Agent-Based Simulation for Indoor Manufacturing Environments---Evaluating the Effects of Spatialization

The paper elaborates on an Agent-based Modeling approach for an indoor manufacturing environment---in particular, a semiconductor production plant. In order to maintain a flexible production ``line'', there is no conveyor belt, and a mix of different products is present in the indoor environment. With the integration of Industry 4.0 or Smart Manufacturing principles, production assets may be transported by autonomous robots in the near future. The optimization of manufacturing processes is challenging and computationally hard. Thus, simulation methods are used to optimize manufacturing plants and the processes. In contemporary literature, the effects of the spatial dimension with respect to the simulation of manufacturing processes is neglected. In this paper, we evaluate on the effects the spatial dimension in an Agent-based Model for indoor manufacturing environments. The Agent-based Model developed in this paper is utilized to simulate a manufacturing environment with the help of an artificial indoor space and a set of test data. Four simulation scenarios---with varying levels of spatial data usable---have been tested using Repast Simphony framework. The results reveal that different levels of available spatial information have an influence on the simulation results of indoor manufacturing environments and processes. First, the distances moved by the worker agents can be significantly reduced and the unproductive movements of worker agents (without production assets) can be decreased.