Development of an automatic, multidimensional, multicriterial optimization algorithm for the calibration of internal combustion engines

The process of engine application is crucial for the fulfillment of exhaust emission limits and for the reduction of fossil fuel consumption. During the past decade, this process became more and more complex, consequently the test bench measurement times become longer and even more expensive. Commercial software tools try to deliver good application methods, that reduce the total application time and application costs. In this work an alternative holistic optimization approach based on integer optimization is presented. Furthermore, we validate the algorithm by testing the approach with test bench data, which have been compressed to a polynomial data model. Finally, we present current results based on the new real world emissions cycle. The numerical results show that this new approach reduces measurement and optimization time in comparison to State-of-the-Art methods, noticeably. Additionally, the results show that one does not need data modeling, in order to solve the multi-criteria optimization problem