Sioux works with a manufacturer of machines for packaging consumer products in boxes, such as bottles of detergent, ice cream cones and bags of crisps. These machines have 5 or 6 robot arms that fill a box simultaneously. Mixed packaging with several types of flavours is the current trend. It is also important for both products and boxes to move through the machine in the same direction in order to save floor space in factories. The old packaging algorithm is unable to meet these requirements.
Through a simulation of the packaging process, Sioux created insight in the underlying scheduling problem. The simulation could be used immediately to compare the quality of different heuristics. In the end, this resulted in the implementation of a custom-made scheduling algorithm that runs in real time on the packaging line.
The new packaging algorithm has a threefold effect. The usability of the machine is improved, the client is able to more thoroughly cater to their customers’ wishes and the machines have an increased output.
Scheduling, heuristical optimisation, search algorithms, mixed integer optimisation, software design, machine interfacing and conceptual visualisations. Mathware technologies: Chi3, C++, Python, Simulink, Simscapea and Modelica.
