Dynamic MCDM with future knowledge for supplier selection

Dynamic multi-criteria decision making (DMCDM) is an emerging subject in the decision-making area and in the last decade the challenge to consider time as an important variable has become important. Some frameworks already exist in this area but when compared with other types of decision-making models, DMCDM needs more work to be applicable in real industrial problems. In this work we extend a dynamic spatial-temporal framework, designed to deal with historical data (feedback), to address the problem of considering future information/knowledge (feed-forward). The main objective is to enrich dynamic decision-making models with explicit knowledge (existing historical data) and tacit knowledge (e.g. expert predictions) in time-evolving problems, such as supplier selection. Considering supplier-predicted information for future situations (e.g. investments in capacity) and, simultaneously, learning from historical data can help a company to find less risky and consistent alternatives. The proposed model is successfully implemented in a real case study for supplier selection in one automotive industry to demonstrate the capability and applicability of the model.