1

A typical Master Production Schedule (MPS) should have:

  1. What you need to produce
  2. How much you need to produce
  3. When you need the items

However, the issue arises when in reality, the jobs in our factory cannot be plotted like the MPS below:

enter image description here

Since we are a job shop, the jobs to be made have unknown production rates due to different specifications (size, design, machine used, and etc.). Basically, each job is almost always unique in some way to other jobs.

So we do not know exactly how much to produce in a day as we just keep working on them until the jobs are finished. This means the MPS as shown above is not going to be practical in our application.

The only information provided to us that I can think of is shown below:

enter image description here

With these information, when to set these jobs and start producing them will depend on the deadline and how much is needed to be produced. So what is a good production schedule for these types of jobs (assuming there are 100 jobs and you cannot just eyeball them like the example shown here)

  • I would record history of production times , how many products etc. then based on the data i could estimate production times , basically 0 -> 5000 products is set one , 5000 -> 10000 is set 2 , number of product can be divided into different sets – mussdroid Sep 7 '19 at 18:38
  • @mussdroid it is very difficult to quantify the production times as it is a job shop. So a new job have no historical data. Although, there may be a way to predict the production rates using design of experiments, but these take a long time and adds in complexity. Although I am more interested in the methodology to sequence the jobs (can assume production rates for the mean time) – Pherdindy Sep 9 '19 at 9:09
1

The production schedule shown at the top of the question is suitable for continuously selling products where the aim of the production is to keep the stock levels in the warehouses stable.

If you instead you have jobs that must be produced in their entirety before date X and then it is done, then I would use a different planning technique. For each job, a bit more information is needed than what you presented here. In particular, you want to know how long it takes to produce one item (so you can calculate how long it takes to produce the whole lot) and which production lines are capable of producing the items.

Then it becomes essentially a classic project planning exercise, where you have a certain number of resources (production lines) and work (the jobs) that must be completed by a deadline. And the work can even be spread over multiple resources without penalty.

  • Right I know my example isn't as comprehensive as adding production rates and machinery that are capable of doing the job will ramp up complexity. I really do want to know what classic project planning exercises is possible to employ and how to apply them in my case. I can add in assumed production rates and machines in the main post if needed. – Pherdindy Sep 9 '19 at 9:20
  • 1
    I believe these types of problems cannot be solved exactly by using optimization techniques and often times heuristics are used so I am not sure if I should have asked this in project management or operations research stack exchange community. I am hoping for the start-to-end example to have a good idea on how it would be scheduled rather than a general idea to it as I also have a general idea but no idea how to approach in detail. – Pherdindy Sep 9 '19 at 9:21
1

Matching many resources with different capabilities with many orders in time is one of the NP-complete scheduling problems. This means that in its general form it cannot be solved in polynomial time.

There are different methods to approach this problem you can use.

If you are looking for solutions, you can use one of the so-called Advanced Planning and Scheduling software tools. The company I work for uses Preactor (which is now acquired by Siemens), but there are other similar pieces of software on the market.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.