Optimization of a furniture factory layout

Abstract

Abstract. This paper deals with the problem of optimizing a factory floor layout in a Slovenian furniture factory. First, the current state of the manufacturing system is analyzed by constructing a discrete event simulation (DES) model that reflects the manufacturing processes. The company produces over 10,000 different products, and their manufacturing processes include approximately 30,000 subprocesses. Therefore, manually constructing a model to include every subprocess is not feasible. To overcome this problem, a method for automated model construction was developed to construct a DES model based on a selection of manufacturing orders and relevant subprocesses. The obtained simulation model provided insight into the manufacturing processes and enable easy modification of model parameters for optimizing the manufacturing processes.Finally, the optimization problem was solved: the total distance the products had to traverse between machines was minimized by devising an optimal machine layout. With the introduction of certain simplifications, the problem was best described as a quadratic assignment problem. A novel heuristic method based on force-directed graph drawing algorithms was developed. Optimizing the floor layout resulted in a significant reduction of total travel distance for the products.

Movement Analytics: Current Status, Application to Manufacturing, and Future Prospects from an AI Perspective

“…The factory layout problem concerns the spatial positioning of work stations, machines, tools, and/or functional areas in order to increase the efficiency of the production flow while respecting the application specific constraints, e.g., building boundaries and pathway dimensions. There is a significant amount work on optimizing the layout by simulation of production schedules on a virtual digital factory environment on simulated data (see, e.g., Centobelli et al [49], Lee et al [135], Kanduč and Rodič [121], Herr et al [114]), which can be partly based on real-time data and/or movement analytics on the current factory setup. Different objectives or combination of have been studied, e.g., minimum travel time/distance (see, e.g., Dzeng et al [84], Kanduč and Rodič [121], Herr et al [114]) and minimum production lead time (see, e.g., Centobelli et al [49]).…”