates a cold job somewhere else
• Align dispatch rules with the firm’s order winners (speed, due-date reliability, or changeover efficiency)

Poor dispatching looks like tribal knowledge: “Run whatever the supervisor shouts.” CPIM expects a rule-based approach connected to performance measures such as tardiness, flow time, and setup hours.

Connecting PAC Methods in One Flow

Consider a stamping plant. MRP (push planning) creates weekly requirements. PAC releases only what I/O control says the press line can absorb. At the press, dispatching sequences dies by setup family within due-date urgency. Downstream welding pulls stamped blanks from a supermarket. When actual press output falls, I/O control cuts further release, and planners reschedule rather than silently growing queues. That hybrid—push planning, gated release, pull between stages, rule-based dispatch—is the mature PAC picture Domain VII points toward.

Common Exam Failure Modes

• Treating PAC as “printing travelers” with no feedback loop
• Calling a system pull when kanban cards are ignored and supervisors still push batches
• Using I/O reports for blame instead of release control
• Dispatching by whoever complains loudest, destroying schedule integrity

Master the vocabulary, then apply it to vignettes: identify whether authorization is push or pull, whether queues are managed by I/O control, and whether dispatching follows an explicit rule. That triad is the heart of scheduling and PAC methods.