Model Parameters Optimisation for an Industrial Application: A Comparison between Traditional Approaches and Genetic Algorithms

Model parameters optimisation is a very common problem when dealing with mathematical models. These models are often designed from theoretical considerations on the physics phenomena and afterwards adapted in order to fit the experimental data that are collected on the real operating scenario. The paper compares different approaches to the problem of finding the optimal values of four parameters characterising a quite complex mathematical model which estimates some important mechanical properties of aluminium killed and interstitial free steels. Several optimisation procedures have been attempted, from traditional methods to genetic algorithms. A comparison of such methods is performed, by illustrating and discussing numerical results.