Question 5 of 7: Production Planning and Control — Volume, Order Types and Dispatching

Question 5: Production Planning and Control — Volume, Order Types and Dispatching (20 marks)

Question text not reproduced: the examination questions are © Engineers and Geoscientists BC. Open the official past paper (linked at the top of this page) to read the question, then follow the worked solution below.

Part (a) — why volume has little effect on the design of a control system

The proposition sounds wrong at first hearing, because a plant making a million flashlights a year plainly does not look like a plant making a thousand. The resolution is that volume acts on the control system only indirectly, through the process choice, and that once the process is fixed the control system is determined by something else entirely: the variety and complexity of what has to be controlled, and the degree of uncertainty in demand, supply and process yield.

Consider what a production planning and control system actually does. It forecasts or accepts demand; it converts that demand into a master schedule; it explodes the schedule against a bill of material and an inventory record to release orders; it loads and sequences those orders against finite capacity; it dispatches work to individual work centres; and it collects progress and feeds the variance back. Every one of those six functions must exist whether the plant makes ten units a month or ten thousand a day. What changes with volume is the frequency with which each is performed and the medium in which it is recorded — a card file, a spreadsheet, or an enterprise system — and neither of those is a change in the design of the system.

What genuinely does change the design is the number of distinct end items, the depth and breadth of the bill of material, the number of routings and the number of work centres they pass through, whether the flow is a line or a jumbled job shop, and how much of the demand is known rather than forecast. A hundred-part product built to order on a job shop routing needs order release, detailed loading, dispatching and expediting no matter how few are made; a single-part product built to stock on a paced line needs a rate, a material feed and very little else no matter how many are made. Volume enters this picture only because high volume usually justifies a line, and a line has a simple control problem — but it is the line that simplified the control, not the volume.

There is a second, subtler reason. The cost of the control system scales with transactions, while the benefit scales with the value of the inventory and capacity being controlled, and both scale roughly with volume, so the economic case for a given level of control sophistication is surprisingly insensitive to output. Doubling the volume doubles both sides of that comparison. This is why small plants that adopt proportionally simpler control systems usually find, on examination, that they have adopted a less formal version of the same system rather than a different one: the same functions, performed by one planner rather than a department, and recorded on fewer documents.

Part (b) — the order types for a small flashlight plant

A flashlight is a short, wide bill of material: a moulded case and lens, a reflector, a switch, a lamp or LED module, a small printed-circuit board on the modern versions, a spring, contacts, and a carton. Some of those are made in-house from moulding and stamping; most are bought. Organising the plant, I would establish four kinds of order, which is the usual answer for a small plant and is the smallest set that keeps the different kinds of commitment distinct.

The first is the shop order (also called a manufacturing, work or production order): the authority for a work centre to make a stated quantity of a stated part to a stated routing by a stated date. It carries the routing, the operation sequence, the standard times and the material requisition, and it is the document against which labour and material are charged and progress is reported. In the flashlight plant one shop order covers a moulding run of cases, another a stamping run of contacts, and another the final assembly and pack.

The second is the purchase order: the same authority directed outward, committing the company to a vendor for a quantity, a price, a specification and a delivery date. Because most of a flashlight is bought, this is the order type that carries most of the plant's material value, and it is the one whose lead times drive the master schedule. For high-usage standard items such as batteries or cartons I would use a blanket purchase order with scheduled releases rather than a separate order per lot, which converts a repetitive purchasing transaction into a single negotiated commitment plus cheap releases.

The third is the tool or capital order, covering the moulds, dies, fixtures and gauges that the product needs before any shop order can run. It is separated from the shop order because its cost is a one-time charge against the product rather than a running cost, because its lead time is much longer, and because it is authorised at a different level in the company.

The fourth is the standing or expense order for work that recurs and is not attached to a unit of product: maintenance, tool sharpening, housekeeping, rework and sampling inspection. Giving this work its own order type keeps indirect cost out of the product cost and stops the shop-order system from being clogged with tiny tickets.

In a plant this size I would deliberately stop at four. A fifth type — separate subcontract orders, or separate rework orders per lot — is defensible in a larger plant, but each additional type is another document to raise, track, close and account for, and in a small plant the administrative burden of the control system is a real cost that shows up directly in overhead.

Part (c) — centralized against decentralized dispatching

Dispatching is the act of deciding, at the moment a work centre becomes free, which of the waiting jobs it will start next, and of releasing the paperwork and material to make that happen. In centralized dispatching, that decision is made in a production control office that holds the queue for every work centre in the plant; in decentralized dispatching, it is made at the department by the foreman or a department dispatcher, working within a schedule the central office has set.

The advantages of the centralized form all flow from having the whole plant in one view. Priorities are consistent across departments, so a job that is late does not get expedited in one department and left in the queue in the next. The office sees the downstream consequences of a sequencing choice and can protect a bottleneck or a committed shipping date that a single foreman cannot see. Records are uniform and complete, which makes load reports, capacity analysis and delivery promising possible. Scarce dispatching skill and, historically, scarce computing capacity are concentrated rather than duplicated. And the control function is independent of the operating departments, so the measurement of performance is not made by the people being measured — an argument that matters more than it first appears.

The advantages of the decentralized form all flow from proximity. The foreman knows which machine is running rough, which operator is best on a difficult setup, which two jobs share a fixture and should therefore run back to back, and which material has just arrived damaged; none of that is in the central file, and all of it changes the right sequence. Response is immediate, with no delay while a request travels to the office and an answer returns, and no paperwork loop for a change that lasts one shift. Setup savings from sequencing similar work together are captured naturally, since only the person at the machine can see them. Responsibility and authority sit in the same place, which is a strong motivator, and the foreman who is accountable for departmental output is the one making the decisions that produce it. Group technology cells make this argument stronger still, because the cell has already been designed so that the interactions a central office would manage happen inside one team.

The two are not exclusive, and the practical answer in most plants is a hybrid: the central office sets and maintains the priorities and the due dates, publishes a dispatch list, and holds the records; the department chooses the sequence within that priority framework and reports back. That arrangement takes the cross-plant consistency of the centralized form and the local knowledge of the decentralized one. The balance shifts toward central control as the routings get longer and more interdependent and as delivery promises get tighter, and toward local control as the plant moves to cells, to flow lines, or to short simple routings — which is exactly the situation of the small flashlight plant in part (b), where decentralized dispatching under a central schedule would be the right choice.