Toggle Main Menu Toggle Search

Open Access padlockePrints

Using a Genetic Algorithm layout design tool to optimise facility layout designs in the capital goods industry

Lookup NU author(s): Professor Christian Hicks


Full text for this publication is not currently held within this repository. Alternative links are provided below where available.


The layout of manufacturing facilities has a major impact of operational effectiveness. A good layout reduces material handling costs and lead time, whilst simultaneously increasing throughput. Capital goods companies supply products in low volume on an engineer-to-order or make-to-order basis. Typical products include power plant and oil rigs. Their main products are complex with many levels of assembly. The demand for products is cyclical and fluctuating, and therefore these companies usually have other businesses, such as spares and subcontract engineering that share common manufacturing resources. Hicks [1, 2] developed the Genetic Algorithm Layout Design Tool (GALDT) that generated layouts that minimised total direct or rectilinear distance travelled. This research extends the GALDT to include additional crossover operators, mutation operators, replacement mechanisms and placement algorithms. A case study used data obtained from a collaborating capital goods company to test the new tool and to evaluate the relative performance of various genetic operators and placement algorithms. The Modified Roulette Wheel (MRW) selection method developed in this research performed particularly well. A relationship between placement algorithms and the shape of the available space and resources was identified. The X-gyration and Y-gyration placement algorithms obtained good solutions and did not place restrictions on the shape of the available space. The Random Selection (RS) method combined the features of different selection operators and obtained an acceptable result. The best configuration of the GA was with the RS crossover operator, the two-operations random swap (2ORS) mutation operator, the Elitist replacement mechanism, and the Y-direction placement algorithm. The best settings for the GA parameters were: 250 generations, a population size of 200, a crossover rate of 90%, and a mutation rate of 18%. The GALDT generated layouts that reduced total rectilinear distance by up to 82.11% compared to the collaborating company's layout.

Publication metadata

Author(s): Lu HC, Hicks C

Publication type: Conference Proceedings (inc. Abstract)

Publication status: Published

Conference Name: The Second International Conference on Operations and Supply Chain Management

Year of Conference: 2008

Pages: 273-278

Publisher: The Chinese University of Hong Kong


Library holdings: Search Newcastle University Library for this item