Output details

Comparing job allocation schemes where service demand is unknown

<13> This is a modelling approach to studying different scheduling techniques. A key result is the observation that the success of the TAGS scheduling scheme is due simply to a high coefficient of variation, rather than a heavy tail property as previously claimed by several other authors. This result has subsequently been acknowledged by Halchol-Balter (a leading authority on scheduling performance and previous proponent of the heavy tail view). The formal modelling approach used here is different from other works in this field, which are largely statistical in nature.