Optimization Guided Scheduling I

Abstract

The resource investment problem as specified by the model formulation (11.1) thru (11.6) seems to have a rather simple structure. On the one hand, there is the scheduling aspect of the problem which has nothing more to take into account than the precedence constraints among the activities. On the other hand, there is the aspect of providing resources which is a trivial problem for a given schedule. Because of this, it seems promising to split the hard to solve resource investment problem up into its subproblems and treat each of it in some sense separately. The underlying idea is to iterate two phases until some stopping criterion is met. Phase one is a scheduling phase, and phase two determines the resource requirements for the resulting schedule. From that solution, one can possibly learn how a better solution may look like to start just another iteration that makes use of this information. Promising techniques for deriving such information are mathematical programming techniques. Such approaches will be studied in this as well as in the next chapter (see also Drexl and Kimms [44]).