A Multi-Objective Optimization Approach for Bulk Material Blending Systems

Nowadays bulk material blending systems still mainly implement the static well-known chevron stacking. Real-time quality measurement techniques as online X-ray fluorescence measurement allow the dynamic adaptation of the blending process to the current quality data. We propose a multi-objective optimization system on the base of steady state evolutionary algorithms using various Baldwinian and Lamarckian repair algorithms and test the algorithms on real world problem data. The optimized solutions generally outperform the standard techniques and furthermore allow insights and deeper understanding of the problem nature and the algorithms presented.