Course Description and Content
The traditional optimization methods have certain limitations. For example, it is typical for them that they can only provide local optima. They also often presuppose such mathematical features as continuity and differentiability for their optimizing functions.
Our course will focus on novel cutting-edge optimization methods. They apply the methods used in evolutionary computing and neural nets, and these can find the global optima in optimization. They are also more robust concerning the mathematical restrictions and thus more applicable in this respect. In particular, we consider genetic, memetic and bacterial-memetic algorithms. We also consider the traditional optimization for the sake of comparison.
We apply these methods to economic model construction and simulations in computer environment. Hence, economical modeling is studied from the theoretical and practically on regression modeling and time series analysis standpoints. In this context, we also study cognitive map modeling which is a convenient technique for examining very complex economic phenomena.
The course focuses on evolutionary computing and traditional optimization methods. It also presents the basic principles of computational intelligence modeling for regression and time series analysis. These methods are applied to economic modeling examples. Corresponding traditional mathematical and statistical methods are also considered to some extent.
- highly specialised, novel, cutting-edge and multi-disciplinary knowledge on optimization in economic applications
- basic knowledge on using computational intelligence in economics
- specialised problem-solving skills for applying cutting-edge optimization and constructing computer models with computational intelligence in economics
- ability to manage and transform novel optimization and computational intelligence modeling to work or study in the contexts which are complex and require novel strategic approaches
For passing the course, a course report should be written (5-10 pages) in which an economic application based on the foregoing methods is presented. These application ideas will be considered and discussed with the participants, and their academic degrees are taken into account in this context. Eg. this application may deal with participant’s other studies or research work.
Adjunct. Prof. Vesa A. Niskanen, has a long international career within computational intelligence, esp. in studies of fuzzy, neuro-fuzzy and genetic-fuzzy systems as well as in cognitive maps. Prof. Laszlo Koczy has a distinguished career in computational intelligence, he is also one of the leading scholars in optimization with evolutionary computing.
The course materials and instructions for passing the course are given in Moodle and public websites which are especially designed for the course. Distant education studies with online services are also possible.
- 10 h independent work prior to course (this specified later)
- 120 h model construction and report writing
- 20 h of lectures
The course is held at the Faculty of Agriculture and Forestry in a computer class on Viikki campus by principally using Matlab software (maybe also with R or Excel), but previous knowledge of this is not required. Basic knowledge on mathematics is recommended.
Admission for NOVA courses is handled by the course organiser/ the NOVA member institution organising the course. Please see the links in the margin for more information.