Over the past several months, we have published a number of articles and blog posts concerning MRP, its origins and its role – if any – in lean manufacturing. More than any other topic we raise, it brings out comments and opinions, as the manufacturing community wrestles with the role of this huge driver of how manufacturing has been managed over the last few decades. With over a billion dollars expected to be spent on ERP systems in the next year, and with thousands of MRP/ERP applications already in place, the significance of this discussion is great.
While we will eventually organize the material concerning MRP in a more structured manner, for now I am using the blog pages to bring together the various MRP related articles, as well as some of the more interesting emails I have received on the subject. In particular I want to thank Gene Thomas of Configuration Management Solutions and Joel Wisner from the University of Nevada – Las Vegas for their valuable additions to the discussion, which you can read in the comments below.
I would like to invite the entire manufacturing community to weigh in on the subject. The goal is to eventually compile the best of all of our thinking into a comprehensive white paper on the role of MRP in lean that can be of value to us all.
Following are links to the documents that we have published in the past:
Farewell To APICS – my blog post discussing the decline of APICS in their influence on manufacturing which I ascribed to their reluctance to replace MRP with lean manufacturing as the core of their body of knowledge
Kevin’s Recent post The False God of the Almighty Algorithm
My article MRP R.I.P.
An excerpt chapter from Rebirth of American Industry entitled The Illusion of MRP
I urge everyone to read the insightful additions made by Gene Thomas and Joel Wisner in the comments, and to add any comments of your own that might advance all of our understanding of this very important topic.
Gene Thomas says
From Gene Thomas
Got a big kick out of your response to my quickie blog entry. I’d be happy to give my recollections of the early MRP era, but I think it will delve into a lot of additional insight into where we will be heading through all the consolidation happening in the industry lately. I intend
to go into quite a bit of specific detailed historical events and it will likely take a couple more weeks to build up. I also have a host of packratted examples copied and scanned over the past 50 years that I need to find the best way to communicate, such as Orlicky’s load v.s. capacity printout of 8/15/67 at J.I. Case.
Preliminarily, MRP was being done by a number of large manufacturers
such as John Deere, J.I.Case, IBM, Allis-Chalmers, Collins Radio etc in the mid ’50’s, but with tape oriented computers. Single level BOM’s & routings were stored electronically on sequential tape and passed level-by-level, netted at low-level code to generate time-phased material and labor requirements. Separate runs were required prior to each requirements generation to maintain a current low-level-code, and monthly generations were the norm with 12 monthly time buckets. Then, drum oriented computers (IBM 650) had enough storage to handle a monthly master schedule, but the BOM’s then had to be turned upside down in a summarized where-used form to pass against the randomized MPS to extend the requirements. This caused an enormous problem of maintenance of the single level parts lists on drawings to develop and maintain the total usage quantities per top level model of the MPS. Then the IBM 305 Ramac was announced, each box consisting of 50 discs and tracks where about 5 million characters were accessible randomly (about one second per seek/read-write-protect cycle).
This is where I came in as the IBM Mfg. Industry Mgr for John Deere and spec’d out the “BOMP” bill-of-material/where-used package in ’61 (which still included some board wiring that wasn’t packagable). In the meantime I worked with Orlicky, Fred Brani (JIC MIS) and our Milwaukee IBM staff (Ted Musial) with implementation of the first net-change Ramac customized MRP system. BOMP wasn’t finished yet (being developed by us with Allis Chalmers and Cutler-Hammer with where-used functionality), so JIC had to upgrade downstream with the larger 1410 Ramac and BOMP. Many subsequent implementations of BOMP were made around Milwaukee and the use of one of its retrievals, the summarized explosion, gave rise to the potential of packaging a requirements planning system.
As a result, in ’63, we packaged the “LAMP” (Labor & Material Planning) system interfacing the BOMP summarized explosion with a time-series item master file (sometimes even on punched cards for smaller companies) to test the viability of getting engineers lined up to maintain single level “parts lists”. We then upgraded these efforts with IBM Industry development of PICS-RPS (Production & Inventory Control System–Requirements Planning System) and published through IBM application package programs and brochures which began in the mid ’60’s.
This is when Ollie/George and Joe came together–Ollie as the world’s best communicator, and Orlicky as the most pragmatic insister of net-change. Ollie sponsored refined educational material taught by his firm and IBM, and convinced Joe to join IBM to further his experiences and prognostications. There will be many war stories to come about the architecture of the MRP-ERP saga that I will try to put together for this audience. I’d suggest that you might want to post this entire communication to the blog to give a chance for interim commentary until I can get more to add.
Thanks for your attention span.
Gene Thomas
Founder Emeritus
Configuration Solutions
847-382-0680
Gene Thomas says
Just a couple quick clarifications of your response: I was an Iowa State IE graduate and then an IBM Systems & Salesman in Des Moines with Maytag and John Deere as early customers. Maytag and Deere-Dubuque was where the 650 drum summarized where-used architecture was in use in the late ’50’s.
I then covered the IBM plant in Rochester, MN as an internal salesman coverage plus Mfg. Industry Mgr covering the Midwest. During this time I published the internal IBM spec’s for BOMP, and worked with Orlicky and our IBM Milwaukee staff on the Case effort for their Claussen works in Racine.
Joe had previously been PC Mgr at Kearney & Trecker and moved to Case insisting on development of a net-change system. It was implemented on the 305 Ramac with customized BOM’s using the Ramac 5-character disc address as part-numbering with the part/drawing number as indicative reference (I think I still have a record layout of the item master and product structure somewhere). It included about 12 time buckets that were used for annual MPS forecasting and then as weekly for actual customer orders. Their configuration process consisted of a 9 character generic code with each position referencing an option modular planning BOM selection. This implementation was in parallel with the local development effort I then managed as the Mfg Sales Mgr for Milwaukee with Allis-Chalmers, consisting of the new where-used function not built into the Case system (a big difference).
I don’t recall “Papa Joe’s” (as we respectfully called him) educational background, but we were working together on IBM’s plan for a humongous centralized planning system for their plants in ’70 when I left IBM to form a manufacturing software/consulting company back in Milwaukee. I’ll attach my bio which gives you a much better background of my career. If you can get these posted to the IW forum, it might be fun to solicit other contributions that have recollections over these eras. I have been very impressed with the submissions of Bill Waddell and Kevin Meyer to this blog, and would like to volunteer some help in resurrecting some of this folklore.
Gene Thomas
Founder Emeritus
Configuration Solutions
847-382-0680
student says
MRP is massively flawed. It does not deserve the name planning system. It is not capable to produce feasible plans.