Modeling Task Allocation with Time using Auction Mechanisms

The e cient allocation of tasks to groups of agents has been studied extensively in various scenarios. In this thesis we look at the task allocation problem with the added parameter of multiple times at which the job may be completed. We consider both the settings where there is a single task to be done and also where there are multiple subtasks of a composite task to be completed. We model the two settings as an auction where workers submit bids for completing individual tasks or bundles of tasks at di erent times. In the combinatorial auction setting, we entertain the idea that workers may consider groups of subtasks to be complementary, and we adjust bids on bundles of tasks using a measure of complementarity. We propose a mechanism for determining the optimal allocation in this setting. Finally, we simulate our mechanism and show that it can e ectively take advantage of the complementarity of subtasks.