Production Scheduling – I

Nothing is perfect.  The production scheduling model is no exception.  The assumption, in other words, is the limitation.  The 2 models have the following assumption:

Single Processing:

1. Materials is assumed to be ready when the production starts
2. The production only involves 1 process to complete or all processes is within a single production line
3. The expected delivery date only mention the date with no concern to the time, let alone shift
4. For the basic model, there are only 2 shifts, day and night, user can change the daily production hours.  The daily timeframe is measured in hours.  Time 0 is the beginning of the shift, time 1 is the end of first hour.
5. There can be day gap between the production completion date and expected delivery date.  Again, it is in terms of day, not hours
6. The allocation mechanism is to fill up one production line capacity before moving the reminants quantity to other line.  As a result, a following situation may occur: Batch C with quantity 10k starts to produce at day 4 in Line I daytime but the deadline is day 5 daytime, the time left can only produce 8k, 2k needs to be produced by Line II.  Since there is no production in Line II, it starts to produce from day 1.  This is an extreme case.
7. Production line is operating in a continuous manner sharply, i.e. for Line 1, day 1, Daytime, if the last batch ends at Time 9.5 while the next batch set up time is 0.8, 0.5 is done on Daytime, 0.3 on the subsequent Nighttime.  Yes, it is theoretically mechanical.

Sequential Processing:

1. Materials is assumed to be ready when the production starts
2. The production scheduling is based on push mechanism, i.e. not start from the delivery date, then packaging process and back to the first process; rather, plan the first process of all production batches, then move forward to subsequent process, finally, determine the completion date
3. In between 2 subsequent processes, there may be time lag from the 1st process to 2nd process, e.g. cooling time or transport time is needed.
4. Each production has pre-defined routing with hourly capacity known in each process for each single product
5. Each order demand is broken down to production batch, e.g. an order of 15k with production batch of 6k, the order will be broken into 3 batches, 6k, 6k and 3k
6. Production batch allocation is on all-or-nothing basis, whole production batch is allocated to a single production line, it will not be further allocated to >1 line

It is a little bit clumsy, agreed

Production Scheduling – 0

Let’s do something Back to Basic.  What should the production scheduling be able to achieve?

1. It should tell either it can meet the expected delivery date or tell when it can be completed
2. It should tell when to do what for each production line
3. It should tell when to change the production line/die cast so that the production people can have better plan for changeover
4. If different product needs different headcount in a single production line, it should tell how many headcount is needed at what time
5. It should be able to tell when there is waiting time (Except for 1-piece flow, it is unavoidable to have waiting time for sequential processing, what we can do is to minimize it)

What factors should be consider in the production scheduling?  (Sorry that due to the shortcoming of my model, I mark those I can handle with blue color while red color which my model cannot.)

1. Time lag between delivery date and production completion date (the perfect situation is that they are the same but a little bit risky)
2. Hourly capacity of a machine/a line
3. Headcount required for the above
4. Set up and set down time, let’s do some elaboration.  Set up is the time spent to lay the production line or upload the die.  Set down is the time to clear the production line / die so that it can be used for the next product
5. Set up and set down time is not required in between 2 batch of production if they are the same product
6. Not everyday is working day.  The working hours can be variables.
7. If a production line is not sufficient to finish the order, it should automatically shift the rest to other production line
8. Availability of mold/tools

If you have more suggestion, please leave your comment.  Many thanks

Adversity is a blessing

I am excited recently because I have resolved all the core technical issues of the production scheduling, which I make use of advanced Excel functions together with a number of Excel/VBA function.  There are 2 models, namely as follows:

1. Single Processing, which can be applied to 1-piece flow, injection machines or any production with only 1 process
2. Sequential Processing, which can be applied to production involved > 1 work centers

Actually, I came across with this idea to realize the production scheduling model on Sep, 2010 when I decided to leave the sports manufacturing company and took up a contract job of supply chain solution(work out a production scheduling solution based on MS Excel) with the client of bottle manufacturer in mainland China.  Before I commenced the contract job, there was 1 month break.  I used the spare time to initiate the models mentioned above, it is really complicated and 1 month is not enough to sketch everything.  Afterwards, I was busy with the contract job and could only spend limited time to finish the models.  The result is not good and my training agent urged me to launch it which I have rejected.  However, when the job is going to be ended, I have more time though I was worried with what I could do next.  I thought being a consultant may be my next career.  I use my spare time to re-visit the draft models.  Many thanks to the consulting job, I got my Excel skill upgraded to another level and I did find out a lot of bugs.  I spent another few weeks to fix them all.  This week, I have finished it after all.

I thought if I were not dropping, I could not afford to spend so many weeks/months to work out the models.

I thought if I were not free, I could not be calm down to think it over the model and sort out the bugs.

I thought if I were not working in the wire harness company, I could not have the opportunity to experience the importance and the work of production scheduling.

Anyway, the models may not be perfect (actually, it cannot be) but it provides a start-up platform/framework for further works.  It is a good experience.  I will do more testing by soliciting help from my network before I roll out for public training. 

If you are interested, drop me a note and I will provide you all the documents/Excel files.  But it will be limited to a few first comers.

If you would like to discuss and know more, you could visit the following link and participate to the discussion:

http://www.linkedin.com/groups/Accountants-in-Manufacturing-Industry-2445034?gid=2445034&trk=hb_side_g

Starting form this week, I will write a number of blog entries to explain some concept I come across with the models.