| Scheduling in an Electronics Manufacturing Service (EMS) environment is one of the most complex tasks encountered by a capacity planner and scheduler. Confronted with a high technology product market, Printed Circuit Board (PCB) manufacturing is becoming increasingly more dynamic and competitive with the introduction of new products in shorter time intervals. In a typical EMS environment, the scheduling problem involves a set of jobs or production orders to be completed, where each production order has a set of operations to be performed. It has been observed that schedules are influenced by factors such as job priorities, release dates, due dates, cost restrictions, production levels, machine and resource capabilities and availabilities. Additionally, the performance criteria involve trade-off analysis between holding costs of the inventory, frequent production changeovers and customer satisfaction.; Due to the aforementioned challenges, enterprises have proceeded to the adoption of Enterprise Resource Planning (ERP) systems. ERP systems are software packages that enable the integration of operations, business policies and functions. However, the majority, if not all, of these systems do not support the production scheduling process, which is a critical aspect of decision making in today's manufacturing and service enterprises. There are a few reasons for the inability of ERP packages to provide production scheduling solutions. First, dynamic production scheduling problems are typically not amenable to solution using optimization theory. Second, dynamic uncertainties, such as machine breakdowns, tool failures, order cancellation, and stochastic demands make the problem more complex. Moreover, there could be more than one feasible production schedule for a specific manufacturing scenario. These constraints and restrictions make it very hard to build production scheduling modules in an information system.; Numerous approaches to incorporate the production scheduling concepts in an ERP system have been well documented in the literature. Approaches/strategies use artificial intelligence (neural networks and fuzzy rule based systems) and genetic algorithms. However, most of them are essentially "stand-alone" systems that cater to specific categories of production scheduling problems. However, in a dynamic domain, such as an EMS provider, the need for real-time decision making is also critical.; This thesis proposes a framework to integrate real-time production data, scheduling techniques and simulation that can provide realistic scheduling policies that can be used for tactical and strategic decision-making. Heuristics will be used to decide the 'best' scheduling policy, which will be subsequently validated by discrete event simulation. It will also be used to model and analyze various 'what-if' scenarios and provide 'optimal' schedules. The 'optimal' schedules will be then communicated to the ERP system for updating its information. The MRP and production schedules will be updated as required, based on this new information. The framework will also incorporate solutions to deal with disturbances and variability on the shop floor.; This research effort also proposes an approach to design and develop a decision support system that can be integrated with an ERP system. The system would help the master scheduler to decide the best scheduling policy for any product mix for future use. In the event of a new product mix, the scheduling module will generate a new dispatch list and the simulation module will be triggered to determine 'best' the scheduling policy. Therefore, the system that is proposed will assist both real-time operational planning and strategic scheduling planning. |
