Welcome to the COMISEF-Wiki

What is COMISEF? COMISEF is a Research and Training Network, co-ordinated by Peter Winker, University of Giessen. The scope of quantitative research in economics and finance broadens due to the increasing availability of data sets. In order to extract relevant information from these data new statistical procedures are developed often resulting in highly complex optimization problems. Novel optimization procedures are required to tackle these problems without imposing unrealistic assumptions. COMISEF will contribute to the development of these methods, their statistical analysis and their application to problems in statistics, econometrics and finance by concentrating on heuristic optimization methods. In this context, COMISEF will conduct individual and joint research and training activities exploiting the specific expertise of all partners. For more information, please visit the COMISEF homepage. External Articles Heuristic optimisation in economics and econometrics. A recent nontechnical account of the research programme. Finding sense in finance. A profile of Dietmar Maringer, one of the founding researchers of COMISEF.

Objectives of the COMISEF Wiki Provide practical guidance into implementing solutions for optimisation problems (in particular heuristic methods) and related issues. To this end, we hope that we can create a quality, highly professional knowledge base that is actively used by a researchers and practitioners alike. Enable dynamic knowledge sharing amongst COMISEF members and provide an area for members to communicate the scope and nature of their research. Ensure the theorectical correctness of content by an editing and peer review process. A logical thematic structure to the way internal pages are linked together, so that a user can easily navigate between problem description, the theorectical ideas behind a problem, and specific problem resolution details.

COMISEF Wiki Feedback At present all content in the Wiki is generated by COMISEF members. However, we actively encourage those external to the project to contact us on any aspect of the Wiki. We are especially keen to learn if any ideas described in the Wiki have been used as a source of reference, or directly and successfully applied to a problem. All feedback is gratefully received. Please click ‘Contact’ on the Topics bar to the left of this page. Alternately, Click here>>. On this page there are instructions on how to contact us here at COMISEF.

COMISEF News For all archive news, Click Here>>. August 2009 Articles Chris Sharpe at University of Giessen has written a piece on Optimised U-Type Designs on Flexible Regions. This is in the field of Experimental Design and applies a new discrepancy measure for differently shaped regions, and implements Threshold Accepting to optimise several design configurations. New Tutorials Enrico Schumann at Université de Genève written a detailed tutorial on Robust regression with Particle Swarm Optimisation in R with the two examples with code for robust regression, namely LTS and LMS. Included are steps to test that any implementation works correctly. New Tip Enrico Schumann has added a code for implementing Vectorised computation of sums of squares in R and Matlab. July 2009 General Wiki Updates: We are now replacing the Tab View at the bottom of content pages with a simple table for reference information. This enables all the information to be shown when printed out. New Member Profiles: The COMISEF member profiles are now complete with the inclusion of Mateusz Gątkowski at CCFEA, University of Essex, UK. He is involved in the field of Agent-Based Modelling applied to Economics and Finance and his main research work is in Contagion in Finance, and Complex Adaptive Systems. Also added this month is the profile of COMISEF fellow Dmitri Blueschke, a researcher based at Klagenfurt University, Austria. The focus of his work is in dynamic game theory; more specifically dynamic tracking games, its theory, and application to real economic models. New Tutorials Enrico Schumann at Université de Genève added two tutorials related to portfolio optimisation, providing implementation solutions in 'R': one is for Tangency Portfolio computation (ie, maximising the Sharpe ratio); and the other demonstrates how to compute the Global Minimum-Variance Portfolio. New Tips Chris Sharpe at University of Giessen provides example Matlab code to render a 3D objective function as a nice, clear monochrome graph.