Comparison of Reinforcement Learning and Combinatorial Auctions

Abstract

The concluding chapter of this thesis will first summarize and compare the properties of the two main methods reinforcement learning and combinatorial auctions with respect to their application for PCRA in distributed computer systems. In a second step the main results of the empirical evaluation and simulations performed within this work will be discussed in terms of their application to the ISIP task allocation problem portrayed in the initial section 1.4. A catalog of recommendations derived from the results will be given for the practical application of the RL and CA methods investigated in real world IT environments. The chapter concludes with a survey of the possible development of this research area and provides a short survey of further questions that arise from the results of this thesis and are worth investigating in future research.