< back

Optimization criteria is the brain of APS

October 23, 2014

Optimization criteria is the brain of APS

Over the past few weeks, we have been exploring the implications of deploying an APS or Advanced Production Scheduling application embedded in the MES. This combination of applications automates scheduling and provides planners and decision makers a truly dynamic and real-time schedule, which is aligned with the company’s strategy. We have seen how experienced planners can use the scheduling functionality as a tool to improve the process performance and thereby contribute to both business retention and expansion. In the era of mass customization and truly global competition, it is imperative to be able to change the production schedule based on external stimuli in real-time, or proactively even.

Today however, our main focus would be on the logic behind these high-tech intelligent and intuitive scheduling applications, which have the capability of saving millions of dollars for your plant if used correctly. In one our recent posts we discussed scheduling as an activity, where we saw the significance of having an optimization criterion or a set of them to decide how the scheduling will be performed at the shop-floor level.

The optimization criteria may be considered the brain of a scheduling application. It is this very criteria which provides the logic to the application to make internal trade-offs, prioritize/delay jobs, change routes or re-schedule a particular order. Planners working in plant operations are very familiar with obtaining/deriving and setting these optimizations, as they prioritize orders and release them for production. For those less familiar with this activity, there might be a few questions they would like answered, such as- Who provides these criteria? What factors influence them? Is there a standard set of criteria, are they static in nature, what are the most common optimization criteria?

Let’s try and address these questions one by one. Firstly, the optimization criteria may hugely depend on the strategic outlook of the company itself, whether it’s expanding volumes/lines, focusing on a target/niche market, giving priority to delivery or order volumes, etc. From a strategic perspective, there may be a single criterion or multiple high level optimization criteria, such as reduction of overall delivery time or maximization of energy utilization. It is also possible that the strategy of the company keeps changing and so would these high level deliverables. Then at a more tactical level, another aspect is the type of process, whether it’s demand driven or production driven, or simply put, is it a pull or a push operation. Few other factors which influence the choice of the optimization criteria are the number of products and product related services being offered and the importance of delivery commitments. As well as what volumes are being produced, the size and span of the customer base, special commitments and so on. All these factors and their mutual interplay might call for a particular optimization criterion or a sort of weighted combination of multiple criteria.

Once the high level criteria is decided, it may be provided to serve as the logic of the scheduling exercise. The planners may be provided the authority to change the criteria combination or the weights of various criteria, depending on the current business situation. They may also consider both business and production aspects, which might include a variety of factors ranging from demand to machine availability. It is this very ability of being able to change the optimization criteria depending on current situations and future predictions, which make automated scheduling both intelligent and valuable. Thereby ensuring that the production activity is always aligned with the strategy and can adjust to contingent situations in real-time. Aside from the factors influencing the optimization criteria, it is also important to have a basic understating about the types of optimization criteria.

Common optimization criteria range from, minimizing process waiting time to maximizing the turnover. Some of the most common ones are as follows (these are, by no means, the only ones):

  1. Minimizing Setup Times
  2. Minimize Delivery Delay Times
  3. Maximize Machine Load
  4. Minimize Biggest Delivery Delay
  5. Minimize Delivery Date Deviations
  6. Maximize Deliver Fulfillment
  7. Maximize Energy Efficiency
  8. Minimize Number of Late Deliveries
  9. Minimize Quantity Weighted Cycle Time
  10. Minimize Priority and Quantity Weighted Cycle Time
  11. Minimize Priority Weighted Delivery Delays.
A planner might choose to use a specific criterion or a combination of criteria based on their weighted contribution to the overall schedule of the production process. Once provided, the optimization criteria acts as the internal logic for the scheduling activity, which coupled with the real-time process data collected from the MES allows the application to provide best possible schedule for an operation.

Having discussed the nitty-gritty of the optimization criteria, it’s critical at this juncture to highlight that no matter how sophisticated the technology or how experienced the planner, it is vital that the optimization criteria be defined and then subsequently refined continually as business priorities and process parameters change. Only then will the scheduling activity of the shop-floor be truly ‘Optimized’.

< back


IIoT Has a "Thing" for MES. Why IoT Platforms Won’t Replace MES for Industry 4.0
by Iyno Advisors and Critical Manufacturing