22-Mec-B4 Integrated Manufacturing Systems · December 2016

Question 6 of 6: Production Planning and Control — Volume, Order Types and Dispatching

Nivaar worked solution (AI-drafted; not reviewed by a licensed engineer)

Notes on this paper

Paper format. 07-Mec-B4 — Integrated Manufacturing Systems, National Exams December 2016. Three hours, open book, any non-communicating calculator permitted. Six questions are printed; any five constitute a complete paper and all questions are of equal value, so each is worth 20 marks on a five-question basis. Only the first five questions appearing in the answer book are marked. All six are solved here.

Reference texts. The paper draws on the operations and facilities side of manufacturing engineering rather than on process metal cutting, so the useful shelf is:

E. S. Buffa and R. K. Sarin, Modern Production / Operations Management, 8th ed. — facility layout and operation sequence analysis, materials handling, inventory systems, production planning and control, dispatching.
M. P. Groover, Automation, Production Systems, and Computer-Integrated Manufacturing, 5th ed. — group technology, cellular manufacturing and rank order clustering (Ch. 15).
R. B. Chase and F. R. Jacobs, Operations and Supply Chain Management, 16th ed. — forecasting model selection, inventory control, statistical quality control.
S. Nahmias and T. L. Olsen, Production and Operations Analysis, 7th ed. — economic order quantity, order intervals, safety stock.
D. C. Montgomery, Introduction to Statistical Quality Control, 8th ed. — inspection strategy, quality information systems, process capability.

Canadian practice is assumed throughout: handling and lifting design is governed by the applicable provincial occupational health and safety regulation and by CSA standards (for example CSA B335 for lift trucks), and quality records are kept to satisfy ISO 9001 as adopted by CSA.

Question 6: 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 production planning and control system. The claim rests on a distinction between the functions a planning and control system must perform and the mechanisms by which it performs them. The functions are invariant. Whatever the volume, someone must translate demand into a production plan; explode that plan into material and component requirements; determine routings and standard times; load the work onto capacity and schedule it against dates; release and dispatch the work in a chosen sequence; follow it up and expedite what is late; and feed actual performance back so that the plan can be corrected. A shop building one prototype and a plant building a hundred thousand units a month both perform every one of those functions, and omitting any of them causes the same failures in both.

What volume changes is where the work of planning is done, not how much of it there is in total. At high volume the plan is embedded once, permanently, in physical arrangements: the routing is built into the line, the sequence is built into the conveyor, the balance is built into the station assignments. The planning effort is large but it is spent once at design time, so the day-to-day control that remains looks simple. At low volume nothing can be embedded, so the same decisions must be made again for every order, and the planning effort appears as continuous paperwork. The total planning content is comparable; only its timing differs. The apparent simplicity of a high-volume control system is therefore an illusion created by looking only at the operating phase.

A second reason is that the information required is attached to the order rather than to the quantity. A shop order carries the same fields — part number, quantity, routing, materials, dates, status — whether it authorises 10 pieces or 10,000, so the clerical and computational cost of control is very nearly per-order, not per-unit. Volume affects only the cost of control per unit produced, which is an economic consequence, not a design difference. It is precisely because the system is the same that a high-volume plant can afford a more elaborate one.

Third, the variables that genuinely do drive the design of a planning and control system are others. Product variety and the stability of the mix determine how much re-planning is needed; the nature of demand, whether made to stock or made to order, determines whether the system is driven by forecast or by order; the depth and complexity of the bill of material determine whether simple order-point logic suffices or requirements planning is necessary; routing complexity determines the difficulty of loading and sequencing; and the process type, from project through job shop and batch to line and continuous, determines almost everything else. Volume enters this list only indirectly, through the process type it makes economic. Once the process type is settled, doubling the volume changes the numbers in the system without changing the system.

The claim should not be pushed past its limit, and saying so is part of a good answer. At very high and very stable volume, control does eventually change in kind rather than in degree: it moves from order-based tracking to rate-based control, in which a takt rate and a pull signal replace shop orders and progress reporting entirely, as in a kanban-controlled assembly line. That is a genuine difference of mechanism. The correct statement is therefore that volume has little effect on what a production planning and control system must accomplish, and a large effect only on the physical form the system takes once the process has been chosen to suit that volume.

Part (b) — types of orders in a small flashlight plant. An "order" in production control is a written authority to commit resources, and each type exists because a different resource is being committed and a different person must act. A flashlight is a simple assembled product: a moulded or drawn barrel, an end cap and switch, a reflector and lens, a bulb or light-emitting diode assembly, a spring and contacts, batteries and packaging. For a small plant making a few models of that product I would use five order types, and no more, because every additional type costs paperwork that a small plant cannot absorb.

The assembly or final production order. This is the master authority for a lot of finished flashlights of one model. It is released against the stock replenishment plan or against a customer order, it states model, quantity and completion date, and it is the document against which finished goods are received into stores. Everything else in the plant is ultimately generated by it.
Shop orders for manufactured components. One per part made in-house — barrel, end cap, reflector, contact strip — each carrying its own routing, quantity, material issue and dates. These are what the machine and moulding departments actually work to, and they are the vehicle for labour and scrap reporting. In a small plant with a stable product these can often be released on a simple two-bin or reorder-point basis rather than exploded from every assembly order.
Purchase orders, raised from purchase requisitions. For everything bought in: bulbs or diodes, batteries, springs, moulding resin, fasteners, printed cartons and blister packs. The requisition is the internal authority created by the planning system; the purchase order is the external contract. For repeating items I would place blanket orders with releases against them, which reduces the number of transactions sharply without losing control.
Standing orders for expense and indirect items. A single open authority covering low-value consumables — adhesives, cleaning materials, hand tools, lubricants — charged to a department account and reviewed periodically rather than ordered item by item. Treating these like production parts would swamp a small purchasing function for no benefit.
Tool and maintenance orders. For the moulds, jigs, fixtures and gauges the product needs, and for plant upkeep. These consume real capacity in the toolroom and must be scheduled alongside production work, so they need a document; but they are not product orders and must not be costed as if they were.

A sixth type, the rework or salvage order, is worth defining even in a small plant, because rework that is done without a document is rework that is never costed and never counted; but it can reasonably be handled as a special class of shop order rather than as a separate form. Deliberately excluded are the sub-assembly order and the engineering change order: with only a few models and short assembly sequences there is no reason to stock sub-assemblies separately, and design changes at this scale can be controlled by revision level on the existing drawings. The principle behind the whole selection is that a small plant should carry the smallest set of order types that still separates the distinct commitments of internal capacity, external money, expense and tooling.

Part (c) — centralized against decentralized dispatching. Dispatching is the act of releasing work to a specific work centre in a specific sequence, together with the papers, materials and tools it needs. It can be done from a central dispatch office holding the orders for the whole shop, or by each department deciding its own sequence from the work it holds.

Centralized dispatching has the advantage of a complete picture. One point knows the status of every order in the shop, which means sequencing decisions can be made in the interest of overall due-date performance rather than of one department, load can be balanced across centres, and a bottleneck can be protected deliberately. Priority rules are applied consistently rather than reinterpreted by each foreman, which matters because locally sensible decisions — running the easy job, or the one that keeps a favourite machine busy — routinely damage the schedule as a whole. Expediting becomes tractable because one office can say where any order is. Material, tooling and drawings can be verified as available before an order is released, which prevents work starting and then stalling. The clerical function is performed once for the whole plant instead of being duplicated in every department, records are not held in several inconsistent places, and the arrangement lends itself naturally to computerisation and to reporting shop-wide performance to management.

Decentralized dispatching has the advantage of speed and local knowledge. The person deciding the sequence is standing in the department and knows that a machine is down, an operator is absent, a fixture is broken or two jobs can share a set-up — facts that reach a central office late if at all. Response to disruption is therefore immediate, and the shop keeps running while a centralized system would still be re-planning. Communication and paperwork are lighter, the delay between a machine falling idle and the next job arriving at it is shorter, and the foreman who is held accountable for departmental output is also given the authority to sequence it, which is sound organisation and improves motivation. The arrangement is also robust: the failure of a central system does not stop the shop.

The comparison resolves according to the structure of the work. Centralized dispatching suits shops with long, complex, interacting routings, many departments, tight due-date commitments and a shortage of experienced supervision — conditions typical of a large job shop. Decentralized dispatching suits short routings, self-contained departments or group technology cells, high disruption rates and experienced foremen — and it is the natural partner of the cellular layout derived in Question 5, since a cell that owns a part family can sequence its own work without reference to anyone. In practice most plants use a hybrid: the central office releases orders, sets priorities and reports status, while sequencing within the released set is left to the department. That combines the global view with the local reaction time, and it is the arrangement I would recommend for any plant large enough for the question to arise.

Topic. Design determinants of a production planning and control system, the order documents it uses, and the organisation of dispatching.

Key relations. Planning and control functions (plan, explode, route, load, schedule, dispatch, follow up, feed back) are invariant with volume; control cost is per-order rather than per-unit; dispatching organisation trades a global view against local reaction time.

Why this works. The unifying idea across the three parts is that a control system is shaped by the variety and complexity of what it must control, not by the quantity. Volume determines the process, and the process determines the control system; the arrow runs that way, which is why volume appears to matter but does not directly. The same idea explains why a small plant needs fewer order types — fewer distinct kinds of commitment to authorise — and why a cellular shop can decentralise dispatching: the cell has already absorbed the interactions that a central office would otherwise have to manage.

Common pitfall. In part (a), agreeing with the proposition without identifying what does drive the design, which leaves the answer with nothing to say. In part (b), listing order types generically instead of committing to a number and justifying each one against the flashlight product — the question asks how many you would use. In part (c), describing the two systems separately without ever comparing them or stating the conditions that favour each.

Source. Buffa and Sarin, Modern Production / Operations Management, 8th ed., production planning and control and shop floor control chapters; Groover, Automation, Production Systems, and CIM, 5th ed., Ch. 25–26 (shop floor control, order release, order scheduling, order progress); Chase and Jacobs, Operations and Supply Chain Management, 16th ed., Ch. 22 on scheduling and priority rules.

← Question 5 Back to the paper →