SSIE 598 Project Report “Resolving the MRP versus Lean Conflict” Student: Michael D. Ford Faculty Advisor: Dr. Nagendra Nagarur Department of Systems Science & Industrial Engineering Binghamton University Spring 2016
SSIE 598 Project Report
“Resolving the MRP versus Lean Conflict”
Student: Michael D. Ford
Faculty Advisor: Dr. Nagendra Nagarur
Department of Systems Science & Industrial Engineering
Binghamton University
Spring 2016
2
Copyright @ 2016 by Michael D. Ford
All Rights Reserved
3
Accepted in partial fulfillment of the requirements for the degree of
Master of Science in Industrial and Systems Engineering
In the Graduate School of
Binghamton University
State University of New York
May, 2016
Dr. Nagendra N. Nagarur,___________________________________, Project Guide, May, 2016
Department of Industrial and Systems Engineering, Binghamton University
Dr. Mohammad T. Khasawneh, _________________________,Committee Member, May, 2016
Department of Industrial and Systems Engineering, Binghamton University
4
Abstract: The “MRP versus Lean” debate has been fermenting for several decades, in fact
evolving into an “ERP versus Lean” debate. Much of this discussion has focused on the similar
“Push versus Pull” argument, but this is only part of the total picture. This involves much more
than an inventory replenishment approach, and yet may not be a total conflict. MRP systems are
a technical application, while lean is a management philosophy. It is possible to utilize both,
when and where appropriate, as part of an enterprise-wide planning and execution system.
A combination of resources have been utilized, to include academic papers, technical
journals, web blogs, operations management texts and the student’s own personal experience in
supply chain operations. This includes hands-on work in the implementation, use and
optimization of MRP/ERP software solutions, as well as extensive discussion with industry
experts who have served in the trenches since the dawn of MRP software in the 1960’s. The
evidence is clear that MRP may retain its role as a planning technique while lean ensures success
of execution. This paper will introduce techniques for improving the performance of MRP/ERP
systems, in essence allowing them to function leaner. This will include an algorithm developed
by the student for optimizing MRP processing and utilizing an MS Excel spreadsheet as the basis
of simulating MRP processing and outputs.
5
Table of Contents
I. Introduction p. 6
II. Literature Review p. 9
III. Methodology p. 17
IV. Results p. 27
V. Conclusions p. 29
VI. Appendices p. 30
VII. References p. 35
6
I. Introduction
The central debate of why material requirements planning (MRP) systems inhibit the
implementation of lean initiatives often centers on criticisms of MRP software. It is useful to
briefly explain what MRP is and what it does for manufacturing entities. MRP is a tool to
calculate the order quantities and timing for needs of dependent demand items, that is, those
represented in a parent item’s bill-of-material (BOM). MRP at the most basic level performs
three critical operations, as follows (Narasimhan, p. 362):
Guides the parent-to-component explosion process by multiplying parent planned order
quantities times the “quantity per” of each component as defined in the BOM
Nets the gross requirements against any existing inventory balances and/or scheduled
receipts in the respective time bucket
Offsets the planned order receipts by the standard lead time to determine the planned
order release.
This process is typically illustrated in textbooks and journals through the use of a grid format,
which will be utilized in multiple examples throughout this paper. It may behoove the reader to
refer to the first two pages of the appendix to become familiar with the grid concept. An example
of such a grid is provided in the first page of the appendix (p. 28), and includes the sequential
calculations for filling in the grid. A more fully detailed explanation of the process is included on
the second page of the appendix (p. 29).
The grid examples demonstrate that basic MRP involves simple arithmetic: adding,
subtracting, multiplying and dividing. Hence, it is very appropriate to utilize MS Excel as a
simulation tool for MRP. This ties into the complaint that MRP requires lots of maintenance, is
data-intensive, and calculation heavy. To be sure, the student has witnessed MRP re-generation
7
(the routinely scheduled system update) that takes overnight to complete all the data processing
requirements.
MRP has evolved since the 1960’s to add more capabilities, such as “where-used”,
“pegging”, exception reporting, etc. Manufacturing resource planning (MRP II) software arrived
in the 1970’s and included the ability to link material requirements with other functions such as
inventory management, forecasting, procurement, master planning and capacity planning,
ensuring a closed-loop process (Narasimhan, p. 351). Enterprise resource planning (ERP)
systems followed in the 1980’s, incorporating all business functions (including human resources,
finance, quality, logistics, distribution, marketing, engineering, etc.) into a single system utilizing
relational databases such that everyone had access to the same data in real time without the need
for scheduled uploads and downloads.
Lean has similarly evolved over the years, from its origins in the just-in-time (JIT)
methodology. JIT originally meant having the right quantity of the right item at the right point in
time, not early nor late. This evolved to the elimination of waste, as it is wasteful activities that
cause long lead times, hence lateness. Whereas the historical viewpoint was that early was good,
it became universally accepted that too early was just as bad as too late. Too early implies wasted
resources to provide materials that are not yet needed, and likewise paying for the space,
opportunity cost and risk associated with inventory holdings.
Lean aspires to remove waste by streamlining activities, for example, minimizing
movements, handling, held inventory, or any activity that does not contribute to adding value to
the product. Value is defined as something the customer is willing to pay for. For example,
consumers are willing to pay more for a restaurant meal than what the food would cost if
purchased at a grocery store because the cooks and wait staff are providing value-added services.
8
On the other hand, stored inventory does nothing to enhance the customer’s perception of value,
in fact may reduce the value due to shelf life concerns and/or technical obsolescence.
Lean initiatives have sought to overcome the frequent complaint that traditional MRP-
based manufacturing is a game of “hurry up and wait.” Orders are expedited at various points,
only to run into bottlenecks, increasing the urgency to expedite further, resulting in arriving at
another bottleneck, etc. Imagine an analogy considering motor vehicles driving through heavy
traffic congestion at intersections regulated via traffic signals. Drivers hit the gas on green, then
slam on the brakes on red. Lean operates more like a roundabout, where drivers keep moving
continuously, perhaps not at the same speed as cars accelerating through a green light, but never
stopping and hence overall moving faster.
Lean manufacturing principles found their way into other areas, leading to lean office,
lean service, lean logistics, lean health care, etc., such that it is most commonly referred to now
as the generic lean operations. The reality is that lean may be applied wherever waste exits, and
“the creation, ordering, and provision of any good or service… can be made to flow” (Womack,
p. 51). This concept of totally eliminating all waste often leads so-called lean gurus to declare
that an organization is either lean or not lean, as if it’s a simple either/or proposition. A more
appropriate perspective would be to recognize whether an organization becomes successively
leaner, suggesting that lean is more of a journey as opposed to a destination. It is with this in
mind that the ultimate question is not “MRP versus lean” but rather if the two techniques may be
utilized in combination such that MRP helps companies become leaner.
9
II. Literature Review
A primary complaint of traditional MRP systems is their inherent dependency on accurate
data with respect to BOM’s and inventory records. It is true that “inventory records are
indispensable to the MRP, and the MRP is only as good as the inventory records” (Arnold, p.
107). A common suggestion is that MRP must have 95%+ accuracy of inventory records,
BOM’s and routings, but even this standard is insufficient. The chance of failure when
considering serial probabilities is equal to the product of those probabilities (assuming
independence). If each item within a BOM had a 95% chance of being in stock, then the odds of
having “n” items is represented by .95n. Any BOM with at least 14 components would have a
less than 50% chance of being completed, as .9514 = .487575.
MRP also presents challenges due to the independent, sequential nature of its processing,
from the master schedule, to the subassemblies, to the purchased parts, all without consideration
of shop floor realities. Many users go outside the system, using substitute processes or end-
arounds to do whatever possible to make it work (Harding, p. 8). Even formal processes to
reduce system nervousness such as lot sizing, input/output analysis, infinite capacity planning,
dispatching, and queue control are considered “Band-Aids” (Kanet, p. 58). This system
nervousness is defined as “significant changes in MRP plans, which occur even with only minor
changes in higher level MRP records” (Vollman, p. 407).
Another criticism of MRP considers the evolution to ERP as making things worse, because
ERP systems have infinitely more data requirements and are even more structured on planning,
versus reacting to demand. “Many companies are stuck with their current ERP system for the
time being due to large investments in the technology and lack of support of top management for
lean tools” (Bradford, p. 30). Companies that expected ERP to make life easier and efficient (i.e.
10
require less staff) found that they needed to hire more workers to manage the system inputs and
outputs, and conduct all the transactions and respective analysis. This was complicated by the
fact that software solution vendors historically viewed lean as something for a “bolt-on”
package, rather than designing lean functionality into their ERP systems (Bradford, p. 30). Dixon
notes that MRP/ERP implementations “too often… actually added cost while doing little to
improve inventory turns or customer service (p. 42). ERP may become a liability as “the
problems it introduces – or perpetuates from the days of MRP – include complex bills of
material, inefficient workflows, and unnecessary data collection” (Bartholomew, p. 2).
An interesting observation, and seeming contradiction, is the fact that lean was originally
most successfully implemented by high-tech industries. Karmarkar notes that Toyota and
Hewlett-Packard are examples of companies that utilized Kanban even though they also
spearheaded the use of advanced computer automation (p. 122).
Lean is not without its own criticisms. Sometimes the failure of the MRP-lean marriage is
primarily based on the latter. One analysis notes that the overwhelming majority of companies
that start to implement lean fail to recognize the anticipated benefits. It is recognized that lean
has not only advantages but risks as well, including potential supply chain disruptions,
significant costs to initiate improvements and employee resistance (Ciarniene, p. 730). Further,
“lean manufacturing can be a great tool for short-term, manufacturing-based planning, but it does
not have the depth to move outside the plant floor or go beyond a one-month period” (Gill, p.
20). There is also the reality that lean cannot replace MRP/ERP’s ability to perform functions
such as maintaining BOM’s, lot tracking for the food and beverage industry, or performing cost
roll-up (refer to page 5 of the appendix for a complete list, p. 32).
11
Push versus pull is a frequent battle that is part of the larger MRP versus lean war.
Traditional MRP has been characterized as the epitome of a push system, with centralized
decisions coming from the office, structuring schedules based on (often inaccurate) forecasts and
releasing materials to the shop floor without regard to downstream conditions. MRP starts with
“sales plans… created independently from constraints on production and material plans”
(Imaoka, p. 1). Ultimately, orders arrive at work centers that are either bottlenecked, shut down
or waiting for tooling, parts, the operator or some other constraint. This is why job shops, the
favored MRP environment, typically result in long manufacturing lead times that are 85% queue
time (i.e. waiting at the work center) as opposed to value added run time. This should not be the
case, as one of the primary tenets of MRP is to NOT release orders without verifying material,
tooling and capacity availability.It is possible to balance the limitations of push by combining it
with pull. One analysis suggests that batch size can be controlled as a means of balancing the
push-pull trade-off, utilizing colors to code the need for replenishment: green is full; yellow
requires an order; red must be expedited (Pitcher, p. 11).
Those who argue for extending pull throughout the supply chain would be advised to observe
the “Beer Game Simulation”, which demonstrates the futility of a pure pull system: the bull whip
effect. This occurs when “small fluctuations in orders at retailers ultimately become wide
fluctuations in orders to suppliers of manufacturers” (Chopra, p. 22). Pull systems work
especially well locally and where consistency exists, but become a challenge with increasing
distance and variability. For example, “pull systems are fine if your McDonald’s franchise is
downtown with a steady stream of customers. But if you’re next to a football stadium, how can a
pull system alone prepare you for the day of the game” (Karmarkar, p. 124). The key is to
balance and manage the push/pull boundary, although there may be debate as to where this
12
boundary exists. Fogarty maintains that distribution requirements planning (DRP) is a push (p.
311), while Ross insists that DRP is a pull system (p. 370)! Qualified operations specialists
would consider reorder point to be a pull system, yet it is classified as part of the MRP-push
system in one article (Abuhilal, p. 52). These examples suggest that the bitter MRP/lean divide is
not necessarily an objective debate.
Another element in the MRP-lean debate is the “Japan v. US-Europe” contrast. Lean is
commonly associated with the Toyota Production System (TPS), although lean “can be traced as
far back as Eli Whitney and Henry Ford” (Miller, p. 1). Ford’s assembly line is often referenced
as a true lean endeavor and in fact Toyota’s management came to the USA to observe Ford
Motor Company operations in the early 1950’s. Cynics have often derided TPS by suggesting
that the USA taught Japan lean and quality so they could teach it back to the USA. The reality is
that Japanese companies more quickly and readily embraced lean as USA and European
companies were focusing on MRP-based approaches, although there has been somewhat of a
reversal. Just as Western companies have been adopting lean, Japanese firms have been adopting
ERP/MRP systems as they “move from high volume standardized goods to customized products”
(Vollman, p. 391).
Some authors have directed their criticisms toward the organization itself, or management in
particular, somewhat alieving MRP of any blame for failed lean implementations. Narasimhan
declares “the major impediment to the revamping of American industry may be the sclerosis of
this management philosophy, rather than any technical barrier” (p. 551). He additionally notes
that “although the practices of lean production do not depend on the use of computers, they are
not inconsistent with CIM technologies and can even be facilitated by them” (p. 552). It is also
noted that management often mistakenly believes that MRP/ERP software or lean initiatives are
13
like a magic potion, a one-size-fits-all solution to whatever ails their company. The reality is that
“no panacea exists, if we blindly follow the philosophy without understanding the process, we
will make cost, delivery, and quality worse, not better” (Sandras, p. 15).
There are two major criteria which may affect the appropriateness of MRP or lean: (1)
planning level, and (2) manufacturing environment. The first point alludes to the fact that MRP
is, by its very definition, a planning technique. Lean is for execution and is “focused on using
simple, visual, shop floor level tools such as Kanban to control the flow of production” (Mclean,
p. 22). This contrast is illustrated in the “Top-down Planning Hierarchy Diagram” on the third
page of the appendix (p. 30). MRP represents the lowest, most detailed level of planning, but is
separated from the execution that follows as work orders are controlled via production activity
control (PAC). MRP may be used as a planning tool at the higher level with lean handling the
execution on the floor so they are complimentary, rather than conflicting.
Different manufacturing environments are described in the “Volume/Variety Matrix” found
on the fourth page of the appendix (p. 31). These include engineer-to-order (ETO), make-to-
order (MTO), assemble-to-order (ATO), make-to-order (MTO) and mass customization (MC).
MRP is most appropriate in the MTO position, the typical job shop environment that creates
unique work orders for each customer. This results in a wide variety of BOM’s, routings, lot
sizes, lead times, etc., as each order is different. Lean does not work as well in this position, but
rather “tends to work best in repetitive environments, where families of products are produced in
cells or production lines” (Steinbrunner, p. 28). The ATO strategy may combine elements of
MRP and lean by using repetitive flow or Kanbans to manage the standard components or
options, and MRP work orders to configure the final assembly. Modular BOM’s also facilitate
lean planning in the ATO environment. A technique known as POLCA (paired-cell overlapping
14
loops of cards with authorization) addresses the challenge of applying lean to the MTO
environment. It utilizes cards, like Kanban, to control the production of high variety products.
“By strictly limiting work-in-process (WIP) inventory between cells, POLCA aims to increase
the speed of job transfer and reduce the unbalances in the manufacturing system” (Powell, p.
397).
Demand-driven MRP (DDMRP) is a popular and relatively recent concept that makes MRP
responsive to demand (pull) as opposed to being driven by a forecast (push). This does not
eliminate all stocked inventory, but rather places strategic buffers, known as de-coupling points,
that should be monitored. “DDMRP uses a new type of strategically positioned and dynamically
managed stock position to dampen variability, compress lead times and reduce working capital
requirements while ensuring unprecedented levels of service” (Harding, p. 11). This approach
recognizes that from a practical standpoint inventory must be kept somewhere in the pipeline. It
recognizes that zero inventory is neither practical nor possible, nor should we keep inventory
everywhere, thus the optimal approach is to keep stock where it best minimizes supply risk and
would result in lower aggregate inventory levels (Ptak, p. 10).
One suggestion for opportunity is the utilization of information technology (IT) resources to
their full potential. The speed of today’s processors and storage capability of current drives are
well suited to performing an array of computational gymnastics on vast volumes of “big data.”
Sinocchi advocates integrating value stream mapping into the software to model the flows of
information and materials (p. 1). Another example is customizing software to fit the combined
needs of an MRP/ERP system and lean. One company reconfigured their software to support one
piece flow, with the general manager commenting “if we didn’t have an ERP system, our lean
initiatives wouldn’t work” (Bartholomew, p. 3). Dixon contrasts the differences between lean
15
and MRP/ERP (p. 42) while suggesting they may fuse their capabilities (p. 45). His suggestions
include: supporting VSM with data capture; utilizing CAD to promote focused layouts;
electronic posting of Kanban signals to facilitate pull; electronic work instructions to support
standard work and error proofing; and real-time measurements to validate takt time. One author
suggests that ERP systems are in fact lean, as they streamline the IT process by utilizing a single
software solution to run the business (Dilhe, p. 1). Reduced IT staffing requires that the system is
capable and fully integrated.
Several authors have adopted an optimistic viewpoint that MRP is a valid technique in its
own right, albeit with qualifying assumptions. Abuhilal found that information sharing provides
the best opportunity to maximize MRP effectiveness, utilizing a computer simulation to analyze
total supply chain cost versus holding cost, resulting in a minimum 22% cost savings (p. 55).
Another approach suggests combining simulation, heuristics and algorithms to gain an
understanding of variation and hence adjust parameters on material plans (Bradford, p. 34).
Bertolini describes a synchro-MRP system incorporating value stream mapping (VSM) as a
solution well suited to the high variety, low volume environments. These somewhat pro-MRP
viewpoints are consistent with the pro-lean (anti-MRP?) positions in that they suggest due
diligence must be given to optimize MRP systems to run effectively and efficiently.
Perhaps the best commentary comes from authors advocating a combination of lean six
sigma (LSS) and ERP. They argue that the functionality of software systems is well suited to
LSS. Sink believes industrial & system engineers “can and should play a large role in ERP
configurations because they can ensure that design is integrated into the configuration…
ensuring that a larger portion of the data elements required for sustainable, continuous
improvement are captured” (p. 37). Nauhria states that an “ERP solution provides capabilities
16
that are core to the success of lean six sigma initiatives: common work processes; real-time data
sharing; application integration; data integrity, inventory visibility; available to promise (ATP);
capable to promise (CTP); real-time order management; proactive exception management;
business intelligence; and analytics” (p. 39-40). This approach suggests that as MRP evolved to
MRPII and ultimately ERP, while JIT evolved to Lean and ultimately LSS, they are now not
only able to coexist, but have a mutual codependency for success.
17
III. Methodology
MS Excel serves as a simulation model to demonstrate how MRP can be more effective,
reducing inventory balances and thus more “lean.” Prior discussion has noted MRP’s
inconsistency with lean, particularly as it enables large lot sizes, poor yields and long lead
times by creating planned orders (PO’s) that allow for those conditions. This is in contrast to
JIT or Lean methodology that aspires to Lot-for-Lot (L4L) quantities, 100% yield and short
lead times. While this may be the utopian scenario, the fact is that many manufacturers exist
in a reality that requires planning for lot sizes and less than 100% yield. This method will
focus on maximizing the effectiveness of MRP planning by carefully analyzing the
processing sequence of yield and lot sizing.
The objectives are as follows:
To illustrate the effect of applying lot size or yield inappropriately
To distinguish where lot sizing should be applied (PO receipt or PO release)
To demonstrate how to use a computer algorithm to optimize MRP processing
There are many environments in which the shop floor must make frequent miscellaneous
issues from (or miscellaneous receipts to) the stockroom. These activities result from over- or
under-issues from the stockroom, even if the stockroom issued the correct quantity given on the
pick lists. How is this possible? The reason is that the MRP-generated quantities reflect the
theoretical requirements, which are often different from the practical need. One popular cause is
that the MRP parameters are not in sync with the “real world” conditions with respect to lot
sizing.
18
For example, MRP may suggest a PO release of 900 feet of cable that is issued on 1,000
ft. spools. In reality, the stockroom practice will be to issue an entire spool, thus requiring 100
feet to be returned after the job. In other cases, the operators may decide to “use up” the entire
quantity as a means of being efficient. This would likely require the stockroom to issue an
additional quantity of the other components. Note that the popular default would be for the floor
personnel to request an additional 10% of each item (100 ft/1000 ft = 10%), but this would be
incorrect. The additional cable was 11.1% extra material (100/900 = 11.1%). The result would be
many small issues back and forth between the stockroom and shop floor to have enough material,
or in some cases leftover would be tossed out, without the proper inventory adjustment. It should
be obvious that the correct procedure would be to adjust the MRP parameters such that the
resulting PO quantities model the manufacturing environment. This is not often the case, as floor
personnel cannot be bothered with working with a computer system or planners who “don’t
understand how operations work.” They will adapt by using the miscellaneous issue system, or
worse yet, by helping themselves to any needed material and/or returning extra without an
accompanying system transaction. It is thus the responsibility of MRP users, or the system gurus,
to ensure that lot sizing parameters are consistent with the physical environment.
The first item of consideration is to understand how MRP calculates lot sizes and yields.
Starting with the Net Requirements, MRP will adjust for the lot size by rounding up to it or a
multiple thereof. In this fashion, MRP is applying the lot size to the PO receipt. MRP will then
consider yield to adjust for scrap loss, resulting in a PO release of some larger quantity. Thus, the
sequence is to apply lot sizing and then yield (see Figure 1). [Note: for simplification all
examples in this study use the following assumptions: on hand = 0, no scheduled receipts, no
19
overdue orders, safety stock = 0, lead time = 1 period, round up to integer values on yields, lot
size = 100 and yield = 90%]
PERIOD 1 2 3 4 5
GROSS REQUIREMENTS 0 0 99 0 0
SCHEDULED RECEIPTS 0 0 0 0 0
PROJECTED BALANCE OHB = 0 0 0 1 1 1
NET REQUIREMENTS 0 0 99 -1 -1
PLANNED ORDER
RECEIPT
Lot Size = 100 0 0 100 0 0
PLANNED ORDER
RELEASE
Yield = 0.90 0 112 0 0 0
(Figure 1)
In this example, the order start quantity is 112 so that the ending quantity will satisfy the lot size
rule of 100. Calculations of this manner are based on the underlying assumption that completed
product (or perhaps a subassembly) will be “lot sized.” Some examples of this case are:
Finished goods packed into a carton or other handling container (boxes, crates, palletized,
etc.)
WIP transferred in standard containers (Kanbans, pallets, drums, etc.)
20
Justification for setup cost (note: this should then be a minimum, not a lot size multiple).
In the first two examples it makes perfect sense to apply the lot size rule to the PO receipt.
The third case is a common misuse of lot sizing. Given the fact that most MRP systems allow for
setting minimums, there is no excuse for applying a lot size rule here. Note the dramatic effect
that can result from this practice (assuming the parameters used in Figure 1): an order for 101
pieces will cause MRP to calculate a PO receipt for 200, and generate a PO release of 223 (see
Figure 2). The next step is to consider what environments justify alternative processing.
PERIOD 1 2 3 4 5
GROSS REQUIREMENTS 0 0 101 0 0
SCHEDULED RECEIPTS 0 0 0 0 0
PROJECTED BALANCE OHB = 0 0 0 99 99 99
NET REQUIREMENTS 0 0 101 -99 -99
PLANNED ORDER
RECEIPT
Lot Size = 100 0 0 200 0 0
PLANNED ORDER
RELEASE
Yield = 0.90 0 223 0 0 0
(Figure 2)
21
There are many examples for which material issued to a job is transferred in a lot size, and
thus the standard MRP processing is not ideal. In these cases it would be preferable to have the
start quantities relate to a multiple. One common scenario is the gateway work center, which
may be fed by vendor lot sizes as follows:
Reels of taped or spooled components
Rolls of paper or film
Tanks or drums of chemicals
Lengths of lumber or bar stock
Pounds (or Kg) of plastic molding compound
Pounds (or Kg) of paint pigment
Any item for which the quantity is “used up.”
Another case where the lot size should be applied to the start quantity is when the prior job
“batches” materials, such as:
Ovens – baked goods, PCB’s, ceramics
Acid bath or plating operation
Standard containers on incoming material
Any circumstance in which the incoming material was batched to “fill something up.”
These situations tend to create miscellaneous material transfers or “ghost” transactions that
are never reflected in the system. The resulting effect on inventory balances results in more
shortages (or overages) until the entire formal planning process has been replaced by a “damn
the pick list, take what you need” substitute. This can be avoided if the MRP processing applied
22
the yield rate first, then lot sized to a multiple (see Figure 3). Using the same numbers as with
Figure 1, we note that the PO release is now 200, satisfying the lot size requirement. Of
particular interest is the fact that in this scenario there is a “recalculated PO receipt” of 180
pieces based on the inflated start quantity, then adjusted for scrap loss. The casual observer could
easily make the mistaken assumption that this processing sequence generates excess inventory.
In reality, the MRP system has been fine-tuned to reflect what is actually happening in practice
on the shop floor. The added knowledge would be a benefit to planning, particularly in terms of
improved capacity requirements planning based on MRP output.
PERIOD 1 2 3 4 5
GROSS REQUIREMENTS 0 0 99 0 0
SCHEDULED RECEIPTS 0 0 0 0 0
PROJECTED BALANCE OHB = 0 0 0 81 81 81
NET REQUIREMENTS 0 0 99 -81 -81
PLANNED ORDER
RECEIPT
Yield = 0.90 0 0 110 0 0
PLANNED ORDER
RELEASE
Lot Size = 100 0 200 0 0 0
RECALCULATED PO
RCPT.
0 0 180 0 0
(Figure 3)
23
It should be apparent that the preferred MRP system would provide for the option of
applying the lot size on either the receipt or the release. As given by the example cases, one
should judge where the lot size is best applied and set the system parameters accordingly. For
example, a woodworking company manufacturing wooden “widgets” could apply a lot size rule
on the start quantity that is consistent with the incoming lengths of lumber (purchase lot size) to
avoid remainder quantities. There is an opportunity to extend this logic to a higher level. What if
it is preferred to set the processing sequence based upon the net requirements, such that the
resulting projected balance was always minimized? This case would exist when the lot size is a
“soft rule” with flexibility, such as minimum quantities for setups that might allow for
adjustment based on a trade-off with the resulting inventory cost, or any time where a conscious
decision can be made to override the lot size rule.
Consider the following examples that highlight the effect of processing sequence. Let us
assume there are two customer orders (CO), one for 25 pieces and another for 99. The following
cases demonstrate the resulting PO’s and remaining inventory balance (after filling the CO) for
each using both methods.
Case 1 (CO for 25 pc., standard MRP processing)
PO Release of 112; scrap 10% and thus receive 100; ship 25 and inventory 75 pc.
Case 2 (CO for 25 pc., apply lot size to release)
PO Release of 100; scrap 10% and thus receive 90; ship 25 and inventory 65 pc.
24
Given these two cases for the 25 pc. order, we would prefer to apply the lot size to the PO
release.
Case 3 (CO for 99 pc., standard MRP processing)
PO Release of 112; scrap 10% and thus receive 100; ship 99 and inventory 1 pc.
Case 4 (CO for 99 pc., apply lot size to release)
PO Release of 200; scrap 10% and thus receive 180; ship 99 and inventory 81 pc.
Given these two cases for the 99 pc. order, we would prefer standard MRP processing. This
suggests that we would like MRP to consider net requirements before deciding on which
processing sequence to use. This can be accomplished by utilizing the following computer
algorithm to calculate the PO’s based upon which technique will produce the minimum release
quantity (and resulting minimum balance).
[Note: The expression “CEILING (X, 1)” will round up the value of X to the next integer value
(i.e. 7.3 rounds up to 8)]
Ford’s MRP Processing Algorithm
IF NR <= 0, REL = 0; RETURN REL
REL1 = CEILING (CEILING (NR/LS, 1)*LS/YR, 1)
REL2 = CEILING (CEILING (NR/YR, 1)/LS, 1)*LS
REL = MIN (REL1, REL2); RETURN REL
RCPT = REL*YR; RETURN RCPT
25
Where the variables are as follows:
NR = Net Requirements
LS = Lot Size
YR = Yield Rate
REL = Planned Order Release
RCPT = Planned Order Receipt
This algorithm was used to produce the MRP report given in Figure 4, which matches the
optimum results from Cases 2 and 3 above.
PERIOD 1 2 3 4 5
GROSS REQUIREMENTS 0 0 99 0 25
SCHEDULED RECEIPTS 0 0 0 0 0
PROJECTED BALANCE OHB = 0 0 0 1 1 66
NET REQUIREMENTS 0 0 99 -1 24
PLANNED ORDER
RECEIPT
Lot Size = 100 100 90
PLANNED ORDER
RELEASE
Yield = 0.90 112 100
(Figure 4)
There are many other factors that can influence the desired goal of optimum MRP
processing. For example, there may be multiple items incoming to a work center with various lot
26
sizes. For the purposes of this analysis, it is assumed that there is a prevailing material that
dominates the process (such as the lengths of lumber in making wooden widgets), but this may
not always be the case. Further, there may be different effects depending on whether a scrap loss
is per assembly or per component. Generally speaking, MRP yield rates consider the scrapping
of the total assembly. If only a particular component is subject to scrap loss, this should be
handled by inflating the “quantity per” on the bill of material (BOM). It may also be useful to
consider the effects of safety stock, lead time and CRP. These considerations should be the basis
of further analysis of the MRP processing model.
The examples given clearly illustrate that standard MRP processing is not always ideal.
When the system fails to reflect the realities of the shop floor, substitute processes prevail.
Worse, the users may continue to put blind faith in computer-generated numbers that
ultimately increase inventories. As with most situations, the fault lies not within the MRP
system, but in the designers who failed to optimize parameters or processes for a given
situation. Systems analysts should look for the opportunities to improve the performance of
their system and make it work for their environment.
27
IV. Results
MS Excel was utilized to demonstrate how MRP may be performing on a suboptimal
basis. The following options were presented as methods of calculating planned orders:
standard MRP processing; applying the lot size to the PO release; and using a computer
algorithm to determine where to apply the lot size rule. Optimizing the calculation of planned
orders reduced both the lot sizes of planned orders and resulting inventory balance. This has
multiple positive effects, first considering the reduced lot size:
Smaller lots implies shorter runs, hence reduced overall manufacturing lead time
Reduced lead time implies less WIP and quicker inventory turns
Building in smaller lots means money was not wasted to create excess inventory
on the shelf, saving the associated materials, labor and overhead
Smaller lots make it easier to quickly identify any quality issues
Smaller lots facilitate easier handling and movement, approaching flow
The effects of reduced inventor balance include:
Savings on warehouse space, labor and overhead costs
Savings on the opportunity cost (cost of capital) avoided
Savings on risk costs (insurance, obsolescence, damage, theft)
The MS Excel spreadsheet can also serve as a tutorial for demonstrating MRP’s key
processing functions (lead time offset, netting of requirements, and parent-to-component
explosion). This tool has been utilized in a multitude of training environments to provide
MRP training to staff representing a wide array of company functions. For example, a major
28
client optimized their MRP processing parameters and projects a 46% reduction in planned
order releases, a vast savings in planner time and effort.
29
V. Conclusions
Many opportunities exist to resolve the MRP versus lean dilemma. A starting
approach is to recognize the advantages and limitations of each technique. MRP serves as
a planning tool that can simulate “what if” scenarios at a higher level. Lean best functions
as an execution tool on the shop floor to facilitate flow.
Opportunities to make MRP “leaner” include the optimization of planning
parameters, reducing lead times, lot sizes and inventory balances, adopting demand-
driven MRP principles, and utilizing modular BOM’s. Opportunities to make lean less
obstructive to the planning process include strategic buffer stocks, value stream mapping,
and adjusting the push/pull intersection.
The ideal scenario may be a combined ERP/LSS model that melds the best of both
techniques, with ERP’s extensive data capturing capabilities serving LSS’s need for
evidence-based inputs/outputs. In this fashion, the organization increases productivity
while improving customer service. Best of all, the employees will work within the system
rather than resisting the change imposed upon them.
30
VI. Appendix
Sample MRP Grid
Key:
OHB – on hand balance
Alloc – allocated quantity
LT – lead time
LS – lot size
S/S – safety stock
YR – yield rate
PAB – projected available balance
PO – planned order
OHB = 75
Alloc = 25
LT = 3
LS = 100
S/S = 10 Periods
YR = 100% 1 2 3 4 5 6 7
Gross Requirements 33 8 47 47 5 63 30
Scheduled Receipts 100
Projected Available 50 17 109 62 15 10 47 17
Net Requirements 63
Planned Order Receipts 100
Planned Order Release 100
PAB (start) = OHB - Alloc = 75 - 25 = 50
PAB (1) = PAB (start) - Gross Req. = 50 - 33 = 17
Math Check PAB (2) = PAB (1) + Sched. Rcpt. - Gross Req. = 17 + 100 - 8 = 109
OHB 75 PAB (3) = PAB (2) - Gross Req. = 109 - 47 = 62
+ Supply 200 PAB (4) = PAB (3) - Gross Req. = 62 - 47 = 15
- Demand 258 PAB (5) = PAB (4) - Gross Req. = 15 - 5 = 10
End Qty. 17 PAB (6) = PAB (5) - Gross Req. = 10 - 63 = -53 (therefore a net req. exists, noting S/S of 10)
Since we have a net req., we must have a planned order, noting lot size of 100.
The planned order release is offset by the lead time of 3 weeks from the receipt. Recalculate PAB.
PAB (6) = PAB (5) + PO Rcpt. - Gross Req. = 10 + 100 - 63 = 47
PAB (7) = PAB (6) - Gross Req. = 47 - 30 = 17
Make sure that each PAB value satisfies the safety stock rule (10).
Make sure that there is a PO Receipt every time there is a net requirement.
Make sure that there is a PO Release offset by the lead time (3 periods) from every PO Receipt.
31
MRP Exercise Calculations
When performing exercises using the MRP planning grids, the following
protocol should be used.
A. Gross Requirements are determined by forecast or actual sales
orders for MPS (low-level code of 0) items. Gross
Requirements for subassemblies & components are the sum of
all PO Releases for the parent items (times qty/per), plus any
independent demand. The Gross Requirements occur in the
same time bucket as the demand (PO Release).
For example, given Component 1, with Parent Items A and B:
Period 1 Gross Req. (Component 1) = Period 1 Assembly A PO
Release (x qty/per) + Period 1 Assembly B PO Release (x qty/per).
B. Net the Gross Requirements against existing inventory
(Projected Available Balance of prior period) and any existing
Scheduled Receipts (Open Orders) of the current time period,
adding in safety stock. This gives the Net Requirements as:
Net Req. = Gross Req. – PAB (prior) – Sched. Rcpt. + S/S
(only consider Net Req.’s w/positive value)
C. Planned Order Receipt = Net Req. if using lot-for-lot; else, use
the lot size or a multiple thereof.
To get the Planned Order Release quantity, divide the PO Receipt by the
Yield Rate (YR = 1 – Scrap Rate). Offset by the lead-time from the PO
Release time bucket to get the correct period to release the order.
32
33
Volume (x-axis)/Variety (y axis) Matrix
1. ETO – fixed position layout; project scheduling techniques such as CPM, PERT, Gantt
charts; high skill labor; general purpose machinery & equipment; high profit margins;
unique customized items; long LT; WIP, generally no stocked inventory; innovative, hi-tech
and/or trendy products. Management style supports empowered and flexible workforce
driven by need to be innovative; customer orders for deliverables, forecasts for budgets,
staffing levels, programs. PLC – introduction.
2. MTO – jobbing/batching; functional or process layout; MRP/CRP; wide variety of
products, lead times, routings, & lot sizes; long queues; lots of NVA (non-value added)
movements, handling and inventory; WIP & RMs; customer orders for end items, forecasts
for RM, staffing and equipment requirements. PLC – growth and/or decline.
3. ATO – could use cellular or job shop layout, possibly some elements of lean/pull; may need
an FAS (final assembly schedule) for customer configuration; inventory includes WIP and
stocked sub-assemblies; combination of forecasts (subassemblies) and customer orders
(FAS). PLC – growth/mature/decline.
4. MTS – repetitive mass production line for discrete items or pure continuous flow for
processes such as pharmaceuticals, food/beverage items, slurries, gases, chemicals, liquids,
etc.; dedicated automated lines; specialized equipment for high volume production of
limited variety items; low margins; short LT; standardized generic commodities; rate-based
scheduling; limited WIP, FG inv. (perhaps at a DC); lean and efficient; low labor skills.
Management style tends to be top-down command and control, driven by cost efficiencies;
forecasts for end items. PLC – mature.
MC – mass customization, attaining both high volume and high variety (ex: Dell computers)
MTO
ATO
MC
MTS
ETO
34
Top Reasons LEAN Environments Use MRP/ERP Systems
To store and maintain costing information.
To store and maintain BOM configurations.
To maintain revision levels and facilitate engineering change.
To comply with federal regulation mandating lot tracking (e.g. food, pharm)
To facilitate time fence policy.
To facilitate customer service with respect to order servicing.
To facilitate customer service with respect to product configuration.
To provide suppliers with a forecast (purchase items’ planned orders).
To facilitate response to changing demand conditions (volatile markets
and/or seasonality).
To allocate inventory to work and/or sales orders.
To maintain safety stocks.
35
VII. References
Abuhilal, Laith, Ghaith Rabadi and Andres Sousa-Poza. “Supply Chain Inventory Control: A
Comparison among JIT, MRP, and MRP with Information Sharing Using Simulation.”
Engineering Management Journal. Jun2006, Vol. 18 Issue 2, p51-57. EBSCO Database.
Arnold, J. R. Tony, Stephen N. Chapman, and Lloyd M. Clive. Introduction to Materials
Management (6th Ed.). Columbus. Pearson, 2008. Print.
Bartholomew, Doug. “Can Lean and ERP Work Together?” Industry Week. Web.
http://www.industryweek.com/systems-integration/can-lean-and-erp-work-together
Bartholomew, Doug. “ERP Learning to be Lean.” Industry Week/IW. Jul2003, Vol. 252 Issue 7,
p19. EBSCO Database.
Bartholomew, Doug. “Lean Gets a Software Assist.” Industry Week/IW. Oct2006, Vol. 255
Issue 10, p25-25. EBSCO Database.
Bartholomew, Doug. “Lean vs. MRP.” Industry Week. Web.
http://www.industryweek.com/articles/lean_vs_erp_413.aspx
Bertolini, Massimo, Marcello Braglia, Giovanni Romagnoli and Francesco Zammori. “Extending
Value Stream Mapping: the Synchro-MRP Case.” International Journal of Production Research.
Sep2013, Vol. 51 Issue 18, p5499-5519. EBSCO Database.
Bradford, Marianne, Tony Mayfield and Chad Toney. “Does ERP Fit in a LEAN World?”
Strategic Finance. May2001, Vol. 82 Issue 11, p28-34. EBSCO Database.
Čiarnienė, Ramunė and Milita Vienažindienė. “Lean Manufacturing: Theory and Practice.”
Economics & Management. 2012, Vol. 17 Issue 2, p726-732. EBSCO Database.
Chopra, Sunil and Peter Meindl. Supply Chain Management: Strategy, Planning, and Operation
(6th Ed.). Boston. Pearson, 2016. Print
Dilhe, Rue. “Leaning Out”. Works Management. Sep2013, Vol. 66 Issue 8, p32-33. EBSCO
Database.
Dixon, Dave. “The TRUCE between LEAN and I.T.” Industrial Engineer: IE. Jun2004, Vol. 36
Issue 6, p42-45. EBSCO Database.
Fogarty, Donald W., John H. Blackstone, Jr., and Thomas R. Hoffman. Production and
Inventory Management (2nd Ed.). Cincinnati. South-Western Publishing Co., 1991. Print.
George, Michael L. Lean Six Sigma: Combining Six Sigma Quality with Lean Speed. New York.
McGraw-Hill, 2002. Print.
Gill, Rebecca. “Lean Manufacturing and ERP Systems: Different by Design.” Ceramic Industry.
Aug2007, Vol. 157 Issue 8, p19-20. EBSCO Database.
36
Harding, Lindsay and Carol Ptak. “Could Demand-driven MRP Be the Solution We Have Been
Looking For?” Operations Management (1755-1501). Dec2012, Vol. 38 Issue 6, p7-12. EBSCO
Database.
Harris, Chris and Chuck Streeter. “A New Purchasing Philosophy.” Industrial Engineer: IE.
Sep2010, Vol. 42 Issue 9, p42-46. EBSCO Database.
Heizer, Jay, and Barry Render. Operations Management (7th Ed.). Upper Saddle River. Pearson,
2004. Print.
“How Demand Driven MRP enables Lean Manufacturing to become practical in complex
environments.” Web. http://www.beyondmrp.com/demand-driven-mrp-lean-manufacturing/
Imaoka, Zenjiro. “Understand Supply Chain Management through 100 words.” Web.
http://www.lean-manufacturing-japan.com/scm-terminology/mrp-materials-requirements-
planning.html
Jones, Daniel T. “Heijunka: Leveling Production.” Manufacturing Engineering. Aug2006, Vol.
137 Issue 2, p29-36. EBSCO Database.
Kanet, J. J., "MRP 96: Time to Rethink Manufacturing Logistics," Production and Inventory
Management, Vol. 29, No. 2 (2nd Qtr. 1988), pp. 57-61. EBSCO Database.
Karmarkar, Uday. “Getting Control of Just-in-Time.” Harvard Business Review. Sep/Oct89, Vol.
67 Issue 5, p122-131. EBSCO Database.
Mclean, Tim. “Now we are Lean - why have ERP systems?” Manufacturers' Monthly. Nov2006,
p22-22. EBSCO Database.
Meyer, Kevin. “MRP versus Lean”. Web.
http://kevinmeyer.com/blog/2006/05/mrp_versus_lean.html
Miller, Jeany. “What is the Difference between MRP and Lean?” Wise Geek. Web.
http://www.wisegeek.com/what-is-the-difference-between-mrp-and-lean.htm
Narisimhan, Sim, Dennis W. McLeavey, and Peter Billington. Production Planning and
Inventory Control (2nd Ed.). Englewood Cliffs. Prentice Hall, 1995. Print.
Nauhria, Yugal, S. Wadhwa and Sunil Pandey. “ERP Enabled Lean Six Sigma: A Holistic
Approach for Competitive Manufacturing.” Global Journal of Flexible Systems Management.
Jul-Sep2009, Vol. 10 Issue 3, p35-43. EBSCO Database.
Pitcher, Michael. “How to Combat Push Inventory Systems.” Official Board Markets.
12/16/2006, Vol. 82 Issue 50, p10-11. EBSCO Database.
Powell, Daryl, Jan Riezebos and Jan Ola Strandhagen. “Lean Production and ERP Systems in
Small- and Medium-sized Enterprises: ERP Support for Pull Production.” International Journal
of Production Research. Jan2013, Vol. 51 Issue 2, p395-409. EBSCO Database.
37
Ptak, Carol and Chad Smith. “The State of Demand Driven MRP.” Demand Driven Institute.
Web. http://www.demanddrivenmrp.com/
Ross, David Frederick. Distribution Planning and Control: Managing in the Era of Supply
Chain Management (2nd Ed.). New York. Springer, 2004. Print.
Sandras, William A., Jr. Just-in-Time: Making it Happen. New York. John Wiley & Sons, Inc.,
1989. Print.
SINK, D. SCOTT. “Finding Solutions with Imperfect Information.” Industrial Engineer: IE.
Jul2014, Vol. 46 Issue 7, p34-39. EBSCO Database.
Sinocchi, M. and Ralph Bernstein. “Lean Versus MRP”. Lean Insider. Web.
http://leaninsider.productivitypress.com/2011/12/lean-versus-mrp-or-lean-and-mrp.html
Steinbrunner, Dan. “The Happy Marriage of Push and Pull.” Industrial Management.
Jan/Feb2005, Vol. 47 Issue 1, p27-30. EBSCO Database.
“The Best of Both Worlds”. Demand Driven Technologies. Web.
http://demanddriventech.com/home/the-best-of-both-worlds/
Vollman, Thomas E., William L. Berry, D. Clay Whybark, and F. Robert Jacobs. Manufacturing
Planning & Control Systems for Supply Chain Management (5th Ed.). New York. McGraw-Hill,
2005. Print.
Womack, James P., and Daniel T. Jones. Lean Thinking (2003 Ed.). New York. Free Press, 2003.
Print.