Simulation Modeling and Analysis of a Door Industry

Introduction

This is an application-based research article focused on production planning with the aid of simulation. As such, it contributes in bridging the gap between theory and practice of production planning. Theoretical planning problems are largely stylized with many unrealistic and/or “convenient” assumptions from the perspective of the system modeler/analyst. An important motivation for this paper is to provide an inside view of a real-world manufacturer and to showcase the particularities and increased complexity of realistic production environments. Another important aim for this paper is to present the application of powerful computational techniques, in this case stochastic simulation, to model and solve production planning problems using a scientific approach rather than intuition and empirical knowledge

In this paper we examine a door manufacturer which is located in northern Greece. The company follows a “hybrid” make-to-stock/make-to-order production model which complicates significantly the production process/planning and it is distinctly differentiated from stylized, theoretical models that can be found in the relevant scientific literature. The make-to-order model generally applies to clients that are large hotel firms who require highly customized products. The make-to-stock model generally applies to individual clients or architectural firms that carry out some construction project (especially a small one) and typically select doors from a range of standardized, product types that are readily available from the door manufacturer

It is of paramount importance for the company to meet the customer due dates, i.e. deliver the ordered products at the promised dates. Furthermore, it is also very important to maintain high productivity levels, i.e. make use of the available resources as efficiently as possible. These goals call for effective production planning and control of production operations. There are several unique features of this door manufacturer that cause production planning to be a daunting task and are not normally found in synthetic/artificial problems that can be found in the relevant literature.

First of all, there is a very strong seasonality component in the demand for finished goods. This is due to the fact that the majority of orders stems from hotel firms and the high season for hotels in Greece is normally between April and September. This means that the bulk of related orders is normally placed in the period October – May causing a notable peak in the workload imposed on the plant at that time. Secondly, the plant constructs a large number of customized doors which in turn means that a significant fraction of its products has never been built before. This makes it very hard for the production planner to forecast actual machining times, personnel requirements etc. and it also entails the related learning curves in respect to the workers who will be engaged in production. Finally, there are significant limitations regarding the available space in the plant, both in terms of storage space as well as available workstations etc. These constraints further complicate the task of production planning, since blocking of some production stages may occur especially in periods when there is a demand surge.

Motivated by the aforementioned problems, this research article makes the following contributions:

• •

  A systematic documentation and standardization of all production operations is carried out

• •

  A realistic simulation model of the plant is developed and verified using real-life data

• •

  A series of simulation experiments is conducted in order to study the effect of varying several production planning parameters on the production system’s performance

• •

  The goal of the authors is to aid the production planning function of the plant, which is largely empirical now, using state-of-the-art, computational methods.