A Stochastic Assignment Approach to Resource-Constrained Multi-Project Scheduling

We study the nonpreemptive resource-constrained multi-project scheduling problem in which activity durations as well as costs are a function of the assigned resource. Regarding projectspecific precedence relations, individual release dates and deadlines per project as well as resource restrictions, the question arises, how and when each activity should be scheduled. The problem may be formulated in terms of a zero-one program. Problems of smaller dimensions can be solved to optimality by branch & bound methods [6]. In this paper we present a (highly efficient) stochastic assignment algorithm, which allows to solve large real world problems approximately.