| As far as implementation issues are concerned, probably the top problems are setting up the detailed rules and objectives correctly. Most APS systems require a model of the production facility -- which needs to be very detailed for proper production scheduling when there are variable changeover conditions, sequencing issues, or high mix. Whether planning or scheduling, setting the weighting of the objectives to which the engine will optimize is not always obvious. The tradeoffs between customer on-time delivery, total production cost, and profitability or plant utilization are tricky to determine in advance of trying to gauge results. As such, once implemented, the model and rules need to be kept up to date. The conditions where APS has been implemented vary widely in a variety of plants and industries -- with successes in each. There have also been less than successful implementations or projects that have never been fully implemented in each industry. Some of the systems are relatively complex and need one or a small group of highly trained people doing the planning & scheduling. The rest of the plant must be willing to work to the schedule when they can and feed back issues when they cannot. The direction that APS is heading is pretty much tied to the vendor's market strategy. Some of the top SCM vendors have focused on building a suite of systems for alerting and monitoring as well as outside of production capabilities, while the more detailed scheduling providers are working to integrate more closely into execution & production performance dashboards. ERP providers are focused on improving the algorithms and the usability of their systems. As companies move to more pure lean manufacturing practices, APS is often used in the feeder lines, with Kanban triggers from a final assembly line serving as the incoming order stream. |