by Vern Puchalski, on November 14, 2018
If you ask 10 people what “dispatching” means to them, you’ll typically get 10 wildly different and sometimes conflicting answers. These are usually filled with technical jargon and reflect some key facet of the individual’s function in the factory (such as “finish my experiment quickly”, “get this lot to the customer on time”, or a whole range of hot-button themes).
To get to the heart of what dispatching really is, consider this definition of dispatching:
Regardless of any specific definition, dispatching is the art of answering the question: “What is the next-best lot(s) to run right now on this specific equipment?”. This includes a host of standard dispatching priority rules but also typically includes the individual operator’s idea of how to manage WIP, efficiently run key equipment and balance a host of post-it style can/can-not runnable rules.
Equipment Specific: The view of “next best” must be associated with the specific equipment that will be used for production.
Real Time: The view is dynamic and must always reflect the current reality. You can’t rely on any plan that is not eternally up-to-date and isn’t reflective of current conditions.
Equipment/Process Capability Aware: The operator needs to see whether the lot is capable of running on this equipment at this moment in a predetermined view. It’s not only unsafe to rely on endless engineering change notices (ECNs) and tribal knowledge that describe the set-up and run conditions necessary for each recipe but it also wastes considerable time. Bottleneck capacity and labor time lost refreshing lists, evaluating capability, etc. can cost millions. Any dispatching system that is unaware of the real-world requirements to run each individual lot is useless!
Evaluates Complex, Conflicting Priorities: It’s often necessary to balance a complex set of individual rules (run the lot with the highest critical ratio unless this lot is not of the same setup as the last lot and unless the downstream bottleneck inventory is less than minimum, etc.). A dispatching system must allow the creation of simple, atomic rules (i.e. “Does this lot require a set-up change?”) and the ability to bundle these simple rules into complex conditions.
Imagine the possibilities such systems allow. Priority rules can be pre-determined and associated with instantaneous WIP and equipment conditions – eliminating the frequent calls to “take lot X off the tool and instead run lot Y”.
Lots are pushed to bottleneck tools keeping them “fed” and running as efficiently as possible. Complex “process capabilities” are pre-evaluated, eliminating common misprocessing errors and tool wait time while an operator searches ECNs. Changing equipment and WIP conditions are automatically handled in real time. Priorities and conditions are formalized in codes and conditions so the factory runs similarly across all shifts and dates.
Such systems can ultimately be fed factory economic data and help minimize cost by minimizing chemical or electricity use, prioritizing lots by customer value, average profit by part number or any other site-critical value. The key is to feed create formal rules, enforce runnable process capabilities and allow the dispatching system access to all relevant system data.
Written by Vern Puchalski
Vern joined SYSTEMA in 2007 with 20 years’ MES and automation experience in silicon wafer and chip manufacturing. He has been involved in numerous projects concentrating on automation-friendly MES implementations and extensions, dispatching, recipe managers and process-capability solutions.
Never miss a blog post again!