What Makes Shop Floor Scheduling So Challenging?

As we entered the late ‘90s, it certainly seemed as though manufacturers had everything they needed to control production, even in the most challenging situations. The material requirements planning (MRP) models of the ‘60s had advanced to include forecasting and master production scheduling, and shop floor scheduling had evolved from simple infinite capacity scheduling to finite capacity scheduling.

Despite these advancements, production control remains, well, out of control.

To better understand why this is, we need to walk through the steps it takes to manage production using these tools, starting with a sales order from a customer for a make to order (MTO) part. For the sake of this exercise, let’s say that the customer has requested that 60 pieces of Part A be delivered exactly one month from now. Here are the production control steps for that MTO part:

1. The MRP system looks at the sales order and creates a planned work order (PWO) for 60 pieces of Part A.
2. The MRP system then looks at Part A’s bill of materials (BOM) and creates a PWO or a planned purchase order (PPO) for every part on the BOM that isn’t readily available from stock or doesn’t have an existing work order or purchase order.
3. The MRP system also calculates the date which each underlying PWO or PPO needs to be either made or received by in order to assemble all the components in Part A for delivery in 30 days.
4. Next, a “planner” looks at each PWO and determines if it is accurate. If a PWO is accurate, the planner will “firm” it and send it to the shop floor to be manufactured.
5. Likewise, a “buyer” does the same for the PPOs, except those go to a supplier to be fulfilled.
6. Any WO that is sent to the shop floor has to have exact operation steps with very precise run times and set up times to ensure that the finite scheduler produces an accurate schedule.
7. Since this is an MTO job, the manufacturing engineer needs to review the operation steps for accuracy and make any necessary adjustments.
8. The WO is finally ready to be produced.

Now, imagine this scenario, except the customer calls back three days later to request 60 pieces of Part A one week earlier; the planner would have to adjust all of the WO quantities and dates, and the buyer would have to do the same for the PPOs.

Remember, modern enterprise resource planning (ERP) systems inherited their MRP logic from models build in the ‘60s. Back then, manufacturing predominantly made to stock (MTS) and there were few changes to quantities or dates once a WO was released to the shop floor. At that point in the production control cycle, a finite scheduling system would simply create a schedule for people and machines on the shop floor and publish a dispatch list — that is, a list of everything that needed to get done that day. Employees on the shop floor would then follow those instructions to the letter.

Seems simple enough, right? It actually wasn’t. In order to produce accurate schedules, finite scheduling systems required precise estimates for:

• Run and setup time — how much time each operation required
• Capacity — how much machine time and people were available each day
• Work completed — the actual work that needed to be completed at each step of the operation

If any one of these elements were less than 90% accurate, the dispatch list would be incorrect, WOs would be in jeopardy of being worked on at the wrong time, and the final part would be late to the customer.

In addition to the accuracy issue, there were other naturally occurring issues that affected scheduling: People would be unavailable for work, machines wouldn’t work quite right the day of, suppliers would deliver incorrect materials or deliver them late, and so on. Finite capacity scheduling systems were unable to anticipate these challenges and instead expected work on the dispatch list to be completed as instructed. This led to WOs starting and ending later than the times listed on the dispatch list, thus rendering it obsolete and requiring the manufacturer to rerun and republish schedules.

And that’s all without considering the underlying material coming from other WOs and PPOs. A bad schedule created bad dates for these materials, too. This would result in parts shortages in the middle of production, causing jobs to stop in the middle of the floor until materials were made or received. Unbelievably, very little has changed in the past 20 years.

Even today, less than 20% of manufacturers who own an ERP system actually use it to schedule their shop floor because ERP scheduling systems are often inaccurate and unable to keep up with changing conditions on the shop floors.

Which brings us to the 2000s: Without an effective shop floor scheduling system to rely on, manufacturers have turned to using spreadsheets, which has, in turn, led to expediting. It’s time for a change: In part two of our series, we’ll take a look at how a next-generation system should solve common challenges associated with shop floor scheduling and production control.