Abstract
Multi-objective and Artificial Intelligence-enhanced optimization methods support the decision makers in finding solutions within trade-offs and conflicting goals, giving rise to a Cognitive Digital Twin framework, capable of simulating, predicting, and prescribing physical behaviors. The research activities on optimization methods in production systems have mainly focused on optimization of individual, single and convergent optimization goals, leaving disruptive external uncertainties to be investigated, such unpredictable climate or social events influencing food supply chains.
This conceptual paper reports the early outcomes of the “Decision Support System for the Life Cycle Optimization (DSS4LCO)” initiative, which aims at implementing a CDT architecture in food supply chains, able to handle multiple data-sources and conflicting goals under uncertainties, combining a DT framework, a lean, agile, resilient, and green (LARG) index and value stream mapping. Adopting the Design Science Research approach, the essay discusses the first 3 steps of the approach, aiming at identifying the research problem, contextualizing this issue within the existing knowledge base, and proposing a solution for the problem. Results discuss the definition of a CDT architecture, introducing the challenges to be faced in the future developments of the research initiative.