The Single-Row Facility Layout Problem with chance constraints

Abstract

In the Single-Row Facility Layout Problem (SRFLP) one has given a set of departments, their lengths and pairwise material flow between the departments. The goal is to find an arrangement of the departments on a single row that minimises the weighted centre-to-centre distances between the departments. Although the majority of the literature assumes that the data are known beforehand, in real-life applications, the material flow may be an uncertain parameter. In order to address real-life requirements, we assume that the material flow is uncertain and follows a multivariate normal distribution. By combining the concept of chance constraints with a well-known model from the literature for the SRFLP, we develop an approach that is easy to implement and does not require specialised software. Furthermore, we generate a comprehensive library of test instances for our problem setting by combining instances from the literature with different variance–covariance matrices. In a computational study we compare our chance-constrained model with the standard model from the literature and explore the limits of our approach.