Graphic representation of a Lean schedule

6 Comments

1. Manny Fernandez, Jr.
   February 12, 2012 @ 9:13 am

   Comment in the NWLEAN discussion group on Yahoo!

   I recall when going through the APICS body of knowledge, the problem that most organizations have is that they do not understand their capacity to accomplish work at the individual work center level. Consequently everything is an estimate. There is no Standard Work, so processes take as long as they take a particular person to do the task. Lead times are estimates and rarely are accurate. Therefore the systems have a lot of garbage in and garbage out. Work is not level loaded, work centers are not maintained. So it should not be a surprise to get what one gets. In order to fix this situation it takes a bit more than just releasing work orders in the correct sequence. Am I missing something here?
   Manny

   Reply

   • Michel Baudin
     February 12, 2012 @ 9:15 am

     The point of using caps on WIP to control production in a job-shop is that you don’t need all the detailed measurements for all routes to make it work. In fact, it works even in service situations where the service time for each customer at each operation varies randomly. I have seen it used successfully in an R&D lab developing a semiconductor process and where every job was an experiment on one or more operations within the process.

     It is Little’s Law at work. Starting from an empty shop, if you release more and more jobs, at first the output increases. Once the shop is saturated, output plateaus and while WIP rise and lead time rise. By capping the WIP at a level that you determine by trial and error, you can set the output rate at the level you want without making the WIP explode, regardless of how the flow is managed inside the shop.

     In a job-shop, each job’s route is planned individually and printed as a traveler, that stays with the job and lets every operator know where the job should go next. What I usually recommend is to process the jobs on a FIFO basis at every operation on their routes. Since the WIP caps keep the overall lead time low anyway, there is not much to be gained by playing
     scheduling games inside the shop.

     Of course, if you cap the WIP, you cause jobs to wait before release, but the only thing that actually waits is the traveler. It is when you release the job to production that the traveler mates with materials, tools, fixtures, jigs, programs, etc. and the consequences of messing with it reach much farther.

     Of course, this method is not appropriate in a shop with high-volume products or product families going through fixed routes, but then the data you need exists and is much smaller than in a job-shop. In an aerospace or automotive machine shop, the WIP cap method would be applicable in the tool room where you make custom fixtures but not on the main floor where you machine parts repetitively in cells.

     Reply

2. Dan Barch
   February 19, 2012 @ 5:38 am

   Comment in the NWLEAN discussion group on Yahoo!

   Hey, Prasad,

   You gave us good detailed observations, I had never seen empirical proof of the concept. But the website also claimed “On the other hand, if jobs are released very early, then WIP and job lead times will increase.” The goal of the question was to find the root problem, and maybe there is another way to fix it. Mr Bauden expressed this with WIP capping, or strict FIFO – if you control your processes, they will not grow and slow no matter what orders are in there – the orders will not be started until the shop is ready no matter of release.. It seems to me that if you confuse people with too many choices, they slow down, and mix things up. Controlled release of WO is one countermeasure. Standard work is another. WIP capping is another. Heijunka levelling, line balancing…….. All of them together start to form a system that is able to ‘self correct’ in that the operators are not confused with orders, and concentrate on working well and better.

   Lean is Learning,
   Dan

   Reply

3. Prasad Velasa
   February 19, 2012 @ 5:39 am

   Comment in the NWLEAN discussion group on Yahoo!

   Hi Dan,

   You raised one good point which I did not cover on my web page.

   For any system with one or more bottlenecks, job releasing beyond a certain rate will increase WIP and lead times. To control WIP and lead times, there are two measures: (1) controlling job releases, (2) proactive dynamic capacity planning, and (3) mix of both. I talked about the first one on the web page. Many people in industry look at the second and third measures also.

   Mr Baudin’s WIP capping plus FIFO rule is helpful in systems with a single process flow or muiltiple parallel process flows (through different cells). However, in job shops where jobs have different process flows and varying process times at different stages and flow through the shop simultaneously competing for some shared resources, this may not be efficient although it nicely appeals to our common sense.

   Regards,
   Prasad

   Reply

4. Michel Baudin
   February 19, 2012 @ 5:52 am

   @Prasad – On Wed, 15 Feb 2012 16:07:40, you wrote:

   > … WIP capping plus FIFO rule is helpful in systems with a single
   > process flow or muiltiple parallel process flows (through different
   > cells).

   This is not what it’s for, nor where I have seen it work. It is for job
   shops. By this I mean shops where you have processing resources through
   which you run jobs, and the jobs are varying networks of tasks that can
   each be performed by some of the resources with varying process times. The
   jobs may even involve dynamic branching in the sense that the outcome of a
   task may change what comes next either for the whole job or for part of it.

   There may be a high-level sequence of phases that jobs go through, but at
   the level of “the herd keeps moving west,” as opposed to “go from this
   drill press to that lathe.”

   Examples of such shops include:
   – The development lab for a manufacturing process.
   – Fixture and jig making in a machine shop.
   – Repair operations.
   – The production of products categorized as “Strangers” in your
   Runner/Repeater/Stranger analysis.

   The approach applies also to situations that are slightly more repetitive
   like machine refurbishment/overhaul, patient processing in health care, or
   final test in manufacturing.

   The WIP capping/FIFO rule requires little data for production control, but
   you still needs the detailed process plan for each job, and it is usually
   on a printed traveler with operator check marks for the work already done
   and specs on the remaining tasks.

   Best regards.

   Reply

5. Graphic Representation of a Lean Schedule « Neovista Newsfeed
   February 27, 2012 @ 8:58 am

   […] on over here to view the graphic. Share this:PrintEmailFacebookLinkedInTwitterLike this:LikeBe the first to like […]

   Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.