Prof. J. Rod Franklin, PhD

Abstract

Despite the increasing academic interest and financial support for the Physical Internet (PI), surprisingly little is known about its operationalization and implementation. In this paper, we suggest studying the PI on the basis of the Digital Internet (DI), which is a well‐established entity. We propose a conceptual framework for the PI network using the DI as a starting point, and find that the PI network not only needs to solve the reachability problem, that is, how to route an item from A to B, but also must confront a more complicated optimality problem, that is, how to dynamically optimize a set of additional logistics‐related metrics such as cost, emissions and time for a shipment. These last issues are less critical for the DI and handled using relatively simpler procedures. Based on our conceptual framework, we then propose a simple network model using graph theory to support the operationalization of the PI. The model covers the characteristics of the PI raised in the current literature and suggests future directions for further quantitative analyses.

Abstract

Increased volume, velocity, and variety of data provides new opportunities for businesses to take advantage of data science techniques, predictive analytics, and big data. However, firms are struggling to make use of their disjointed and unintegrated data streams. Despite this, academics with the analytic tools and training to pursue such research often face difficulty gaining access to corporate data. We explore the divergent goals of practitioners and academics and how the gap that exists between the communities can be overcome to derive mutual value from big data. We describe a practical roadmap for collaboration between academics and practitioners pursuing big data research. Then we detail a case example of how, by following this roadmap, researchers can provide insight to a firm on a specific supply chain problem while developing a replicable template for effective analysis of big data. In our case study, we demonstrate the value of effectively pairing management theory with big data exploration, describe unique challenges involved in big data research, and develop a novel and replicable hierarchical regression‐based process for analyzing big data.

Abstract

Wir stellen einen experimentellen Vergleich von Prognosetechniken für das Predictive Business Process Monitoring vor. Ausgehend von unseren Experimentergebnissen schlagen wir eine geeignete Kombination von Prognosetechniken vor.

Abstract

Predictive business process monitoring aims at forecasting potential problems during process execution before they occur so that these problems can be handled proactively. Several predictive monitoring techniques have been proposed in the past. However, so far those prediction techniques have been assessed only independently from each other, making it hard to reliably compare their applicability and accuracy. We empirically analyze and compare three main classes of predictive monitoring techniques, which are based on machine learning, constraint satisfaction, and Quality-of-Service (QoS) aggregation. Based on empirical evidence from an industrial case study in the area of transport and logistics, we assess those techniques with respect to five accuracy indicators. We further determine the dependency of accuracy on the point in time during process execution when a prediction is made in order to determine lead-times for accurate predictions. Our evidence suggests that, given a lead-time of half of the process duration, all predictive monitoring techniques consistently provide an accuracy of at least 70%. Yet, it also becomes evident that the techniques differ in terms of how accurately they may predict violations and nonviolations. To improve the prediction process, we thus exploit the characteristics of the individual techniques and propose their combination. Based on our case study data, evidence indicates that certain combinations of techniques may outperform individual techniques with respect to specific accuracy indicators. Combining constraint satisfaction with QoS aggregation, for instance, improves precision by 14%; combining machine learning with constraint satisfaction shows an improvement in recall by 23%.