Non-linear optimization of track layouts in loop-sorting-systems

Optimization used for enhancing geometric structures is well known. Applying obstacles to the shape optimization problem is on the other hand not very common. It requires a fast contact search algorithm and an exact continuous formulation to solve the problem robustly. This paper focuses on combining shape optimization problems with collision avoidance constraints by which a collision detection algorithm is presented. The presented method is tested against the commercial loop-sorting-system used for sorting of medium sized items. The objective is to minimize price and footprint of the system while maintaining its functionality. Contact constraints are in this context important to include as various obstacles may surround the loop-sorting-system.

► Test cases shows that an optimal length of the layout can be obtained. ► A collision algorithm proves efficient to avoid obstacles and building walls. ► It is possible to find a discrete solution close to the optimum. ► A gradient based algorithm provide the best results.