Allocation of in-house services: Experimental comparison of allocation mechanisms

In this paper we consider the problem of allocating scarce resources in a divisionalized company; the resources are made available by the headquarters and requested by profit centers (PCs). This problem was observed at a large german insurance company (LAGIC). Currently, the resources are allocated for free; for the future, LAGIC proposes to use an iterative allocation mechanism to solve the allocation problem. Alternatively, we present a combinatorial auction-based allocation mechanism. Within experimental tests the performance of these two mechanisms is analyzed. To this end, instances are generated motivated by the problem size and structure occurring at LAGIC.