Can scheduling work outside the MES?
September 12, 2014
Production planning and scheduling have always been very complex processes, especially at the shop-floor level where the activities planned on a piece of paper or spread-sheets are actually executed. What makes scheduling for actual production complex is the fact that the planner or planning team actually scheduling jobs making many trade-offs and considering priorities, have very little or no insight of what is transpiring on the shop-floor and further have no way to control such events. Any discrepancy made in one operation can derail the entire schedule prepared by the planners due to the ‘knock-on’ effect and cause the plant significant damage. Previously we have discussed many features of the MES application and how it is the ideal system to plan, schedule, orchestrate and optimally execute a production process.
Today we will zero-in on the scheduling part of the MES and how it is imperative that the scheduling/planning function be directly embedded in the MES application, therefore enabling truly optimized planning and subsequent production, which provides excellence in both service and efficient execution.
MES applications from their very existence, have been applications which bridge the gap between the planning and the execution layer. They have a capability of capturing data from the shop-floor and provide actionable information to other applications such as the MRP or ERP in real time. MES applications are also capable of modeling a production process to its every specific detail, right from the total equipment base to the number of employees in a shift, to the total possible recipes in an operation and their possible execution routes. Let’s understand how the two features mentioned above make the MES the best possible vehicle to carry a scheduling module for any plant floor. Scheduling, as an activity, is far more complex than it is given credit for. A master schedule first considers all prime components to ensure smooth production right form material availability, to demand, to cycle times, to personnel availability and so on. This master schedule is then converted into an operational schedule which is more specific and routes specific orders through specific machine/personnel in a time-bound manner, to ensure final production and resultant cycle time is compliant with the master schedule. Now depending on the process, the schedule may be decided for various time frames, daily, weekly, monthly or more, what adds complexity. Here is that the schedule needs to be dynamic, considering actual and predicted demand, customer service priority, optimal production efficiency, production cost, material use optimization, change-over times. Also other variables which keep changing constantly, such as material availability, machine uptime, operator availability etc.
Now imagine a process which produces 5 different products, all of which can be customized. Each product involves hundreds of components, some being commonly used, others used specifically for a particular product. All products follow different production routes, but these routes intersect at times, meaning the same piece of equipment may be required by two or more products. Operators are skilled, but can’t be switched in operations, which means unavailability of personnel may cause down-time. With just the simple snap-shot of a modern process, it becomes clear how complex a scheduling activity might become if it’s executed manually through spread-sheets or even through a separate software application, which is capable of performing multiple complex trade-offs and optimizations. Also, it becomes an even more complex activity if the actual scheduling is not embedded in the application which is modeling the process, executing the planned activities and capturing each shop-floor transaction as it happens, which in most cases is the MES application.
For being even remotely effective, the scheduling activity needs to be proactive. It needs to equate demand and priorities with cost of production machine/manpower availability, cost of production and overall cycle time, that too in real-time. The entire scheduling operation might be jeopardized if a break-down which happened to a bottle-neck equipment/tool at 3 am in the night is captured and reported to the planner at 8 am next morning. This is where having an MES application capable of scheduling jobs becomes extremely important, as it perfectly models a manufacturing process and captures data as events happen on the floor. When a break-down occurs, the application alerts not only the planner/plant manager, but can provide contingent plans to the operators, which may range from getting the machine back up, to taking an alternate route, to switching the recipe or trying out another equipment and compromise on the spec (within tolerance levels of-course!!). A sort of dynamic schedule becomes possible when data from the shop-floor provides inputs for planning, as the planner now has access to the ERP/MRP data, which helps prioritize jobs in short-term, plan material and support customers. Also process data which provides status of all process bottlenecks, operator availability, current schedule compliance, deviation from schedules, in the form of Gantt charts, which are generated automatically and in real-time. The activity of scheduling and re-scheduling becomes very scientific and fact based when the MES is used of scheduling as, it is the same application which is responsible for executing the schedule, managing/training personnel, routing/re-routing jobs, ensuring asset availability, reporting KPIs and providing management reports based on actual performance.
Without the process rich data generated by the MES application, the planner or planning team, no matter how sophisticated their algorithms, calculations or software might be, will remain at best reactive and would definitely experience the knock-on effect sooner or later. There is a common yet flawed notion in many organizations that an Advanced Planning and Scheduling software may be used for production scheduling, as it has been equipped with many complex algorithms and optimizations. This may be used to make the service-efficiency trade-off and yet provide the best possible schedule allowing factories to meet standard lead times and yet insure quality. To this end, it becomes critical to consider the arguments made above, about MES and how it should carry the APS or operational scheduling module, simply because any advanced software without data would just provide it’s users garbage, since the input to the software is old/outdated process data, which is collected manually or through spread-sheets. By the time it is fed into the modern software with all possible logic, the actual constraints of the production process have changed on the actual production floor, the system which gets a garbage input will only give a garbage output. Modern automated plants dealing with thousands on inputs and many possible customizations to produce a plethora of diverse products using sophisticated multipurpose equipment, need scheduling to be real-time and be triggered by events as they happen, at both the planning and the production levels. This avoids unnecessary costs, shortages, down-times and to reach an optimal equilibrium between service and efficiency on a weekly, daily and hourly basis.
For such optimization to ever take place, the only real possibility is to have the production scheduling functionality embedded in a bigger software module such as the MES which may not only model the production process, but integrate the shop-floor to the top-floor which is the ERP/MRP. This ensures that the strategy, which was translated to a tactical schedule, may be properly executed through a real-time, automatically adjusting and dynamic operation schedule.
DOWNLOAD NEW WHITE PAPER:
IIoT Has a "Thing" for MES. Why IoT Platforms Won’t Replace MES for Industry 4.0
by Iyno Advisors and Critical Manufacturing