Tools for perfect combinations
In the Configuration Technologies Cluster, CT experts in Vienna, Austria, and Brasov, Romania, are developing software tools that use clever algorithms to optimize complicated processes and create the best possible combination of all the modules in large installations such as railroad switch towers.
The 21st century is a dynamic age. Companies that have to react flexibly to the needs of the market are quite conscious of this fact. They want their products to match the individual preferences of their customers as closely as possible. That applies not only to cars, built-in kitchens, and PCs, but also to contracts and service packages that are tailored to meet specific customer needs.
This may be a relatively simple matter in the case of consumer products, but it becomes a huge challenge when complex industrial installations are involved. For an electronic railroad switch tower, for example, there are about 20,000 components to consider, as well as some 100,000 constraints under which these components can be used and combined. A correctly functioning system of this kind, which also takes into account all of the customer’s individual preferences, can only be implemented through the use of a knowledge-based expert system – in this case, a “configurator”.
Knowledge-based expert systems originated in the area of artificial intelligence research. They are software systems that calculate solutions for certain problems on the basis of expert knowledge. The relevant expert knowledge is kept separate from the processes and algorithms that are used to calculate the solution, and all of the information about the components under consideration is stored in a knowledge database.
In the case of a railroad switch tower, these components are the possible switches, signals, and tracks, as well as the computer cabinets, control modules, cables, software, and permissible routes. The configurator combines these components in a way that satisfies all relevant conditions – that is, it takes into account which components are compatible with one another under which circumstances. This could mean that all the signal components of a tower are controlled by the same computer, or that access to any valid route must be restricted by the appropriate signals.
In the Configuration Technologies Cluster, CT experts in Vienna, Austria, and Brasov, Romania, are developing software tools that use clever algorithms to optimize complicated processes and create the best possible combination of all the modules in large installations such as railroad switch towers.
New languages
As the basis of their configurators, experts at the Constraint-based Configurator GTF have developed a basic framework that can be used to quickly model and configure systems of any size. In practice, however, it is not often necessary to design complex installations from the ground up. Usually the problem is how to extend a solution that already exists. In order to be able to formulate an adequate description of this task as well, CT experts are working with scientists at the universities in Klagenfurt, Germany, and Oxford, UK, on the development of a new description language and the associated software tools. The name of the project is “Reconcile” (Reconciling Legacy Instances with Changed Ontologies).
If a railroad switch tower needs to be adapted to new country-specific safety regulations, for example, the technical description of some of the components will necessarily change. With a solution developed by Reconcile, the configuration of the entire installation can be automatically adapted to the new requirements in such a way that the costs of the necessary conversions are minimized.
The degree of interest in a universal software tool of this kind can be gauged from the fact that the Austrian Research Promotion Agency FFG named the project proposal of the Reconcile collaboration the best research proposal of 2009 in the “Semantic Systems” category of the FIT-IT program.
Although configurators have been used for a large variety of applications for many years, this is currently a very active and exciting field of research, especially in combination with new technological trends such as sustainable and resource-conserving planning for urban areas and traffic systems, personalized solutions involving autonomous software agents, semantic systems, and the Internet of the future.
The CON Cluster has earned great distinction in this area, in part because of its many years of experience and its expertise in the field of configurators, but also because of its practical research projects and its ability to rapidly implement the results of its research through powerful applications.