| Summary: | 碩士 === 國立臺灣大學 === 商學研究所 === 89 === In recent years, due to the dynamic market environment, the quick response capability becomes a key determinant for an enterprise to fulfill customer orders effectively. In order to overcome the huge external variation, order splitting mechanism provides enterprises with a new way of thinking to improve their capacity of on-line order fulfillment.
In this thesis, two-phase dynamic order splitting and scheduling model are proposed. In the first phase, the optimal economic production interval decision model that applies the Theory of Constraint to building up a cost model on the bottleneck machine is developed. This model takes the bottleneck’s capacity as a constraint to find the optimal economic production interval and the max number of suborders. In the second phase, the dynamic suborder sizing and scheduling model is formulated. This model uses integer linear programming as a tool to formulate the dynamic scheduling problem. Four objective functions and eight types of constraints are considered in our model. The optimal suborder’s size and each suborder’s plan start time and completion time are determined by this OR model according to different objectives. After two-phase dynamic order splitting and scheduling model is constructed, the performance of our model is compared with several other business practices. In order to evaluate our performance, several performance indices and scenarios are designed and followed by ANOVA statistic testing.
From our testing results, in terms of the fixed order splitting model, original order due date setting performs better than other suborder due date setting methods. However, in terms of the dynamic order splitting model, suborder due date setting is secondary and has no significant impact on the performance. Among all the order splitting models, dynamic order splitting model always performs better than others. The selection of scheduling objective should coincide with the chosen of performance indices. Besides, dynamic order splitting model also deliveries better performance on most of the performance indices. Hence, this model is a better order splitting model for enterprises to reduce total production cost and enhance the customer satisfaction level.
|
|---|
