Colloquium IS (April 6, 2018) – Feeding Evolutionary Algorithm with Column Generation output

Dear all,

We are pleased to invite you to our next Colloquium IS that will take place on Friday,

April 6, 2018, 12:30 – 13:30 (Paviljoen K.16).

Speaker: Murat Firat

Title: Feeding Evolutionary Algorithm with Column Generation output

Abstract:

Nowadays companies in telecommunication, logistics, and airport operations face large scale optimization problems.  These problems are usually scheduling and planning problems. Solving these problems to optimality can be exhaustive, even impossible due to the exponential size of their feasible solution sets. At this point, column generation comes to help by providing us good-quality lower bounds for the relaxations of Mixed Integer Linear Programming (MILP) formulations.  In my talk, I will firstly mention the basics of the Column Generation, and secondly explain how the output of Column Generation method can be used as input for an Evolutionary Method.

Best Paper Award for 2017

Qing Chuan Ye and Yingqian Zhang received a “Best Paper Award for 2017” from Omega-International Journal of Management Science.

The award-winning paper “Fair task allocation in transportation”, co-authored with Rommert Dekker, is available at https://doi.org/10.1016/j.omega.2016.05.005.

The paper studies a max-min fair, minimum-cost task allocation problem in transportation. It proposes an efficient algorithm for allocating tasks and resources in a fair way among players. The experiments show that often fairness comes with a very small price in terms of cost. In many cases such as applications in Sharing Economy, such a fair allocation algorithm is more socially desired than the classical cost minimization ones.

 

New Employee Reza Refaei Afshar

 

 

 

 

 

I received my Bachelor’s degree from Ferdowsi University of Mashhad in 2012 and then, my Master’s degree from University of Tehran in 2015. After that I worked for 2 years as a data scientists in Tehran.

Now, I am a PhD student at TU/e in Information Systems (IS) group. My research focus is on the programmatic advertising decision system project which involves collaboration with high tech industry. My research interests also include data mining, data analysis, machine learning and social networks analysis.

New Employee Paulo De Oliveira Da Costa

 

 

 

 

 

I obtained my Bachelor’s degree in Computational and Applied Mathematics in 2010 from the University of Campinas (Unicamp), with a specialisation in Operational Research. After my bachelor studies, I worked in Data Analytics roles for 4.5 years at two major companies in the banking sector, based in São Paulo, Brazil.

I then moved to Dublin, Ireland to pursue a Master’s degree in Business Analytics at the University College Dublin (UCD). After completing the programme in 2016, I extended my stay in Ireland working as Data Scientist.

At TU/e I will work within the Information Systems (IS) and the Operations, Planning, Accounting and Control (OPAC) groups as PhD student on the Real-time data-driven maintenance logistics project (WP1), which aims to leverage dynamic maintenance logistics policies supported by real-time data. The topic is focused on the integration of machine learning and optimisation models for more efficient and real-time decision making. I am excited to start working on the topic and looking forward to the next years of learning and collaboration.

Colloquium IS (March 2, 2018)

Dear all,

We are pleased to invite you to our next Colloquium IS that will take place on Friday,

March 2, 2018, 12:30 – 13:30 (Paviljoen K.16).

Speaker: Estefania Serral Asensio

Title: Integrating BPM and IoT

Abstract:

Many researchers have already shown the numerous benefits that Internet of Things (IoT) can bring to the Business Process Management (BPM) field. However, few companies have done the transition from traditional BPs to IoT BPs. They lack the needed guidelines and tools to embrace this challenging BP redesign.

This presentation will give an overview of my recent and ongoing research on supporting the integration of IoT in BPM and will also introduce CAPN, a tool-supported formalism that allows the modelling and simulation of IoT BPs. This tool enables a more accurate and reliable simulation of such processes improving the way organizations can evaluate their To-Be processes before they are put into production. The contribution of CAPN is two-fold: 1) it allows IoT BPs to be simulated and 2) it improves BPs simulation by enabling the consideration of all levels of context.

Colloquium IS (February 2, 2018)

Dear all,

We are pleased to invite you to our next Colloquium IS that will take place on Friday,

February 2, 2018, 12:30 – 13:30 (Paviljoen K.16).

Speaker: Maryam Razavian

Title: Empirical Research Design for Software Architecture Decision Making:
An Analysis

Abstract: Software architecture decision making involves humans, their behavioral issues and practice. As such, research on decision making needs to involve not only engineering but social science research methods. Despite past empirical research in software architecture decision making, we have not systematically studied how to perform such empirical research.

This talk is about the research methods have been used to study human decision making in software architecture. We analyzed research papers from our literature review on software architecture decision making. We classified the papers according to different facets of empirical research design like research logic, research purpose, research methodology and process. We derive lessons learned from existing studies and discuss open research issues inspired by social science research. We found predominant choices for the strategic research design and a variety of tactical design, operational design and study foci. We therefore introduce the focus matrix and the decision making research cycle to help researchers to position their research clearly. Thereby we provide a retrospective for the community and an entry point for new researchers to design empirical studies that embrace the human role in decision making.

 

Colloquium IS (November 3, 2017) – Interplay between learning and optimization

Dear all,

We are pleased to invite you to our next Colloquium IS that will take place on Friday,

November 3, 2017, 12:30 – 13:30 (Paviljoen K.16).

Speaker: Yingqian Zhang

Title: Interplay between learning and optimization

Abstract:

There are increasing interests in combining machine learning and optimization. In this talk, I will introduce two ongoing work in this research line.

In most existing approaches of using data to solve optimization problems, predictive (machine learning) models serve as decision variables, input parameters, or solution evaluation functions. In our work, we show how to use the internal structure of predictive models in optimization process, and demonstrate how the proposed approach helps to find better solutions.

Predictive models such as decision trees are typically built using sub-optimal algorithms, which often aim at optimizing loss functions (i.e., accuracy). These algorithms are not flexible when a different learning objective rather than accuracy is desired. We propose to transfer the decision tree learning problem to a mathematical optimization problem. In this way, different learning objectives, such as minimizing discrimination or false positive errors, can be easily specified for constructing optimal predictive models.

I will use online auction as an example to illustrate both approaches. 

 

Colloquium IS (October 6, 2017) – Leveraging linguistic modeling for clinical decision support

Dear all,

We are pleased to invite you to our next Colloquium IS that will take place on Friday,

October 6, 2017, 12:30 – 13:30 (Paviljoen K.16).

Speaker: Uzay Kaymak

Title: Leveraging linguistic modeling for clinical decision support

Abstract:
As electronic medical records become the norm for documenting medical history of patients, reuse of the data generated during the care process opens new ways of supporting clinical decisions. Advanced data analysis techniques, machine learning and data mining models that make secondary use of medical data are accepted more and more in clinical applications. Despite the advent of data-driven models, the practitioners find it important to have transparent models whose behavior can be understood well. In this respect, natural language is an effective means for communicating model behavior to the users. We argue that linguistic models based on fuzzy set theory form an excellent bridge between the data-driven modeling and the transparency required by the users in the clinical domain. In this presentation, we discuss data science challenges in the healthcare domain and consider several modeling approaches that use fuzzy set theory to develop models for supporting clinical decisions and improving the care process.

 

Micro to Macro in IS – at Which Level Should we Think ?

Dear all,

We are pleased to invite you to our next Colloquium IS that will take place on Friday,

September 1, 2017, 12:30 – 13:30 (Paviljoen K.16).

Speaker: Paul Grefen

Title: Micro to Macro in IS – At Which Level Should we Think?

Abstract:
IBM already contemplated the issue of aggregation levels (admittedly from a slightly different angle) a little while ago (actually, before some of us were born …): https://www.youtube.com/watch?v=0fKBhvDjuy0. This presentation applies this line of thinking to IS design (taking the world of MPMS as an example context) and intends to make you consider where your considerations are, or rather where your considerations should be.

Colloquium IS (July 7, 2017) – Compliance verification of inter-organizational workflows

Dear all,

We are pleased to invite you to our next Colloquium IS that will take place on Friday,

July 7, 2017, 12:30 – 13:30 (Paviljoen K.16).

Speaker: Pieter Kwantes

Title: Compliance verification of inter-organizational workflows

Abstract:
Inter-organizational workflows (IOW) are commonplace in industry. Often they are facilitated by an industry body for standardization of message formats, enabling electronic message exchanges between the participants of the IOW. The SWIFT organization, created in 1973 and responsible for maintaining message standards for the Financial services industry, is (to my knowledge) one of the first examples of such an industry body but there are many more (eg. HL7 for Healthcare, Rosetta in Electronics, EDSN for the Dutch energy market and so on). Making a design for an IOW and verifying its correctness is much more challenging than for an intra-organizational workflow and the stakes are higher. Ongoing research into this subject aims to find methods and tools to support this design and verification process. One question investigated in [1] is how to verify, using an automated procedure, that the design of a local workflow of an organization participating in an inter-organizational workflow is in compliance with the globally specified rules of that workflow. In this talk the results of [1] will be presented and future work will be discussed.

[1] Kwantes, P.M., Gorp, P.V., Kleijn, J., Rensink, A.: Towards compliance verification between global and local process models. In: Graph Transformation – 8th International Conference, ICGT 2015).

Bio:
Pieter Kwantes is a senior business consultant and project manager specialized in Investment Banking and holds a Master in Economics from the University of Amsterdam and a Master in Computer Science from the University of Leiden and is part-time Phd student also at the University of Leiden.