Rise Sics

MIM-Method Support for Industrial Modeling

The project is a framwork for conducting shorter theoretical studies to gain understanding of what is needed, and to develop and/or adapt the components that are needed in our industrial projects.

Sub-projects:
Modeling and analysis of nonlinear and chaotic systems
 

MIM-Method Support for Industrial Modeling

The project is a framwork for conducting shorter theoretical studies to gain understanding of what is needed, and to develop and/or adapt the components that are needed in our industrial projects.

Sub-projects:
Modeling and analysis of nonlinear and chaotic systems
It has been shown that the nonlinear and so called chaotic systems play a major role in many application areas, such as machine learning and in measurements of computer communications. We therefore intend to study modeling methods that can be used to effectively analyze and manage such systems. The result of the study is a report and / or a demonstration to illustrate important principles in an accessible way.

Specifically, we have examined various filtermetoders usefulness for certain prediktionsproblem in machine learning. Thereby, we have particularly studied so called particle properties and learning problems where the system model consists of nonlinear differential equations. In connection with this, we have shown that there is a class learning problem that is difficult in the sense that learning can not be done for a time horizon longer the time for which the system can be approximated by linear differntialekvationer.

Strategies for traffic management based on capacity analysis
The project studies in the longer term, the possibility of a paradigm shift in the approach to planning and operational control of the railway.

With effective methods and tools for performance analysis can determine which traffic arrangements best suited to a given traffic situation. One can, for example, in some cases, show that for some sparlayouter and some flows of traffic, there are strategies that allow any interference by a train in one flow, can not interfere with the trains of the second stream. By doing this kind of active choice in the operational stage increases flexibility and thus the competitiveness of rail transportsysytemet. It also enables an increased capacity utilization of the track, which in turn leads to reduced investment costs in rail infrastructure as a result of that you avoid building excess capacity due to poor utilization.

The project has studied how policies created some simple sparlayouter and what the essential characteristics of a particular traffic pattern is that makes it fit in some driving situations. Max plus algebra notation has proved suitable for the relevant analyzes of strategies, such as consequences of disturbances.