Semi-Automatic Generation of a Fuzzy Inference System in a Reshoring Context

Abstract

Reshoring can be regarded as offshoring in reverse. While offshoring mainly has been driven by cost aspects, reshoring considers multiple aspects, such as higher quality demands, faster product delivery and product mass-customization. Where to locate manufacturing is usually a purely manual activity that relies on relocation experts, hence, an automated decision-support system would be extremely useful. This paper presents a decision-support system for reshoring decision-making building a fuzzy inference system. The construction and functionality of the fuzzy inference system is briefly outlined and evaluated within a high-cost environment considering six specific reshoring decision criteria, namely cost, quality, time, flexibility, innovation and sustainability. A challenge in fuzzy logic relates to the construction of the so called fuzzy inference rules. In the relocation domain, fuzzy inference rules represent the knowledge and competence of relocation experts and are usually generated manually by the same experts. This paper presents a solution where fuzzy inference rules are automatically generated applying one hundred reshoring scenarios as input data. Another important aspect in fuzzy logic relates to the membership functions. These are mostly manually defined but, in this paper, a semi-automatic approach is presented. The reshoring decision recommendations produced by the semi-automatically configured fuzzy inference system are shown to be as accurate as those of a manually configured fuzzy inference system.