The problems of servicing of the binary object flow in system with refillable storage component

We consider a model for single-stage servicing of a finite deterministic stream of objects by a processor with a storage component, a reservoir of bounded capacity. The stream consists of a stream of objects that fill the reservoir and a stream of objects that are filled from the reservoir. With each object, we associate a linear individual penalty function for the time it spends in the servicing system. We study the problem of constructing a schedule that minimizes total penalty over all objects in the stream. Our algorithms are based on dynamic programming, branch and bound techniques, and their joint implementation. We show results of computational experiments.