Solving Complex Optimization Problems with Bacterial Memetic Algorithms

Dr. Janos Botzheim, a Hungarian researcher at the Department of Automation (Szechenyi Istvan University, Györ), was invited by the Fuzzy Logic Laboratorium Linz (FLLL) and the Software Competence Center Hagenberg (SCCH) to give a talk about the topic „Bacterial Memetic Algorithms. Introduction and Applications“.

Nowadays, various forms of genetic algorithms enjoy a great attraction for solving complex optimization problems in industrial processes and (quality) control systems for increasing product quality while minimizing the costs.
This is simply because the optimization problems are often too complex to be solvable with classical analytical techniques, especially when constraints (arising due to limitations of the operating system) appear in the optimization problem. Genetic algorithms are leaned on Darwin’s evolution theory and as such they are reproducing biological processes such as mutation and cross-over of genes in individuals.

In his talk, Dr. Janos Botzheim presented the following aspects

“Bacterial Memetic Algorithms (BMA) – Basic Aspects”

Bacterial algorithms are specific genetic algorithms, which are based on mutation and gene transfer in the same way as they are appearing in bacteria. The combination with the memetic concept opens the possibility to locally refine rough solutions of the optimization problem in order to enhance the performance of conventional bacterial algorithms.

“Fuzzy Rulebase Optimization with BMA”

One concrete application case of bacterial memetic algorithms is the identification of fuzzy rule bases on the basis of data (such as measurements, signals, image features etc.), where an input-output relationship is extracted and represented as an approximation model. The problem to be optimized includes the least squares error measure between predicted and real measured values and a complexity term containing the number of rules in the fuzzy system.

“Solving Travelling Salesman Problem with BMA”

The second application case was dealing with the famous Travelling Salesman Problem (finding the shortest way between a list of cities on a two-dimensional map), which represents a very difficult problem (NP-hard) and can be solved with the help of genetic algorithms within a reasonable time period and with high precision.

Further information: http://info.sze.hu/index.php?option=com_content&task=view&id=10&Itemid=21#automat, www.flll.jku.at, und www.scch.at

The Software Competence Center Hagenberg (SCCH) is one of the largest independent software research centers in Austria and a harbinger in software-related, technological research & development trends. The SCCH guarantees internationally recognized research & development services. The combination of scientific disciplines around computer science and mathematics makes SCCH unique. The focus is on the following areas:
Process and Quality Engineering
Integrated Software Engineering Tools
Enterprise Applications
Industrial Data Warehousing
Industrial Data Mining
Biomedical Data Mining
Knowledge-Based Vision Systems for Industrial Applications

Software Competence Center Hagenberg GmbH
Mag. Martina Höller, Marketing & PR Manager
Softwarepark 21, A-4232 Hagenberg, Austria
Tel. +43 7236 3343-882
Fax +43 7236 3343-888
mailto:martina.hoeller@scch.at
www.scch.at/

This release was published on openPR.