One Decade of Experience in Coupling a Genetic Algorithm with Casting Process Simulation (V. Kokot)

Abstract

More than 10 years ago, MAGMA launched an optimization module in casting simulation. A genetic algorithm has been integrated, which in principle allows the processing of any kind of typical questions for casting simulation applications. Despite the high cost for evaluation of a single design by the simulation, this approach has led to successful solutions of practical problems, partly with surprising results, partly with results reflecting the well-known principles and physical laws.

One of the greatest obstacles in introducing optimization to classical simulation is not the technology (e.g. calculation time) but changing the mindsets of the users. The variant calculation with statistical assessment capabilities opens up a new view to application of simulation and requires a different approach when dealing with problems from practice. This change in thinking and the reduction of prejudices and concerns among users is the most time-consuming and difficult component to establish optimization in simulation applications.

MAGMA can look back on more than one decade of experience. Today, the optimization is an integral part of our software, which is available to all our users.