- Email: [email protected]
Takt Planning: A technique for materials requirement and production planning
Author’s Accepted Manuscript Cover-Time Planning/Takt Planning: A technique for Materials Requirement and Production Planning Anders Segerstedt
www.elsevier.com/locate/ijpe
PII: DOI: Reference:
S0925-5273(17)30124-X http://dx.doi.org/10.1016/j.ijpe.2017.04.006 PROECO6700
To appear in: Intern. Journal of Production Economics Received date: 18 April 2016 Revised date: 23 February 2017 Accepted date: 12 April 2017 Cite this article as: Anders Segerstedt, Cover-Time Planning/Takt Planning: A technique for Materials Requirement and Production Planning, Intern. Journal of Production Economics, http://dx.doi.org/10.1016/j.ijpe.2017.04.006 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Cover-Time Planning/Takt Planning: A technique for Materials Requirement and Production Planning Anders Segerstedt Industrial logistics, Luleå University of Technology, SE-971 87 Luleå, Sweden [email protected] Abstract Cover-Time Planning, or Takt Planning, is presented. It is a system for calculating material requirements and start of purchases and production. Requested production rates of sales items, or alternative prefabricated modules in stock, are “broken down”, exploded, to create the need for components, for all underlying items (articles) in the Bill of Material. Inventory and already ordered replenishments are compared with the item's desired production rate. How long already made actions are expected to cover the desired expected sales and production rates is compared with the item's lead time; if a forward future shortage is likely the article is signalled for a refill. With examples is described how make-to-order production is done easily. The method is a type of reorder point system, but with time instead of quantity as decision variable. Unlike a traditional reorder point system increases and decreases of production can be planned. Future work load in various production sections can be estimated. It is described how an “Available-to-promise”-system should be designed and used. Cover-Time Planning (CTP) is a complete alternative to Materials Requirements Planning (MRP). CTP responds faster than MRP, since MRP for each structural level uses a batch size to "break down" and to magnify the need at the underlying level. In the end a large material acquisition needs to be ordered maybe just to build a single end item. CTP uses the end item requirement rates on all structural levels shifted with the lead times, when customer demand changes, the entire production chain react simultaneously. Keywords: Multi-level inventory and production control, Cover-Time Planning, Takt Planning, Materials Requirements Planning, Make-to order, Modules
1. Introduction and background Companies want to deliver to their customers with short delivery times, it is a competitive tool. Companies do not want to have huge inventories and capital tied up in unsold material that suppliers have already been paid for. In order to achieve a short and competitive delivery time purchase, production and preparation must have started before the final customer order arrives; sometimes long before. - What should we start, and after what decision rules should we start? It is a question that a lot of companies struggle with. What can we store? What production will start before the final customer orders come in? Just those items/products that we can appreciate a demand for; a demand rate at which it is desired to satisfy customer demand. If we do not know exactly the pace customers will ask for we must decide the rate we want to accomplish if customers would like to have the product. A way to keep down the amount of capital tied up in inventory and work-in-process (WIP) as well as to establish short delivery times is to create modular products. When the customer order arrives, the sales and end item is assembled from modules. The same module will be used in several end items; thereby the modules help to create a variety of end items. Modules are kept in store and a large stock of different end items is avoided. Many manufacturing companies have this option or opportunity but they must also have an administrative tool and control system that makes this possible and efficient. For some companies, if there is no final product finished, there is nothing to sell. The market demands that the product is available the instant the customer wants the product; sausages, cultured spices, chainsaws, some type of electric motors etc. – This article presents a technique to start production before the final
customer order arrives for both types of products and companies. This is important because it is never financial goals that create profitability, it is how effectively the company can manage its processes and satisfy current and changing demand. To give a background to and argument for this technique I start with a historical look back over my previous activities and experiences: 1.1 Ericsson Information Systems (1984-86) A problem that we had when I was working as a manager of Production Control at Ericsson Information Systems’ plant in Linköping 1984-86 was low turnover of inventories of components and semi-finished products. We said at that time that this was due to the very long delivery times from suppliers and that we had no reporting of each operation on production orders; but there were other reasons as well. The MRP execution results "conned" us to start unnecessary purchase orders, and production orders of the semi-finished and finished products. We started not only proposals for the current week but proposals for several weeks’ ahead (made planned orders to scheduled receipts). Thus, it was not just suggested planned orders it became real purchase orders. With these purchase orders the production control department could “chase” the purchasers; “where do you have this material?” and by that avoid future material shortages. The MRP run rescheduling meant that the earlier-and postponement list (rescheduling- in and reschedulingout) was huge, the only thing that there was time to correct was some rescheduling in. The result was much work-in-process (WIP) and long lead times. At Ericsson I came in contact with the consulting company Mysigma and their ideas. At that time Mysigma analysed their customers' inventory; how many days or months of demand each items’ inventory corresponded to. The items were also sorted in order of this cover-time and/or run-out time: a simple but brilliant idea. Mysigma’s founder Lars O. Södahl (Södahl, 1984) introduced the Toyota Production System from Shingo (1981) in Scandinavia. 1.2 Benzlers (1986-91) In 1986 I moved to the company Benzlers where I continued working with production, planning and production control. Benzlers was at that time a well-run, well invested company with wire guided automatic forklifts, with a modern computer system reporting every operation in the production orders etc. Benzlers produced mechanical gears and its components; gearwheels and pinions with many operations in long production chains with turning, external hardening, grinding etc. (At Benzlers, and Ericsson, I learned a lot about operations that was not covered as a part of my education at Linköping Institute of Technology (1969-1973, MSc Industrial Engineering); which is still considered to be one of the top educations in Sweden. E. g., Material Requirements Planning (MRP) was not taught at that time in Linköping. Even though MRP is available and used in most computer systems for material and production control most textbooks used at universities and high schools present a superficial introduction to MRP; although Jacobs et al. (2011) is an exception.) Now in retrospect, an error that we also made at Benzlers was that too much WIP was released in production. But we felt compelled to do as the MRP runs told us to do; how could we do otherwise? An exception was castings, bulky and expensive materials, an awake buyer did not want to fill our warehouse; planned orders were ignored and he bought new castings
for the gear housings according to current inventories and our Master Production Schedule (for end items). This was possible because the end items had a limited number of gear houses that could be overviewed manually. When we registered and tried to dispatch an assembly order of a gear, we discovered that it was not possible to release the order without a material shortage due to one pinion being missing at the same time as there were two order quantities or batches, of a different pinion or gearwheel! The Bill of Material (BOM) was correct; the MRP calculation knew and expected that we needed two pinions and two gearwheels, of various dimensions, to put together the complete gear. But the production flow had become unbalanced, an assembly was not possible!? One reason for this was that the routings of the items were very long; they contained many operations and also external heat treatments. Therefore and because of much WIP, individual production orders of the same item had very long lead times (throughput times). When different orders of the same item appeared together in the gear hobbing operation, the operators saved setup time by manufacturing both batches at the same time. Thereby it delayed other orders and items. We noticed this habit and made the disadvantage of it clear; but still very long and varying lead times continued. (Merger of production orders does not exist today? Yes, the same things still happens in lean oriented large well-known successful Swedish companies (Source: Master Theses)).
The problem with the lack of an article and two available order quantities of another article made me consider the advantages and disadvantages of MRP’s. I did a simulation model, with relatively simple product structures (BOMs), with variations in the lead times, and with net requirements planning exercises every week for a long simulation time with lead times uniformly distributed. As in most real cases, the MRP-calculation’s proposals on advancing and/or postponement of already registered manufacturing orders (rescheduling of scheduled receipt) could not be addressed. The results showed that there was always a "delay" for the Master Production Schedule (MPS), production quantities could not be delivered in time, a significant part only in past time. From time to time material was missing and it was not possible to complete the final end item in time because no safety stocks were used. The lateness was the “safety stock” necessary to accomplish the production rate. The accomplished production rate corresponded to the production rate(s) of the MPS; this delay was a consolation for that we were always accused of being not competent due to the delay in our MPS at Ericsson and Benzlers. At the same time I tested another idea than MRP, Cover-Time Planning (CTP); description of this method or technique is the purpose of this article. The study showed that with variations in demand and lead time, a simpler model than MRP is able to provide the same and better results in terms of lateness, inventories, WIP and its variety and at the same time achieve the same production rate. This became my Licentiate thesis (Segerstedt, 1991). Today if I am studying my licentiate thesis again I see opportunities for improvement; computer capacity has increased, I was not well informed about "Little's formula" and I did not mention in detail a major problem with MRP, make-to-order; how to deal with customer order production. Cover-Time Planning has further been described internationally (Segerstedt, 1995 and 2006), but mostly only in Sweden under the name “täcktidsplanering” in textbooks and popular science journals (e.g. Segerstedt 2005, 2009 and 2016). The idea has been further developed so the description presented here differs and goes deeper compared to early presentations.
1.3 Last 20 years After my Licentiate degree I started as a teacher at Mälardalen University, both in business administration and material and production control. Over the last 20 years I have supervised and taken part of a considerable amount of master theses in material and production control/logistics. This has given me insight and information from a large number of companies, some I have also visited. It is not a revolutionary development I have experienced. Real time computer solutions have been introduced in software for material and production control but the steering principles are basically the same as in the mid-1970s. Kanban, Theory of Constrains (The Goal from Goldratt and Cox (1984)) was very interesting readings but I have seen few successful implementations. Trends have swept through industry since my entrance in the mid-1970s; old things have got new names to be emphasized by consultants. In 1980s previous possibilities to reduce taxes by inventory write-downs disappeared (in Sweden) so a great campaign started for capital rationalization and to decrease inventories. 15 years after I heard/read the word "lean" it became a trend in Swedish industry and is often presented as a type of new science. 5s, eight wastage, etc., I have read about in numerous student works. But what's really new science since Henry Ford's Assembly line and Taylor (1911)? Many peer reviewed articles have been published but few of them have changed the behaviour and routines in industry. Orlicky (1975) and MRP on the contrary has influenced behaviour and routines concerning operations all over the world -, supply chain- and logistics management to an extent that has not really been highlighted (exception Jacobs et al. (2011)). Jonsson and Mattsson (2008, 2014) point nor at any revolutionary development; they show that the update of parameters of material and production control is very rarely but they see an improvement in comparison with previous studies. 1.4 Make-to-Order; Customer orders and production When a company has registered a sales order (customer order) they want it to take part in the MRP run so that no material procurement or production is forgotten. Then the Master Production Schedule (MPS), a forecast of future deliveries, must be reduced so that in the next MRP calculation material for the new customer order and the old MPS are not both generated. This would soon fill up WIP and inventories. If you have a limited number of end items, you can do this manually, as we did at Benzlers, by reducing the forecast in the interval of time in which the customer order is to be delivered and the amount the customer order contains (see: Segerstedt, 2002 (case Volvo CE); Segerstedt, 2009). If the number of end items are larger the MRP user requires a Master Production Schedule Planning (MPSP)system that consumes the forecast (Jacobs et al. (2011) and also helps to establish a new MPS for the next MRP calculation (Segerstedt, 2002 (case ABB Motors); Segerstedt, 2009). With Cover-Time Planning/Takt Planning the customer orders interfere in the calculations in a different way than with MRP which will be shown in the forthcoming text. Then the customer orders do not have to consume the plan; the forecasted demand and production rate is consumed by time. If the amount of customer orders does not come up to the expected demand rate, the replenishment of the item decreases immediately; the inventory and the WIP is enough for the future demand rate. Thus WIP is immediately somewhat faster than MRP.
2. Total lead time or cumulated lead time To describe CTP/Takt Planning we start with a simple Bill of Material (BOM) or product structure, see figure 1. E101
S207
S202
S301
B401
B305 B400
B400 Figure 1. Schematic product structure (BOM)
The end item article E101 is assembled from two semi-manufactured items (S202, S207) and one raw material (B401, a buy article); reality is generally much more complicated with many more articles, but this serves as an example. The number of structural levels is 4. The total or cumulated lead time for E101 starts with article B400, from this raw material article S301 is manufactured and from there S202. If we shall increase production of E101, the ordering of B400 must increase at once. Then we have to wait until the increased amount arrives, when it does we can increase the production of S301. Then it takes some time (lead time of S301) until we can start an increased production of S202. When we have finished the increased production of S202 (the lead time of S202) we can start the increased production of E101. Then, when this production is finished, at last we have increased the production of E101! Using MRP requires that a lead time is estimated and registered for every item (in Figure 1). To suggest an increase of production for E101 in close time means that MRP will correctly suggest the start of planned orders in past time! The sum of the lead times for items E101, S202, S301 and B400 must be kept track of and within this timeframe the production rate of E101 should not be increased; this to avoid proposed purchase of B400 in past time. This time interval is the total or cumulated lead time for E101. But if for e.g. B305 had a very long lead time then the longest “way” through E101’s BOM could be the sum of E101, S202 and B305; the cumulated lead time for E101. Using Cover-Time Planning or Takt Planning demands, like MRP, an estimated lead time for all items involved in the materials requirement planning. Disregarding the total lead time and increasing the production rate in near time will become a failure, with severe delivery delays despite the technique used. Decreasing production rates in near time will not create the desired consequences either; already ordered raw material and semi-manufactured items will continue to flow in. The used planning horizon must exceed the longest cumulated lead time of all end items independent of the technique for multi-level inventory control that is used.
3. Cover-time and Cover-Time Planning (Takt Planning) For how long will steps already taken cover expected demand for the item in question? The answer to this question we call cover-time. A more precise definition of cover time will be presented below, to make it distinct from run-out time. If the cover time of the article is shorter than the lead time there is certainly a high risk of a future shortage of the article: a risk that must be avoided. On the contrary, if the cover-time is very long it presents a signal that the inventory may need to be scrapped. Cover-Time Planning is a technique that converts expected demand rates for end items into decided production rates and explodes this demand rates’ gross to all components in the Bill of Material. Cover-Time Planning is a type of reorder system based on time instead of quantity and with a forward visibility (like MRP). A “master production schedule” in a “takt”, units per production day, is reported for all “end items”. In such a way that a production rate is supposed to be valid until another production rate is reported. Thus it is possible to plan ups and downs of production. From these “master production schedules” desired production rates are calculated for all articles considering lead times and connections in the BOMs (product structures). 3.1 Example: Cover-Time Planning The following product structures are prevailing between the articles S202, S207, S301 and B400. (The BOMs for S202 and S207 put together):
S202
S207
S301
B400 Figure 2. Product structures in our example
S202 and S207 in Figure 2 are now end items. (An explanation of why this is will be presented in section 4 Make-to-Order.) Raw material B400 is used for both S207 and S301 (thereby it is also a part of S202). To produce one unit of S202 requires two units of S301, and to produce one unit of S301 requires one unit of B400. To produce one unit of S207 requires two units of B400. Table 1 presents further information: Table 1. Used order quantity and lead time Product/ article
Lead time in time periods
Order quantity
S202 S207 S301 B400
4 5 10 15
15 10 40 50
Articles S202 and S207 have desired and decided production rates (shown below in Table 2 and 3). Considering already existing inventories and scheduled issues as well as scheduled receipts: when must the production of S202, S207 and S301 and the purchase of B400 start in order to fulfil planned production rates for S202 and S207? This will be presented in the forthcoming text.
Table 2. Situation S202 period 1
Article S202 Demand rate Sch. issues Sch. receipts Inventory "Supply" Pl. orders
14 2
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
5
5
5
5
5
5
5
5
5
5
6
6
6
6
6
6
6
6
6
6
5
4
7
1
5
6
11
7
0
2 15 13
12 31
7
10
3
1
1
-9
-9
-9
-9
-12
-12
…
(11)
(6)
(1)
(11)
(5)
(14)
(11)
(11)
(11)
15
15
15
15
A-T-P
2
1
-2
3
4
0
-1
5
-5
5
6
6
3
6
…
Cum. A-T-P
2
3
1
4
8
8
7
12
7
12
18
24
27
33
…
15
15
We are in the absolute beginning of time period (day) 1.”Supply” is already taken actions to cover future demand. “Supply” is equal to: current available inventory plus all scheduled receipts (cumulated) minus scheduled issues in past time. Scheduled issues in past time should have been delivered and they are not future demand, they are past demand, therefore available inventory must be reduced. Due to that the customer does not want a partial delivery, and/or another article which should also be send to the customer is currently in shortage, it is possible that we have a positive available inventory and still scheduled issues in past time. At the moment “Supply” is 2+14 + 15 = 31 (which is shown in Table 2). In table 2 it can be seen that “Supply” is placed in period (day) 5; because the lead time of a new replenishment of article S202, a scheduled receipt, that starts now can first be delivered in period (day) 5.
Supply will cover the demand rate for H202 during 31/5= 6.2 time units. Cover time (=6.2) is larger than the lead time plus inspection time (4+1= 5), therefore no replenishment is necessary at the moment. The expected demand that we should prepare for from now, including period 5, is 25 (= 5· 5), and actions already taken add up to 31. So at the moment no more replenishment is necessary. The first thing that happens in period 1 is that a decision is made as to if replenishment is needed or not. Using Cover-Time Planning a signal for replenishment is created when: (1) Cover-time < lead time including inspection interval + possible buffer time (2) Planned inventory within lead time plus inspection time < 0 Number (1) is the main principle. Number (2) is used when there is no forecasted demand rate but scheduled issues (customer orders), and also to complement number (1) and alert for possible mistakes that have been made. What deliveries, new scheduled issues, of S202 can be promised without causing shortages or late deliveries? The principle is: from now and in the future cumulated scheduled issues (customer orders) may not exceed cumulated forecasted demand rates; otherwise customer
delivery that is not planned for, or has not been forecasted, is promised. Then this material is not on its way downstream, shortages and late deliveries will occur. Additionally, a new replenishment cannot happen within the lead time; i.e. no new scheduled issues that make planned inventory in period 5 negative can be promised. Available to promise (A-T-P) in first period (day) is: Current available inventory (on hand) – Scheduled issues in past time + Demand rate in period 1 – Scheduled issue in period 1 (S202: 2 -0 +5 -5 =2). Scheduled receipts are not included in the calculation of available to promise (A-T-P); but cumulated scheduled receipts must exceed the cumulated demand rate. Cumulated scheduled issues (customer orders) may not exceed cumulated demand rates. Scheduled issues in past time should have been delivered and they are not future demand, therefore they must reduce current available inventory. A-T-P in other periods is: Demand rate in that period – Scheduled issue in the same period. A-T-P can be negative and must be allowed to be negative because a delivery larger than original forecasted demand during that period (day) must be possible. Cumulated available to promise (Cum A-T-P) cannot be allowed to be negative, it must always be nonnegative. If it is not, a scheduled issue that has not been forecasted before is promised but these units are not on their way downstream in the supply chain. No delivery of S202 can be promised in Period 1, 2 and 3. If it were the inventory within its lead time would become negative (according to Table 2). In period 5, 12 units seem to be available for deliveries. A promised delivery of 12 units in period 5 immediately changes planned inventory, A-T-P and Cum A-T-P (see Table 3).
Table 3. Situation S202 period 1 (12 extra units period 5) 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Demand rate
5
5
5
5
5
5
5
5
5
5
6
6
6
6
6
6
6
6
6
6
Sch. issues Sch. receipts Inventory "Supply"
5
4
7
13
5
6
11
7
0
2 15 13
0 31
-5
-11
-11
-21
-21
-21
-21
-24
-24
…
A-T-P
2
1
-2
3
-7
0
-1
5
-5
5
6
6
3
6
…
Cum. A-T-P
2
3
1
4
-3
-3
-4
1
-4
1
7
13
16
22
…
-
14 2
10
3
Therefore, promising 12 units in period 5 is nothing to recommend. The scheduled issues in period 6 and 7 will probably be delayed. There is still time to register and start a scheduled receipt (replenishment) for period 6, but downstream components are not prepared for this increase and a possible shortage of components will delay the replenishment. Many of the forecasted demand rates are consumed, so a new delivery larger than one unit cannot be promised until period 11 (see table 3). A delivery of 12 units should therefore not be promised until period 10, according to Cum A-T-P in table 2. In the case where there are never any delays, never any mistakes in reservations of orders, no scrap rejects or other disturbances; then ”Available-to-Promise” could be described rather simply. What is described here is logic and a technique that must deal with stochastic events, even correcting previously made mistakes; which makes Cum A-T-P the important decision variable.
Information about starts and deliveries of S202 must be created in order to be able to estimate future work load. Future planned orders for replenishment can be calculated from ”supply” and already decided scheduled receipts and the existing demand rate. The next replenishment of S202 must come in period 7. There will therefore be a fictive planned inventory of 31+15(5· 7) = 11 units at the end of period 7. (Total demand from period 1 to 7: 5· 7= 35; total “supply” in period 7: 31 +15 = 46; 46 -35 = 11) This means that the next incoming delivery of 15 units must take place in period 10 (see table 2). At the end of period 10, fictive planned inventory is 11; hence the next replenishment must come in period 12. The end of period 12 inventory is 14 units etc. Planned orders are only calculated when an analysis of work load is necessary and done. The lead time of the article determines the start of the planned order, and when different operations are performed, distributed between the start and completion time (like the usual procedure when using MRP). Table 4. Situation S207 period 1
Article S207 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
3
2
3
3
7
2
-1
4 10 5
2
2
2 14
0
0
-7
-7
…
A-T-P
4
-1
-2
-1
2
3
-2
3
-4
3
3
3
3
…
Cum. A-T-P
4
3
1
0
2
5
3
6
2
5
8
11
14
…
Demand rate Sch. issues Sch. receipts Inventory "Supply"
5
7
”Supply” is 7-3+10= 14 units. A scheduled issue in past time should have been delivered; therefore it is not included in ”supply”. “Supply” is the current amount already decided we have for satisfying future demand. Cover-time for S207: 5+ (14- 5·2)/3 = 5+ 4/3 = 6.33; expected demand we should prepare for until period 6 is 13 units. ”Supply” is 14; thus replenishment can wait until next period. But, with a buffer time, necessary as a type of safety stock, replenishment for period 6 should be performed now. A scheduled issue larger than 5 cannot be promised until period 10. If it is possible an order should perhaps also be made in part deliveries to enable deliveries to other customers. Article S301 The demand rate for S301 depends on S202’s demand rate. S202’s demand rate increases in period 11 and therefore the demand rate for S301 increases four periods earlier (due to the lead time of S202); in period 7. Since two units of S301 require producing one unit of S202 the demand rate is twice as large as for S202.
Table 5. Situation S301 period 1 Demand rate Sch. issues Sch. receipts Inventory "Supply" Pl. orders
30 38 20
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
10
10
10
10
10
10
12
12
12
12
12
12
12
12
12
12
12
12
12
12
28
28
28
40 68
68
69
68
108
108
108
108
…
108
(16)
(2)
(30)
(18)
(6)
(34)
(22)
(10)
(38)
40
40
40
40
40
”Supply” is 20-30+38+40+40= 108 units. The scheduled issue in past time is the manufacturing order of S202 (15 units in period 4); not yet started with a withdrawal of S301. Cover-time for S301: 6+ (108- 6 10)/12 = 6+ 48/12 = 10 < 10 +1; expected demand we should prepare for from now until period 11 is 120 units; ”Supply” is 108. A new scheduled receipts for delivery in period 11 must by registered and started at once. Article S301 is manufactured in the own production facility where sometimes there is a need to analyse future workload. At the moment there are 3 scheduled receipts, next replenishment originating from those, ”supply” and the expected demand rate is required in period 11, then in period 14 and 17. Article B400 The demand rate for B400 depends on the demand rates of both S207 and S301. S207’s demand rate increases in periods 6 and 16. Therefore the demand rate of B400 increases in periods 1 and 11. The demand rate of S301 increases in period 7 and therefore the demand rate of B400 increases in period (7 -10 = -3), i. e. it has already happened. The demand rate in period 1 is the sum of the demand rate for S207 multiplied with 2 in period 6 and the demand rate for S301 in period 11. Table 6. Situation K400 period 1 Demand rate Sch. issue Sch. receipts Inventory "Supply"
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
18
18
18
18
18
18
18
18
18
18
20
20
20
20
20
20
20
20
20
20
50
50
50
50
95 295
Cover-time for K400 is 10+ (295- 10 13)/20 = 10+ 115/20 = 15.75; 15.75 15+1, a new purchase order must be ordered at once. 4. Make-to-Order, Production of customer orders Figure 1 consciously shows E101 as an end item the company delivers to its external customer. E101 is design and assembled from different modules. Like a Wheel Loader at Construction Equipment or an electric motor at ABB Motors. E101 is unique and we do not know in advance when and how the customer wants to compound and equip the product. The article E101 does not actually exist until we have got the order from the customer. As we do not know when and how, we have no estimated demand rate for E101. But for the modules S202, S207 and several other items we have estimated demand rates and materials are on their way downstream in the supply chain to allow a short delivery time for E101 Suppose the company gets an order of 5 units of a product equipped like E101, if this order is attempted to be registered and started earlier than period 10 then the system will/should say it is not possible! Because first in period 10 Cum A-T-P for S207 is equal or larger than 5 units. A lead time for E101 may be estimated to be two time periods. That means that a delivery date of E101 before time period 12 should be signalled as not possible. This will avoid a future late delivery. Segerstedt (2002) shows how both ABB Motors (ABBM) and Volvo Construction Equipment (VCE) worked with fictive end items. Why, because they worked with MRP to accomplish short delivery times they had to start production before the customer orders arrive. They had
so many end items and variants of end items it was not possible to make a master production schedule for all end items. ABBM had only a few fictive items (7) on the highest structural level (0); those were expanded one structural level with fixed percentage distributions. This structural level (1) still consisted of only fictive items. Structural level 2 was also created from level 1 with fixed percentage distribution; structural level 2 consisted of real sellable end items (equipped electric motors), approximately 2000 different products. To promise delivery dates ABBM used a Master Production Schedule System (”available-topromise”-system) in their Baan Triton installation (their ERP (Enterprise Resource Planning)system). The system helped ABBM to construct the next MPS and to set delivery dates that did not override current and previous MPS. In their MRP calculation VCE used two fictive articles (L50 and L70) and several ”options” containing possible modules in a wheel loader; excavator, gear box, motor, sits, air conditioners. The options, with article numbers that all began with ”9-”, were complex products with their own Bill of Material. Those “9-numbers” contained a credit quantity for the article in the fictive end items (L50 or L70) they were substituting. That was necessary to avoid material requirement of both the 9-article and the one in the fictive product structure. 9articles were handled manually in the same way as described earlier in relation to Benzlers. Every 9-article had a MPS, and the MPS’s were decreased if a customer order arrived equipped with this 9-article If ABBM and VCE should use Cover-Time Planning or Takt planning as it is described here; what should they do? ABBM were already working with takt. Every month their sales and operations planning ended up with a decision how many units that should be produced of seven different aggregate (fictive) electric motors per day. 2-300 article numbers are standard products and must be delivered immediately from (two) different sales inventories. A number of aggregate fictive products should be used to explode to real articles in sales inventory like before. Because 2-300 articles is difficult to handle with individual manually set forecasted demand rates. The electric motors are built from modules; modules consumed of customer orders, not stored in sales inventory require a separate expected demand rate. But forecasting modules instead of finished motors would decrease the number of fictive article a lot. VCE does not need any fictive articles; every “9-number” gets an expected demand rate, also the “option” in the previous fictive articles. When a customer order is registered, no quantity has to be decreased from any plan. This will radically decrease the (manual) administration. 5. End comments 5.1 Differences Some find it difficult to see the difference between MRP and Cover-Time Planning (CTP). The main differences are: demand rate instead of a special quantity in a special time; the demand rate ”rules” from now and into the future, the past is ”lost”; no automatic rescheduling of scheduled receipt and their connected schedule issue (only warnings: this is late!); a new customer order does not have to decrease the expected demand rate (=MPS)! The example shows that the customer orders contribute to the calculations in a totally different way than with MRP. The new customer order does not have to “turn off” the old plan (or MPS); instead if no new customer orders arrive CTP quickly “turns off” new
replenishments. (CTP reacts more quickly than MRP because MRP has an order quantity (size) for every structural level that ”presses down” and increase the demand on the lower structural level. In the end there can be a large purchase requirement just to build one single end item. 5.2 Buffering, variants and/or takt We talked about buffering, variants and/or takt in the 1980s when I worked at Ericsson. With a variant buffer we meant an inventory of semi-manufactured items that made it possible to quickly deliver just that specific variant of the product, e.g. a special printer. The total semimanufactured inventory could be allowed to be larger than the total sales of printers. Thus also a type of takt buffering was accomplished; it could be possible to deliver more than the forecast of the product. But we could not figure out how we should solve this with our MRPsystem. With CTP the problem is easily solved; the demand rate for the modules is set a bit larger than expected average demand for the end product. Tied up capital will increase a bit but so does delivery performance. 5.3 Material control and Production Control CTP is not only a system for material control; it also presents information for production control. The estimated demand and production rates used must be possible to achieve both in near time and in future time (due to capacity). Therefore it is important that CTP can create information about workload in different machines and production facilities. The shorter cover time an article has the more important is it that the manufacturing or purchase order of the item makes ready and finishes. Cover-time becomes a complement priority magnitude in which sequence orders should be manufactured in a work station. CTP has a CONWIP attribute: inventory and work-in-process of item i is constrained between a multiple of the used order quantity { (n 1) Qi WIPi I i n Qi ( n 1, 2 , )}. (CONWIP is presented e.g. Spearman et al (1990), Hopp and Spearman (2008), Hopp (2008)) Hopp (2008) define push and pull “A pull system is one in which work is released based on the status of the system and thereby places inherent limit on WIP. A push system is one in which work is released without consideration of system status and hence does not inherently limit WIP.” There are many previous attempts to define push and pull that exist, cf. Pyke and Cohen (1990) and Bonney et al. (1999). Bonney et al. (1999) point out that the definitions of push and pull are inconsistent between different researchers and arguments about performance are peculiar, if the performance of a pull system is poor then it may be suggested that this is because the fundamentals of just-in-time are not being observed. Whereas, if the performance of a push system is poor it is because it is a push system. It seems like Hopp’s definition of limited work-in-process is important for the success of a system and not push or pull. Nicolin (1959) explained that too much WIP would lead to future delays of customer orders. ‘‘Balance flow and not capacity’’ is an OPT-strategy (cf. Goldratt and Cox (1984); Goldratt and Fox (1986)) that implicates control of WIP. Silver et al. (1998) and Hopp and Spearman (2008) show that after WIP has reached a certain level, more WIP does not increase the throughput rate. It only prolongs the lead-time. In addition, Pettersen and Segerstedt (2009) show that much WIP not only prolongs the lead-time or throughput time it also increases its variation. Therefore the clear CONWIP attribute of CTP is important and an advantage compared to MRP.
5.4 Future for Cover-Time Planning!? What is described here is a computer system. When I, many years ago, in the early 1990s, was in contact with and was contacted by, software developers for production and material control, I felt doubt and misbelief. They had invested in MRP and CTP was unproven. But now Jonsson and Mattson (2008, 2014) state that many companies use “run-out time planning” (Jonsson and Mattsson (2008) term run-out time planning is synonymous to cover time planning (täcktidsplanering)) In the form that it is presented here we cannot be sure that it is in many installations; but the idea has spread in an extent that I did not know before (Jonsson and Mattson, 2014). In a given computer system the calculations do not take place in tables like they have been presented here but with entirely different solutions; linked lists etc. A demand rate is in order as long as there is no other. The absolute majority of functions in an ordinary ERP-system can still be used without changes; order entry, routings, in and out of stock, BOM etc. So it is rather easy and requires little work to change to CTP. Make-to-Order creates no problem; inventory and WIP for every item are restricted. It creates information for calculating work load; it is possible to plan future increases or decreases of production! CTP is a type of reorder system, with time instead of quantity as a decision variable. Therefore it should also be compared with other reorder systems. (Companies change from MRP to reorder systems due to dissatisfaction.) CTP avoids reorder system’s disadvantages, forecasts for every item and no future visibility. The author is thankful for valuable comments from reviewers. Referenser Bonney M. C., Zhang Z., Head M. A., Tien C. C., Barson R. J. (1999). Are push and pull systems really so different?. International Journal of Production Economics, 59 (1), 53–64. Goldratt E. M., Cox J. (1984). The Goal - Excellence in Manufacturing. New York: Creative Output/North River Press. Hopp W. J., Spearman M. L. (2008). Factory Physics (3ed.). Boston : McGraw-Hill Publishing. Hopp W. J. (2008). Supply Chain Science. New York : McGraw-Hill Irwin. Jacobs F. R., Berry W. L., Whybark D. C., Vollman T. E. (2011). Manufacturing Planning & Control. New York : McGraw-Hill Irwin. Jonsson P., Mattsson S.-A. (2008). Inventory practices and their impact on perceived planning performance. International Journal of Production Research, 46 (7), 1787-1812. Jonsson P., Mattsson S-A. (2014). ’Best practice’ vid lagerstyrning i svensk industri (in Swedish: Best practise inventory control in Swedish industry), Forsk-ningsrapport. Chalmers Tekniska Högskola och Plan. (http://publications.lib.chalmers.se/records/fulltext/192457/local_192457.pdf) Nicolin C. (1959). Planering av en verkstad med blandad tillverkning - matematiska tumregler till hjälp för att organisera effektiv verkstadsplanering (in Swedish: Planning a Workshop – Mathematical Rules of Thumb for Efficiency; Internal Working Paper). Finspong: Arbetsrapport PM VD4a/59. deLaval Jungström. Orlicky J. (1975). Material Requirements Planning. New York: McGraw-Hill. Pettersen J.-A., Segerstedt A., 2009. Restricted work-in-process: A study of differences between Kanban and CONWIP. International Journal of Production Economics, 118 (1), 199–207 Pike D. F., Cohen M. A., 1990. Push and Pull in Manufacturing and Distribution Systems. Journal of Operations Management, 9 (1), 24-43. Segerstedt A. (1991). Cover-Time Planning - an Alternative to MRP, Linköping: PROFIL 10 (ISBN 91-970074-5-5).
Segerstedt A. (1995). Cover-Time Planning, A method for Calculation of Material Requirements. International Journal of Production Economics, 41, 355-368, 1995 Segerstedt A. (2002). Production and Inventory Control at ABB Motors and Volvo Wheel Loaders, Two Examples of MRP in Practical Use. Production Planning & Control, 13 (3), 317-325. (https://www.researchgate.net/publication/237218389_Production_and_inventory_control_at_ABB _Motors_and_Volvo_Wheel_Loaders_two_examples_of_MRP_in_practical_use) Segerstedt A. (2005). Täcktidsplanering underlättar kundorderstyrd produktion och kontroll av produkter i arbete (in Swedish: Cover Time Planning; facilitates make-to-order production and control of work in process). Bättre Produktivitet, No 3. Segerstedt A. (2006). Master Production Scheduling and A Comparison of Material Requirements Planning and Cover-Time Planning. International Journal of Production Research, 44 (18-19), 3585-3606. Segerstedt A. (2009). Logistik med fokus på material- och produktionsstyrning (in Swedish: Logistics with focus material- and production control). Malmö: Liber (ISBN 91-47-08900-0). Segerstedt A. (2016). Täcktidsplanering/Taktplanering (in Swedish: Cover-Time Planning/Takt Planning). Bättre Produktivitet, No 2. Shingo S. (1981). A Study of the Toyota Production System. Tokyo: Japan Management Association. Silver E. A., Pyke D., Peterson R. (1998). Inventory Management and Production Planning and Scheduling. New York: Wiley. Spearman M. L., Woodruff D. L., Hopp W. J. (1990). CONWIP: A pull alternative to Kanban. International Journal of Production Research, 28 (5), 879–894. Södahl L. O. (1984). Den nya japanska produktionsfilosofin (in Swedish: The new Japanese philosophy of production). Stockholm: MGruppens förlag, Stockholm Taylor F. W. (1911). The principles of Scientific Management. New York: Harper & Brothers
www.elsevier.com/locate/ijpe
PII: DOI: Reference:
S0925-5273(17)30124-X http://dx.doi.org/10.1016/j.ijpe.2017.04.006 PROECO6700
To appear in: Intern. Journal of Production Economics Received date: 18 April 2016 Revised date: 23 February 2017 Accepted date: 12 April 2017 Cite this article as: Anders Segerstedt, Cover-Time Planning/Takt Planning: A technique for Materials Requirement and Production Planning, Intern. Journal of Production Economics, http://dx.doi.org/10.1016/j.ijpe.2017.04.006 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Cover-Time Planning/Takt Planning: A technique for Materials Requirement and Production Planning Anders Segerstedt Industrial logistics, Luleå University of Technology, SE-971 87 Luleå, Sweden [email protected] Abstract Cover-Time Planning, or Takt Planning, is presented. It is a system for calculating material requirements and start of purchases and production. Requested production rates of sales items, or alternative prefabricated modules in stock, are “broken down”, exploded, to create the need for components, for all underlying items (articles) in the Bill of Material. Inventory and already ordered replenishments are compared with the item's desired production rate. How long already made actions are expected to cover the desired expected sales and production rates is compared with the item's lead time; if a forward future shortage is likely the article is signalled for a refill. With examples is described how make-to-order production is done easily. The method is a type of reorder point system, but with time instead of quantity as decision variable. Unlike a traditional reorder point system increases and decreases of production can be planned. Future work load in various production sections can be estimated. It is described how an “Available-to-promise”-system should be designed and used. Cover-Time Planning (CTP) is a complete alternative to Materials Requirements Planning (MRP). CTP responds faster than MRP, since MRP for each structural level uses a batch size to "break down" and to magnify the need at the underlying level. In the end a large material acquisition needs to be ordered maybe just to build a single end item. CTP uses the end item requirement rates on all structural levels shifted with the lead times, when customer demand changes, the entire production chain react simultaneously. Keywords: Multi-level inventory and production control, Cover-Time Planning, Takt Planning, Materials Requirements Planning, Make-to order, Modules
1. Introduction and background Companies want to deliver to their customers with short delivery times, it is a competitive tool. Companies do not want to have huge inventories and capital tied up in unsold material that suppliers have already been paid for. In order to achieve a short and competitive delivery time purchase, production and preparation must have started before the final customer order arrives; sometimes long before. - What should we start, and after what decision rules should we start? It is a question that a lot of companies struggle with. What can we store? What production will start before the final customer orders come in? Just those items/products that we can appreciate a demand for; a demand rate at which it is desired to satisfy customer demand. If we do not know exactly the pace customers will ask for we must decide the rate we want to accomplish if customers would like to have the product. A way to keep down the amount of capital tied up in inventory and work-in-process (WIP) as well as to establish short delivery times is to create modular products. When the customer order arrives, the sales and end item is assembled from modules. The same module will be used in several end items; thereby the modules help to create a variety of end items. Modules are kept in store and a large stock of different end items is avoided. Many manufacturing companies have this option or opportunity but they must also have an administrative tool and control system that makes this possible and efficient. For some companies, if there is no final product finished, there is nothing to sell. The market demands that the product is available the instant the customer wants the product; sausages, cultured spices, chainsaws, some type of electric motors etc. – This article presents a technique to start production before the final
customer order arrives for both types of products and companies. This is important because it is never financial goals that create profitability, it is how effectively the company can manage its processes and satisfy current and changing demand. To give a background to and argument for this technique I start with a historical look back over my previous activities and experiences: 1.1 Ericsson Information Systems (1984-86) A problem that we had when I was working as a manager of Production Control at Ericsson Information Systems’ plant in Linköping 1984-86 was low turnover of inventories of components and semi-finished products. We said at that time that this was due to the very long delivery times from suppliers and that we had no reporting of each operation on production orders; but there were other reasons as well. The MRP execution results "conned" us to start unnecessary purchase orders, and production orders of the semi-finished and finished products. We started not only proposals for the current week but proposals for several weeks’ ahead (made planned orders to scheduled receipts). Thus, it was not just suggested planned orders it became real purchase orders. With these purchase orders the production control department could “chase” the purchasers; “where do you have this material?” and by that avoid future material shortages. The MRP run rescheduling meant that the earlier-and postponement list (rescheduling- in and reschedulingout) was huge, the only thing that there was time to correct was some rescheduling in. The result was much work-in-process (WIP) and long lead times. At Ericsson I came in contact with the consulting company Mysigma and their ideas. At that time Mysigma analysed their customers' inventory; how many days or months of demand each items’ inventory corresponded to. The items were also sorted in order of this cover-time and/or run-out time: a simple but brilliant idea. Mysigma’s founder Lars O. Södahl (Södahl, 1984) introduced the Toyota Production System from Shingo (1981) in Scandinavia. 1.2 Benzlers (1986-91) In 1986 I moved to the company Benzlers where I continued working with production, planning and production control. Benzlers was at that time a well-run, well invested company with wire guided automatic forklifts, with a modern computer system reporting every operation in the production orders etc. Benzlers produced mechanical gears and its components; gearwheels and pinions with many operations in long production chains with turning, external hardening, grinding etc. (At Benzlers, and Ericsson, I learned a lot about operations that was not covered as a part of my education at Linköping Institute of Technology (1969-1973, MSc Industrial Engineering); which is still considered to be one of the top educations in Sweden. E. g., Material Requirements Planning (MRP) was not taught at that time in Linköping. Even though MRP is available and used in most computer systems for material and production control most textbooks used at universities and high schools present a superficial introduction to MRP; although Jacobs et al. (2011) is an exception.) Now in retrospect, an error that we also made at Benzlers was that too much WIP was released in production. But we felt compelled to do as the MRP runs told us to do; how could we do otherwise? An exception was castings, bulky and expensive materials, an awake buyer did not want to fill our warehouse; planned orders were ignored and he bought new castings
for the gear housings according to current inventories and our Master Production Schedule (for end items). This was possible because the end items had a limited number of gear houses that could be overviewed manually. When we registered and tried to dispatch an assembly order of a gear, we discovered that it was not possible to release the order without a material shortage due to one pinion being missing at the same time as there were two order quantities or batches, of a different pinion or gearwheel! The Bill of Material (BOM) was correct; the MRP calculation knew and expected that we needed two pinions and two gearwheels, of various dimensions, to put together the complete gear. But the production flow had become unbalanced, an assembly was not possible!? One reason for this was that the routings of the items were very long; they contained many operations and also external heat treatments. Therefore and because of much WIP, individual production orders of the same item had very long lead times (throughput times). When different orders of the same item appeared together in the gear hobbing operation, the operators saved setup time by manufacturing both batches at the same time. Thereby it delayed other orders and items. We noticed this habit and made the disadvantage of it clear; but still very long and varying lead times continued. (Merger of production orders does not exist today? Yes, the same things still happens in lean oriented large well-known successful Swedish companies (Source: Master Theses)).
The problem with the lack of an article and two available order quantities of another article made me consider the advantages and disadvantages of MRP’s. I did a simulation model, with relatively simple product structures (BOMs), with variations in the lead times, and with net requirements planning exercises every week for a long simulation time with lead times uniformly distributed. As in most real cases, the MRP-calculation’s proposals on advancing and/or postponement of already registered manufacturing orders (rescheduling of scheduled receipt) could not be addressed. The results showed that there was always a "delay" for the Master Production Schedule (MPS), production quantities could not be delivered in time, a significant part only in past time. From time to time material was missing and it was not possible to complete the final end item in time because no safety stocks were used. The lateness was the “safety stock” necessary to accomplish the production rate. The accomplished production rate corresponded to the production rate(s) of the MPS; this delay was a consolation for that we were always accused of being not competent due to the delay in our MPS at Ericsson and Benzlers. At the same time I tested another idea than MRP, Cover-Time Planning (CTP); description of this method or technique is the purpose of this article. The study showed that with variations in demand and lead time, a simpler model than MRP is able to provide the same and better results in terms of lateness, inventories, WIP and its variety and at the same time achieve the same production rate. This became my Licentiate thesis (Segerstedt, 1991). Today if I am studying my licentiate thesis again I see opportunities for improvement; computer capacity has increased, I was not well informed about "Little's formula" and I did not mention in detail a major problem with MRP, make-to-order; how to deal with customer order production. Cover-Time Planning has further been described internationally (Segerstedt, 1995 and 2006), but mostly only in Sweden under the name “täcktidsplanering” in textbooks and popular science journals (e.g. Segerstedt 2005, 2009 and 2016). The idea has been further developed so the description presented here differs and goes deeper compared to early presentations.
1.3 Last 20 years After my Licentiate degree I started as a teacher at Mälardalen University, both in business administration and material and production control. Over the last 20 years I have supervised and taken part of a considerable amount of master theses in material and production control/logistics. This has given me insight and information from a large number of companies, some I have also visited. It is not a revolutionary development I have experienced. Real time computer solutions have been introduced in software for material and production control but the steering principles are basically the same as in the mid-1970s. Kanban, Theory of Constrains (The Goal from Goldratt and Cox (1984)) was very interesting readings but I have seen few successful implementations. Trends have swept through industry since my entrance in the mid-1970s; old things have got new names to be emphasized by consultants. In 1980s previous possibilities to reduce taxes by inventory write-downs disappeared (in Sweden) so a great campaign started for capital rationalization and to decrease inventories. 15 years after I heard/read the word "lean" it became a trend in Swedish industry and is often presented as a type of new science. 5s, eight wastage, etc., I have read about in numerous student works. But what's really new science since Henry Ford's Assembly line and Taylor (1911)? Many peer reviewed articles have been published but few of them have changed the behaviour and routines in industry. Orlicky (1975) and MRP on the contrary has influenced behaviour and routines concerning operations all over the world -, supply chain- and logistics management to an extent that has not really been highlighted (exception Jacobs et al. (2011)). Jonsson and Mattsson (2008, 2014) point nor at any revolutionary development; they show that the update of parameters of material and production control is very rarely but they see an improvement in comparison with previous studies. 1.4 Make-to-Order; Customer orders and production When a company has registered a sales order (customer order) they want it to take part in the MRP run so that no material procurement or production is forgotten. Then the Master Production Schedule (MPS), a forecast of future deliveries, must be reduced so that in the next MRP calculation material for the new customer order and the old MPS are not both generated. This would soon fill up WIP and inventories. If you have a limited number of end items, you can do this manually, as we did at Benzlers, by reducing the forecast in the interval of time in which the customer order is to be delivered and the amount the customer order contains (see: Segerstedt, 2002 (case Volvo CE); Segerstedt, 2009). If the number of end items are larger the MRP user requires a Master Production Schedule Planning (MPSP)system that consumes the forecast (Jacobs et al. (2011) and also helps to establish a new MPS for the next MRP calculation (Segerstedt, 2002 (case ABB Motors); Segerstedt, 2009). With Cover-Time Planning/Takt Planning the customer orders interfere in the calculations in a different way than with MRP which will be shown in the forthcoming text. Then the customer orders do not have to consume the plan; the forecasted demand and production rate is consumed by time. If the amount of customer orders does not come up to the expected demand rate, the replenishment of the item decreases immediately; the inventory and the WIP is enough for the future demand rate. Thus WIP is immediately somewhat faster than MRP.
2. Total lead time or cumulated lead time To describe CTP/Takt Planning we start with a simple Bill of Material (BOM) or product structure, see figure 1. E101
S207
S202
S301
B401
B305 B400
B400 Figure 1. Schematic product structure (BOM)
The end item article E101 is assembled from two semi-manufactured items (S202, S207) and one raw material (B401, a buy article); reality is generally much more complicated with many more articles, but this serves as an example. The number of structural levels is 4. The total or cumulated lead time for E101 starts with article B400, from this raw material article S301 is manufactured and from there S202. If we shall increase production of E101, the ordering of B400 must increase at once. Then we have to wait until the increased amount arrives, when it does we can increase the production of S301. Then it takes some time (lead time of S301) until we can start an increased production of S202. When we have finished the increased production of S202 (the lead time of S202) we can start the increased production of E101. Then, when this production is finished, at last we have increased the production of E101! Using MRP requires that a lead time is estimated and registered for every item (in Figure 1). To suggest an increase of production for E101 in close time means that MRP will correctly suggest the start of planned orders in past time! The sum of the lead times for items E101, S202, S301 and B400 must be kept track of and within this timeframe the production rate of E101 should not be increased; this to avoid proposed purchase of B400 in past time. This time interval is the total or cumulated lead time for E101. But if for e.g. B305 had a very long lead time then the longest “way” through E101’s BOM could be the sum of E101, S202 and B305; the cumulated lead time for E101. Using Cover-Time Planning or Takt Planning demands, like MRP, an estimated lead time for all items involved in the materials requirement planning. Disregarding the total lead time and increasing the production rate in near time will become a failure, with severe delivery delays despite the technique used. Decreasing production rates in near time will not create the desired consequences either; already ordered raw material and semi-manufactured items will continue to flow in. The used planning horizon must exceed the longest cumulated lead time of all end items independent of the technique for multi-level inventory control that is used.
3. Cover-time and Cover-Time Planning (Takt Planning) For how long will steps already taken cover expected demand for the item in question? The answer to this question we call cover-time. A more precise definition of cover time will be presented below, to make it distinct from run-out time. If the cover time of the article is shorter than the lead time there is certainly a high risk of a future shortage of the article: a risk that must be avoided. On the contrary, if the cover-time is very long it presents a signal that the inventory may need to be scrapped. Cover-Time Planning is a technique that converts expected demand rates for end items into decided production rates and explodes this demand rates’ gross to all components in the Bill of Material. Cover-Time Planning is a type of reorder system based on time instead of quantity and with a forward visibility (like MRP). A “master production schedule” in a “takt”, units per production day, is reported for all “end items”. In such a way that a production rate is supposed to be valid until another production rate is reported. Thus it is possible to plan ups and downs of production. From these “master production schedules” desired production rates are calculated for all articles considering lead times and connections in the BOMs (product structures). 3.1 Example: Cover-Time Planning The following product structures are prevailing between the articles S202, S207, S301 and B400. (The BOMs for S202 and S207 put together):
S202
S207
S301
B400 Figure 2. Product structures in our example
S202 and S207 in Figure 2 are now end items. (An explanation of why this is will be presented in section 4 Make-to-Order.) Raw material B400 is used for both S207 and S301 (thereby it is also a part of S202). To produce one unit of S202 requires two units of S301, and to produce one unit of S301 requires one unit of B400. To produce one unit of S207 requires two units of B400. Table 1 presents further information: Table 1. Used order quantity and lead time Product/ article
Lead time in time periods
Order quantity
S202 S207 S301 B400
4 5 10 15
15 10 40 50
Articles S202 and S207 have desired and decided production rates (shown below in Table 2 and 3). Considering already existing inventories and scheduled issues as well as scheduled receipts: when must the production of S202, S207 and S301 and the purchase of B400 start in order to fulfil planned production rates for S202 and S207? This will be presented in the forthcoming text.
Table 2. Situation S202 period 1
Article S202 Demand rate Sch. issues Sch. receipts Inventory "Supply" Pl. orders
14 2
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
5
5
5
5
5
5
5
5
5
5
6
6
6
6
6
6
6
6
6
6
5
4
7
1
5
6
11
7
0
2 15 13
12 31
7
10
3
1
1
-9
-9
-9
-9
-12
-12
…
(11)
(6)
(1)
(11)
(5)
(14)
(11)
(11)
(11)
15
15
15
15
A-T-P
2
1
-2
3
4
0
-1
5
-5
5
6
6
3
6
…
Cum. A-T-P
2
3
1
4
8
8
7
12
7
12
18
24
27
33
…
15
15
We are in the absolute beginning of time period (day) 1.”Supply” is already taken actions to cover future demand. “Supply” is equal to: current available inventory plus all scheduled receipts (cumulated) minus scheduled issues in past time. Scheduled issues in past time should have been delivered and they are not future demand, they are past demand, therefore available inventory must be reduced. Due to that the customer does not want a partial delivery, and/or another article which should also be send to the customer is currently in shortage, it is possible that we have a positive available inventory and still scheduled issues in past time. At the moment “Supply” is 2+14 + 15 = 31 (which is shown in Table 2). In table 2 it can be seen that “Supply” is placed in period (day) 5; because the lead time of a new replenishment of article S202, a scheduled receipt, that starts now can first be delivered in period (day) 5.
Supply will cover the demand rate for H202 during 31/5= 6.2 time units. Cover time (=6.2) is larger than the lead time plus inspection time (4+1= 5), therefore no replenishment is necessary at the moment. The expected demand that we should prepare for from now, including period 5, is 25 (= 5· 5), and actions already taken add up to 31. So at the moment no more replenishment is necessary. The first thing that happens in period 1 is that a decision is made as to if replenishment is needed or not. Using Cover-Time Planning a signal for replenishment is created when: (1) Cover-time < lead time including inspection interval + possible buffer time (2) Planned inventory within lead time plus inspection time < 0 Number (1) is the main principle. Number (2) is used when there is no forecasted demand rate but scheduled issues (customer orders), and also to complement number (1) and alert for possible mistakes that have been made. What deliveries, new scheduled issues, of S202 can be promised without causing shortages or late deliveries? The principle is: from now and in the future cumulated scheduled issues (customer orders) may not exceed cumulated forecasted demand rates; otherwise customer
delivery that is not planned for, or has not been forecasted, is promised. Then this material is not on its way downstream, shortages and late deliveries will occur. Additionally, a new replenishment cannot happen within the lead time; i.e. no new scheduled issues that make planned inventory in period 5 negative can be promised. Available to promise (A-T-P) in first period (day) is: Current available inventory (on hand) – Scheduled issues in past time + Demand rate in period 1 – Scheduled issue in period 1 (S202: 2 -0 +5 -5 =2). Scheduled receipts are not included in the calculation of available to promise (A-T-P); but cumulated scheduled receipts must exceed the cumulated demand rate. Cumulated scheduled issues (customer orders) may not exceed cumulated demand rates. Scheduled issues in past time should have been delivered and they are not future demand, therefore they must reduce current available inventory. A-T-P in other periods is: Demand rate in that period – Scheduled issue in the same period. A-T-P can be negative and must be allowed to be negative because a delivery larger than original forecasted demand during that period (day) must be possible. Cumulated available to promise (Cum A-T-P) cannot be allowed to be negative, it must always be nonnegative. If it is not, a scheduled issue that has not been forecasted before is promised but these units are not on their way downstream in the supply chain. No delivery of S202 can be promised in Period 1, 2 and 3. If it were the inventory within its lead time would become negative (according to Table 2). In period 5, 12 units seem to be available for deliveries. A promised delivery of 12 units in period 5 immediately changes planned inventory, A-T-P and Cum A-T-P (see Table 3).
Table 3. Situation S202 period 1 (12 extra units period 5) 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Demand rate
5
5
5
5
5
5
5
5
5
5
6
6
6
6
6
6
6
6
6
6
Sch. issues Sch. receipts Inventory "Supply"
5
4
7
13
5
6
11
7
0
2 15 13
0 31
-5
-11
-11
-21
-21
-21
-21
-24
-24
…
A-T-P
2
1
-2
3
-7
0
-1
5
-5
5
6
6
3
6
…
Cum. A-T-P
2
3
1
4
-3
-3
-4
1
-4
1
7
13
16
22
…
-
14 2
10
3
Therefore, promising 12 units in period 5 is nothing to recommend. The scheduled issues in period 6 and 7 will probably be delayed. There is still time to register and start a scheduled receipt (replenishment) for period 6, but downstream components are not prepared for this increase and a possible shortage of components will delay the replenishment. Many of the forecasted demand rates are consumed, so a new delivery larger than one unit cannot be promised until period 11 (see table 3). A delivery of 12 units should therefore not be promised until period 10, according to Cum A-T-P in table 2. In the case where there are never any delays, never any mistakes in reservations of orders, no scrap rejects or other disturbances; then ”Available-to-Promise” could be described rather simply. What is described here is logic and a technique that must deal with stochastic events, even correcting previously made mistakes; which makes Cum A-T-P the important decision variable.
Information about starts and deliveries of S202 must be created in order to be able to estimate future work load. Future planned orders for replenishment can be calculated from ”supply” and already decided scheduled receipts and the existing demand rate. The next replenishment of S202 must come in period 7. There will therefore be a fictive planned inventory of 31+15(5· 7) = 11 units at the end of period 7. (Total demand from period 1 to 7: 5· 7= 35; total “supply” in period 7: 31 +15 = 46; 46 -35 = 11) This means that the next incoming delivery of 15 units must take place in period 10 (see table 2). At the end of period 10, fictive planned inventory is 11; hence the next replenishment must come in period 12. The end of period 12 inventory is 14 units etc. Planned orders are only calculated when an analysis of work load is necessary and done. The lead time of the article determines the start of the planned order, and when different operations are performed, distributed between the start and completion time (like the usual procedure when using MRP). Table 4. Situation S207 period 1
Article S207 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
3
2
3
3
7
2
-1
4 10 5
2
2
2 14
0
0
-7
-7
…
A-T-P
4
-1
-2
-1
2
3
-2
3
-4
3
3
3
3
…
Cum. A-T-P
4
3
1
0
2
5
3
6
2
5
8
11
14
…
Demand rate Sch. issues Sch. receipts Inventory "Supply"
5
7
”Supply” is 7-3+10= 14 units. A scheduled issue in past time should have been delivered; therefore it is not included in ”supply”. “Supply” is the current amount already decided we have for satisfying future demand. Cover-time for S207: 5+ (14- 5·2)/3 = 5+ 4/3 = 6.33; expected demand we should prepare for until period 6 is 13 units. ”Supply” is 14; thus replenishment can wait until next period. But, with a buffer time, necessary as a type of safety stock, replenishment for period 6 should be performed now. A scheduled issue larger than 5 cannot be promised until period 10. If it is possible an order should perhaps also be made in part deliveries to enable deliveries to other customers. Article S301 The demand rate for S301 depends on S202’s demand rate. S202’s demand rate increases in period 11 and therefore the demand rate for S301 increases four periods earlier (due to the lead time of S202); in period 7. Since two units of S301 require producing one unit of S202 the demand rate is twice as large as for S202.
Table 5. Situation S301 period 1 Demand rate Sch. issues Sch. receipts Inventory "Supply" Pl. orders
30 38 20
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
10
10
10
10
10
10
12
12
12
12
12
12
12
12
12
12
12
12
12
12
28
28
28
40 68
68
69
68
108
108
108
108
…
108
(16)
(2)
(30)
(18)
(6)
(34)
(22)
(10)
(38)
40
40
40
40
40
”Supply” is 20-30+38+40+40= 108 units. The scheduled issue in past time is the manufacturing order of S202 (15 units in period 4); not yet started with a withdrawal of S301. Cover-time for S301: 6+ (108- 6 10)/12 = 6+ 48/12 = 10 < 10 +1; expected demand we should prepare for from now until period 11 is 120 units; ”Supply” is 108. A new scheduled receipts for delivery in period 11 must by registered and started at once. Article S301 is manufactured in the own production facility where sometimes there is a need to analyse future workload. At the moment there are 3 scheduled receipts, next replenishment originating from those, ”supply” and the expected demand rate is required in period 11, then in period 14 and 17. Article B400 The demand rate for B400 depends on the demand rates of both S207 and S301. S207’s demand rate increases in periods 6 and 16. Therefore the demand rate of B400 increases in periods 1 and 11. The demand rate of S301 increases in period 7 and therefore the demand rate of B400 increases in period (7 -10 = -3), i. e. it has already happened. The demand rate in period 1 is the sum of the demand rate for S207 multiplied with 2 in period 6 and the demand rate for S301 in period 11. Table 6. Situation K400 period 1 Demand rate Sch. issue Sch. receipts Inventory "Supply"
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
18
18
18
18
18
18
18
18
18
18
20
20
20
20
20
20
20
20
20
20
50
50
50
50
95 295
Cover-time for K400 is 10+ (295- 10 13)/20 = 10+ 115/20 = 15.75; 15.75 15+1, a new purchase order must be ordered at once. 4. Make-to-Order, Production of customer orders Figure 1 consciously shows E101 as an end item the company delivers to its external customer. E101 is design and assembled from different modules. Like a Wheel Loader at Construction Equipment or an electric motor at ABB Motors. E101 is unique and we do not know in advance when and how the customer wants to compound and equip the product. The article E101 does not actually exist until we have got the order from the customer. As we do not know when and how, we have no estimated demand rate for E101. But for the modules S202, S207 and several other items we have estimated demand rates and materials are on their way downstream in the supply chain to allow a short delivery time for E101 Suppose the company gets an order of 5 units of a product equipped like E101, if this order is attempted to be registered and started earlier than period 10 then the system will/should say it is not possible! Because first in period 10 Cum A-T-P for S207 is equal or larger than 5 units. A lead time for E101 may be estimated to be two time periods. That means that a delivery date of E101 before time period 12 should be signalled as not possible. This will avoid a future late delivery. Segerstedt (2002) shows how both ABB Motors (ABBM) and Volvo Construction Equipment (VCE) worked with fictive end items. Why, because they worked with MRP to accomplish short delivery times they had to start production before the customer orders arrive. They had
so many end items and variants of end items it was not possible to make a master production schedule for all end items. ABBM had only a few fictive items (7) on the highest structural level (0); those were expanded one structural level with fixed percentage distributions. This structural level (1) still consisted of only fictive items. Structural level 2 was also created from level 1 with fixed percentage distribution; structural level 2 consisted of real sellable end items (equipped electric motors), approximately 2000 different products. To promise delivery dates ABBM used a Master Production Schedule System (”available-topromise”-system) in their Baan Triton installation (their ERP (Enterprise Resource Planning)system). The system helped ABBM to construct the next MPS and to set delivery dates that did not override current and previous MPS. In their MRP calculation VCE used two fictive articles (L50 and L70) and several ”options” containing possible modules in a wheel loader; excavator, gear box, motor, sits, air conditioners. The options, with article numbers that all began with ”9-”, were complex products with their own Bill of Material. Those “9-numbers” contained a credit quantity for the article in the fictive end items (L50 or L70) they were substituting. That was necessary to avoid material requirement of both the 9-article and the one in the fictive product structure. 9articles were handled manually in the same way as described earlier in relation to Benzlers. Every 9-article had a MPS, and the MPS’s were decreased if a customer order arrived equipped with this 9-article If ABBM and VCE should use Cover-Time Planning or Takt planning as it is described here; what should they do? ABBM were already working with takt. Every month their sales and operations planning ended up with a decision how many units that should be produced of seven different aggregate (fictive) electric motors per day. 2-300 article numbers are standard products and must be delivered immediately from (two) different sales inventories. A number of aggregate fictive products should be used to explode to real articles in sales inventory like before. Because 2-300 articles is difficult to handle with individual manually set forecasted demand rates. The electric motors are built from modules; modules consumed of customer orders, not stored in sales inventory require a separate expected demand rate. But forecasting modules instead of finished motors would decrease the number of fictive article a lot. VCE does not need any fictive articles; every “9-number” gets an expected demand rate, also the “option” in the previous fictive articles. When a customer order is registered, no quantity has to be decreased from any plan. This will radically decrease the (manual) administration. 5. End comments 5.1 Differences Some find it difficult to see the difference between MRP and Cover-Time Planning (CTP). The main differences are: demand rate instead of a special quantity in a special time; the demand rate ”rules” from now and into the future, the past is ”lost”; no automatic rescheduling of scheduled receipt and their connected schedule issue (only warnings: this is late!); a new customer order does not have to decrease the expected demand rate (=MPS)! The example shows that the customer orders contribute to the calculations in a totally different way than with MRP. The new customer order does not have to “turn off” the old plan (or MPS); instead if no new customer orders arrive CTP quickly “turns off” new
replenishments. (CTP reacts more quickly than MRP because MRP has an order quantity (size) for every structural level that ”presses down” and increase the demand on the lower structural level. In the end there can be a large purchase requirement just to build one single end item. 5.2 Buffering, variants and/or takt We talked about buffering, variants and/or takt in the 1980s when I worked at Ericsson. With a variant buffer we meant an inventory of semi-manufactured items that made it possible to quickly deliver just that specific variant of the product, e.g. a special printer. The total semimanufactured inventory could be allowed to be larger than the total sales of printers. Thus also a type of takt buffering was accomplished; it could be possible to deliver more than the forecast of the product. But we could not figure out how we should solve this with our MRPsystem. With CTP the problem is easily solved; the demand rate for the modules is set a bit larger than expected average demand for the end product. Tied up capital will increase a bit but so does delivery performance. 5.3 Material control and Production Control CTP is not only a system for material control; it also presents information for production control. The estimated demand and production rates used must be possible to achieve both in near time and in future time (due to capacity). Therefore it is important that CTP can create information about workload in different machines and production facilities. The shorter cover time an article has the more important is it that the manufacturing or purchase order of the item makes ready and finishes. Cover-time becomes a complement priority magnitude in which sequence orders should be manufactured in a work station. CTP has a CONWIP attribute: inventory and work-in-process of item i is constrained between a multiple of the used order quantity { (n 1) Qi WIPi I i n Qi ( n 1, 2 , )}. (CONWIP is presented e.g. Spearman et al (1990), Hopp and Spearman (2008), Hopp (2008)) Hopp (2008) define push and pull “A pull system is one in which work is released based on the status of the system and thereby places inherent limit on WIP. A push system is one in which work is released without consideration of system status and hence does not inherently limit WIP.” There are many previous attempts to define push and pull that exist, cf. Pyke and Cohen (1990) and Bonney et al. (1999). Bonney et al. (1999) point out that the definitions of push and pull are inconsistent between different researchers and arguments about performance are peculiar, if the performance of a pull system is poor then it may be suggested that this is because the fundamentals of just-in-time are not being observed. Whereas, if the performance of a push system is poor it is because it is a push system. It seems like Hopp’s definition of limited work-in-process is important for the success of a system and not push or pull. Nicolin (1959) explained that too much WIP would lead to future delays of customer orders. ‘‘Balance flow and not capacity’’ is an OPT-strategy (cf. Goldratt and Cox (1984); Goldratt and Fox (1986)) that implicates control of WIP. Silver et al. (1998) and Hopp and Spearman (2008) show that after WIP has reached a certain level, more WIP does not increase the throughput rate. It only prolongs the lead-time. In addition, Pettersen and Segerstedt (2009) show that much WIP not only prolongs the lead-time or throughput time it also increases its variation. Therefore the clear CONWIP attribute of CTP is important and an advantage compared to MRP.
5.4 Future for Cover-Time Planning!? What is described here is a computer system. When I, many years ago, in the early 1990s, was in contact with and was contacted by, software developers for production and material control, I felt doubt and misbelief. They had invested in MRP and CTP was unproven. But now Jonsson and Mattson (2008, 2014) state that many companies use “run-out time planning” (Jonsson and Mattsson (2008) term run-out time planning is synonymous to cover time planning (täcktidsplanering)) In the form that it is presented here we cannot be sure that it is in many installations; but the idea has spread in an extent that I did not know before (Jonsson and Mattson, 2014). In a given computer system the calculations do not take place in tables like they have been presented here but with entirely different solutions; linked lists etc. A demand rate is in order as long as there is no other. The absolute majority of functions in an ordinary ERP-system can still be used without changes; order entry, routings, in and out of stock, BOM etc. So it is rather easy and requires little work to change to CTP. Make-to-Order creates no problem; inventory and WIP for every item are restricted. It creates information for calculating work load; it is possible to plan future increases or decreases of production! CTP is a type of reorder system, with time instead of quantity as a decision variable. Therefore it should also be compared with other reorder systems. (Companies change from MRP to reorder systems due to dissatisfaction.) CTP avoids reorder system’s disadvantages, forecasts for every item and no future visibility. The author is thankful for valuable comments from reviewers. Referenser Bonney M. C., Zhang Z., Head M. A., Tien C. C., Barson R. J. (1999). Are push and pull systems really so different?. International Journal of Production Economics, 59 (1), 53–64. Goldratt E. M., Cox J. (1984). The Goal - Excellence in Manufacturing. New York: Creative Output/North River Press. Hopp W. J., Spearman M. L. (2008). Factory Physics (3ed.). Boston : McGraw-Hill Publishing. Hopp W. J. (2008). Supply Chain Science. New York : McGraw-Hill Irwin. Jacobs F. R., Berry W. L., Whybark D. C., Vollman T. E. (2011). Manufacturing Planning & Control. New York : McGraw-Hill Irwin. Jonsson P., Mattsson S.-A. (2008). Inventory practices and their impact on perceived planning performance. International Journal of Production Research, 46 (7), 1787-1812. Jonsson P., Mattsson S-A. (2014). ’Best practice’ vid lagerstyrning i svensk industri (in Swedish: Best practise inventory control in Swedish industry), Forsk-ningsrapport. Chalmers Tekniska Högskola och Plan. (http://publications.lib.chalmers.se/records/fulltext/192457/local_192457.pdf) Nicolin C. (1959). Planering av en verkstad med blandad tillverkning - matematiska tumregler till hjälp för att organisera effektiv verkstadsplanering (in Swedish: Planning a Workshop – Mathematical Rules of Thumb for Efficiency; Internal Working Paper). Finspong: Arbetsrapport PM VD4a/59. deLaval Jungström. Orlicky J. (1975). Material Requirements Planning. New York: McGraw-Hill. Pettersen J.-A., Segerstedt A., 2009. Restricted work-in-process: A study of differences between Kanban and CONWIP. International Journal of Production Economics, 118 (1), 199–207 Pike D. F., Cohen M. A., 1990. Push and Pull in Manufacturing and Distribution Systems. Journal of Operations Management, 9 (1), 24-43. Segerstedt A. (1991). Cover-Time Planning - an Alternative to MRP, Linköping: PROFIL 10 (ISBN 91-970074-5-5).
Segerstedt A. (1995). Cover-Time Planning, A method for Calculation of Material Requirements. International Journal of Production Economics, 41, 355-368, 1995 Segerstedt A. (2002). Production and Inventory Control at ABB Motors and Volvo Wheel Loaders, Two Examples of MRP in Practical Use. Production Planning & Control, 13 (3), 317-325. (https://www.researchgate.net/publication/237218389_Production_and_inventory_control_at_ABB _Motors_and_Volvo_Wheel_Loaders_two_examples_of_MRP_in_practical_use) Segerstedt A. (2005). Täcktidsplanering underlättar kundorderstyrd produktion och kontroll av produkter i arbete (in Swedish: Cover Time Planning; facilitates make-to-order production and control of work in process). Bättre Produktivitet, No 3. Segerstedt A. (2006). Master Production Scheduling and A Comparison of Material Requirements Planning and Cover-Time Planning. International Journal of Production Research, 44 (18-19), 3585-3606. Segerstedt A. (2009). Logistik med fokus på material- och produktionsstyrning (in Swedish: Logistics with focus material- and production control). Malmö: Liber (ISBN 91-47-08900-0). Segerstedt A. (2016). Täcktidsplanering/Taktplanering (in Swedish: Cover-Time Planning/Takt Planning). Bättre Produktivitet, No 2. Shingo S. (1981). A Study of the Toyota Production System. Tokyo: Japan Management Association. Silver E. A., Pyke D., Peterson R. (1998). Inventory Management and Production Planning and Scheduling. New York: Wiley. Spearman M. L., Woodruff D. L., Hopp W. J. (1990). CONWIP: A pull alternative to Kanban. International Journal of Production Research, 28 (5), 879–894. Södahl L. O. (1984). Den nya japanska produktionsfilosofin (in Swedish: The new Japanese philosophy of production). Stockholm: MGruppens förlag, Stockholm Taylor F. W. (1911). The principles of Scientific Management. New York: Harper & Brothers