A benchmark problem in online scheduling

The problem considered here is a modification of a flexible manufacturing cell at a plant that manufactures turbochargers. The original production data for one week were given by P. Solot and J.M. Bastos (1987). The problem was modified to make it more demanding by removing one loading station and by changing the part mix to get a more uniform distribution of the load in the system. Two different sets of problems are described. In the first case, the original loading times and continuous order release during the shifts are assumed. This results in moderate average queue lengths. In the second case, the loading times are increased; different load levels for the machines are generated; and the orders are released only at the beginning of each shift. This results in a considerably more difficult decision problem at the loading station, where long queues would exist and the so-called shadowing phenomenon would occur.<
>