Abstract
A conveyor belt system consisting of a dozen inserters ("hoppers") placed one item in each belt "slot" to produce batches of fixed length and order. Because there were usually more hoppers than item types, production could be speeded by loading more than one item per slot, with the constraint that at the end of the belt, the slot contents would stack in proper order. This led to a conceptually difficult design problem for the controller, which was solved by simulating an idealized machine that produced the desired results. A further improvement in throughput was obtained by using a simple random optimization scheme to find near optimum ordering of hopper contents.
Get full access to this article
View all access options for this article.
