Assigning Multiple Job Types to Parallel Specialized Servers

In this paper methods of mixing decision rules are investigated and applied to the so-called multiple job type assignment problem with specialized servers which is modeled as continuous time Markov decision process. Performance optimization is difficult for this assignment problem, but optimization over the class of static policies is tractable. By applying the described mixing methods a suitable static decision rule is mixed with some dynamic decision rules which are easy to describe and implement. For the discussed mixing methods optimization is performed over corresponding classes of so-called mixing policies. These mixing policies are still easy to describe and implement and for all investigated instances the optimized mixing policies perform substantially better than optimal static policies. Moreover, the optimized mixing policies perform better than stationary dynamic policies which apply at decision epochs one of the dynamic rules to which the mixing methods have been applied.