Optimization of Information Acquisition for Decision-Intensive Processes
Simon Voorberg defended his PhD thesis at the department of Industrial Engineering and Innovation Sciences on May 9th.
When we look at the current economy and its goals, we can say that the modern labour force is more and more shifting in the direction of knowledge-intensive and decision-intensive jobs. Such jobs are based on the capabilities of workers to make decisions based on collected knowledge or information. Decision-intensive business processes consist of a phase where information is collected before a final decision is taken in the subsequent phase.
In his research Simon Voorberg focusses on dynamically optimizing how much information to collect in which order (Phase I), followed by optimizing the decision (Phase II).
Model the complex behaviour
These two phases are dependent on each other. Collecting more information in Phase I is costly and asks for efficiency. The challenge is to choose which information to collect and, based on that new information, how to continue. Also, extra information allows to make a better decision in Phase II to improve the effectiveness. That is why we build models that can model the complex behaviour of the decision-intensive process and optimize such processes.
These models are built, beginning by turning any practical decision-intensive process into a defined business process model named Case Management Modeling Notation. The result is an approach that returns what we define as an optimizable decision-intensive process (ODIP). Subsequently, this ODIP is a correct input for the optimization using Markov Decision Process techniques. We show both the feasibility of the whole approach and the efficiency of the methods of optimization. Moreover, we study different directions in which we extend this basic approach.
Title of PhD thesis: 'Optimization of Information Acquisition for Decision-Intensive Processes.' Supervisors: Geert-Jan van Houtum, Willem van Jaarsveld and Rik Eshuis.