Material Shortages Propagation: Using Network Science to Evaluate Inventory Efficacy

The ability to manage the effects of supply chain (SC) disruptions has played a major role in industrial environments in recent years. Pandemics, natural disasters, and political tensions increased difficulties in procuring components, resulting in delays in delivering finished products and thus in lost sales, along with other relevant effects that strained companies’ survivability (e.g. delayed financial flows). Companies have tried to avoid material shortages by increasing inventory levels, with increased costs and further risks (e.g., inventory devaluation, obsolescence), but with hard-to-evaluate benefits. Efficient resilience requires these cost-benefit analyses: their computation is even harder in today’s complex and intertwined supply networks (SN), where it is also hard to align SC players with contrasting goals. In this paper, we aim to understand the effect of increasing inventory levels in two different network types. Hence, we simulate the shortages of materials propagating through the SN. For the first time in this area of study, we extend the common SIR (susceptible, infectious, recovered) model to a SEIR (susceptible, exposed, infectious, recovered) model, where the state ‘E’ helps in modelling the time-dependent effect of different inventory levels to preserve operations. This paper’s practical value lies in offering a way for professionals to estimate the benefit coming from the investment in higher inventory levels, towards a choice that balances resilience and efficiency. On the theoretical side, we show that increasing inventory levels is more effective in SNs corresponding to scale-free network type (closer to automotive industry), compared to SNs corresponding to small-world type (typical of electronics industry).