The present study aims to propose a constraint aware hybrid automated layout design program-computerised relative allocation of facilities technique (ALDEP-CRAFT) framework for multi-floor facility layout optimisation using data from a multi-floor warehouse. By using data-based relationship (REL) modelling approach in ALDEP and implementing CRAFT algorithm with multiple possible swapping, the present framework solves the limitations of expert subjectivity and reliance on over simplified centroid swapping present in traditional construction and improvement algorithm.
The study includes a data-based REL modelling approach that learns closeness relationships from empirical material flow data. A multi-start ALDEP construction helps achieve different feasible layout configurations in a variety of geometric and operational constraints. CRAFT algorithm further improves layout quality with swap, relocation and rotation neighbourhood operators. The framework is evaluated in the real-world warehouse layout problem with multi-floor material movement, fixed location and different department sizes.
The results show significant improvement in transportation costs. The improved CRAFT phase reduced transportation cost from 5438 to 4000.2 units, a 26.4% improvement over the ALDEP layout. A best-performing layout with a transportation cost of 3131.9 units was discovered in batch experiments (42.4%). The framework is verified for using several independent runs that obtained a significant reduction in transportation costs (t = 13.92 and p < 0.001) and a large practical effect size (Cohen’s d = 3.11).
The framework is scalable, objective and computationally efficient. It can be replicated for optimising the layout of a warehouse and manufacturing facilities.
