Project scheduling under partially renewable resource constraints

We consider a generalization of the classical resource constrained project scheduling problem. We introduce so-called partially renewable resources by assuming for each resource a capacity on subsets of periods. The concept of partially renewable resources is a fundamental tool in order to make, e.g., timetabling and shift scheduling aspects, amenable to project scheduling. In addition, partially renewable resources serve to model complicated labor regulations. Furthermore, they cover traditional renewable and nonrenewable resource constraints as special cases. We consider makespan minimization as objective. For the exact solution of the problem we employ a basic enumeration scheme. In order to speed up convergence, we formulate bounds which take into account future resource consumption of partially renewable resources. Moreover, we generalize the serial scheduling scheme in order to get fast approximation methods. A rigorous assessment of the procedures is provided by solving ProGen instances generated under a full factorial test design. Besides the well-known problem parameters we employ additionally three parameters which control the generation of partially renewable resources.