Bayesian network construction from event log for lateness analysis in port logistics

• We use Bayesian network to analyze lateness in port logistics.
• We detect loops from dependency graph and decompose it into Bayesian networks.
• We add attribute of event as a state in the Bayesian network.
• The structure is verified using chi-square and local Markov property analysis.
• The model shows the probability of lateness and the influence factor of the activities.

The handling of containers in port logistics consists of several activities, such as discharging, loading, gate-in and gate-out, among others. These activities are carried out using various equipment including quay cranes, yard cranes, trucks, and other related machinery. The high inter-dependency among activities and equipment on various factors often puts successive activities off schedule in real-time, leading to undesirable activity down time and the delay of activities. A late container process, in other words, can negatively affect the scheduling of the following ones. The purpose of the study is to analyze the lateness probability using a Bayesian network by considering various factors in container handling. We propose a method to generate a Bayesian network from a process model which can be discovered from event logs in port information systems. In the network, we can infer the activities’ lateness probabilities and, sequentially, provide to port managers recommendations for improving existing activities.