BOTTLENECK ORIENTED CAPACITY PLANNING THROUGH ERP SYSTEMS AT AN EMS PROVIDER
Authors: S. Ramakrishnan and K. Srihari Company: Department of Systems Science & Industrial Enginee Date Published: 1/17/2006
Pan Pacific Symposium
Abstract: Customer oriented manufacturing and production flexibility are ‘critical success factors’ in today’s electronics manufacturing domain. This calls for a new form of manufacturing systems organization and control. The prime foci of such manufacturing organizations are adherence to the due date and the reduction of cycle time for lower capital commitment. At an Electronics Manufacturing Service (EMS) provider, the determination of manufacturing plans needed to manufacture forecasted demands is complicated due to the sensitivity of the product mix, uncertainty of demands, lead times, and tool costs. The product demands in an EMS environment are highly volatile, resulting in difficulty to predict the demand profile over several months. This could result in a large gap between the planned and needed capacity, when actual demand is materialized. The current turbulent market trends call for precise and fast predictions of both manufacturing planning and scheduling. This is a critical aspect considering the dynamic nature of plant conditions. Enterprise Resource Planning (ERP) vendors have been providing planning solutions that often do not create realistic solutions for such situations due to a variety of reasons. Hence, it is imperative for organizations to be aware of the plant conditions for effective planning. Linear Programming (LP) or Discrete Event Simulation (DES) could be used as a proactive tool to identify these opportunities. This research endeavor focused on integrating ERP systems with linear programming models for effective capacity planning, using the concepts of Line Requirements Planning (LRP). A framework that will retrieve the plant conditions through the ERP system is developed. This framework feeds a linear programming model and could be used for robust capacity planning under demand uncertainty and underutilized bottleneck resources. Given a demand scenario, the proposed model would generate the ‘best’ product mix that maximizes the utilization of the bottleneck resources, increasing the overall revenue. Multiple demand scenarios are considered along with the resource availability of the process centers, especially the system bottlenecks. The determination of the product mix is also dependent on the Pricing Index (PI), a metric used to assign priority for a product. The best product mix that is obtained from this model is the basis for scheduling the orders. Hence, the proposed methodology would effectively release orders to minimize system nervousness and inventory. The proposed methodology would also help in understanding the effect of LRP as opposed to Materials Requirement Planning (MRP). Inadequate consideration of bottlenecks in most MRP and ERP systems needs to be addressed. Most ERP vendors have bottleneck-oriented procedures for production control. However, it is very expensive to purchase and customize these modules. The proposed architecture would provide a cost-effective solution that is easy to implement while concurrently catering to the specific needs of an EMS production environment.