Exploiting Big Data in Logistics Risk Assessment via Bayesian Nonparametrics

摘要

In cargo logistics, a key performance measure is transport risk, defined asthe deviation of the actual arrival time from the planned arrival time. Neitherearliness nor tardiness is desirable for customer and freight forwarders. Inthis paper, we investigate ways to assess and forecast transport risks using ahalf-year of air cargo data, provided by a leading forwarder on 1336 routesserved by 20 airlines. Interestingly, our preliminary data analysis shows astrong multimodal feature in the transport risks, driven by unobserved events,such as cargo missing flights. To accommodate this feature, we introduce aBayesian nonparametric model -- the probit stick-breaking process (PSBP)mixture model -- for flexible estimation of the conditional (i.e.,state-dependent) density function of transport risk. We demonstrate that usingsimpler methods, such as OLS linear regression, can lead to misleadinginferences. Our model provides a tool for the forwarder to offer customizedprice and service quotes. It can also generate baseline airline performance toenable fair supplier evaluation. Furthermore, the method allows us to separaterecurrent risks from disruption risks. This is important, because hedgingstrategies for these two kinds of risks are often drastically different.