Computational aspects of assigning agents to a line
Working paper
Issue number:
2016/54
Series:
CORE Discussion Papers
Publisher:
Université Catholique de Louvain
Year:
2016
We consider the problem of assigning agents to slots on a line, where only one agent can
be served at a slot and each agent prefers to be served as close as possible to his target. We
introduce a general approach to compute aggregate gap-minimizing assignments, as well
as gap-egalitarian assignments. The approach relies on an algorithm which is shown to be
faster than general purpose algorithms for the assignment problem. We also extend the
approach to probabilistic assignments and explore the computational features of existing,
as well as new, methods for this setting.