News

We are undergoing a transition period where computer intelligence is rising to dominance. It will become widely available to everyone at an unprecedented level of intelligence and also at low cost. Such transition periods are often highly chaotic and short-term predictions may hence be rather inaccurate. Based on first principles, however, we can predict what ultimately will prevail after the transition has taken place. This keynote shall provide such an analysis to answer the most relevant questions from the perspective of system simulation: Will the use of simulation rise or fall? Will there be a need for system simulation? What will be the use of Modelica in the future? Will mathematical modeling still be a thing for humans and for computers? The keynote will also explain why it uses the term computer intelligence rather than artificial intelligence.

This paper presents progress toward the simulation of large-scale networks in a strict subset of Modelica using Linear Implicit Equilibrium Dynamics (LIED). A specialized compiler generates modular simulation code that scales to millions of equations — beyond conventional monolithic approaches. Demonstrated with a ThermoLiquidStream library for district heating, fractal networks exceeding 10 million equations are successfully simulated on standard hardware. The results confirm that LIED-based modular compilation offers a practical, industrially relevant path for large network modeling while remaining fully compatible with existing Modelica tools.

This session is chaired by: