This paper presents a simulation-based optimal design and analysis method for designing a train overhaul maintenance facility. Because the train is composed of one or more cars, the authors design the simulation model after analysing the operation of the train according to the train's following parts: train, car, car body, and wheel. The authors consider the critical (dependent) factors and design (independent) parameters for the simulation analysis. Therefore, multi-criteria decision making (MCDM) is proposed for selecting alternatives, and simulation optimization is used for finding the optimal design parameters of the selected alternative. A case study for the above approach is applied to the electric locomotive overhaul maintenance facility. This paper provides a comprehensive framework for a train overhaul maintenance facility design using MCDM and the simulation optimization.

In a rapid prototyping process, the time required to build multiple prototype parts can be reduced by building several parts simultaneously in a work volume. Interactive arrangement of the multiple parts, called three-dimensional nesting, is a tedious process and does not guarantee the optimal placement of all the parts. The three-dimensional nesting is well known as a problem requiring intense computation. Thus, an ecient algorithm to solve this problem is still under investigation. This paper presumes that the three-dimensional packing problem can be simpli®ed into a set of two-dimensional irregular polygon nesting problems for each layer to take advantage of the characteristic of a rapid prototyping process, i.e. the process eventually uses two-dimensional slicing data of the STL ®le. The proposed algorithm uses a no-®t polygon (NFP) to calculate the allowable locations of each slice of a part such that it does not overlap other existing slices in the same z level. Then the allowable position of the part with respect to other parts already located in a work volume can be determined by obtaining the union of all NFPs that are obtained from each slice of the part. Additionally, a genetic algorithm is used to try and determine the various orders of the placement of the part and the various orientations of each part for ecient packing. Various orientations of a part are examined by rotating it about the normal vector of the slice in ®nite angles and by inversion. The proposed algorithm can be applied to a rapid prototyping process that does not use support structures.