Decision-making using agent-based modeling - A case study of complexity at Unilever Bestfoods

Abstract: Problem: The industry structure could be viewed as changing. A problem for some mangers is the knowing how the consequences of their decisions affect their supply chain in advance. Purpose: The purpose is divided into two parts. (1) To examine the possibility of applying the complexity theory through an agent-based model at a supply chain. (2) To investigate and evaluate different future scenarios at Unilever Bestfoods Nordic's supply chain through an agent-based model. Method: A case study method was chosen and the data has been collected through interviews, observations and document studies. The data was finally analyzed through a custom-made agent-based computer model. Conclusions: The parts of the supply chain studied at Unilever Bestfoods Nordic’s (UBFN) displays internal properties in the same way as heterogeneous agents exist in a complex adaptive system (CAS). These agents all have different rules guiding them as well as some degree of freedom, or dimensionality, which enable them to self-organize. When the agents struggle to improve their own fitness they at the same time change the conditions for other agents. The authors believe that UBFN can certainly be viewed as a CAS. The computer model was verified through historical data to make sure that it was valid for answering what-if questions. Manipulating input parameters and observing the outcome has strongly convinced the authors that the constructed ABM is suitable as a basis for decision. The results of the simulations, in several cases both novel and interesting, were well received at UBFN at lead to further inquiries about the future and discussions about change at UBFN.