Altruism in Facility Location Problems

We study the facility location problems (FLPs) with altruistic agents who act to benefit others in their affiliated groups. Our aim is to design mechanisms that elicit true locations from the agents in different overlapping groups and locate a facility to serve agents to approximately optimize a given objective based on agents' costs to the facility. Existing studies of FLPs consider myopic agents who aim to minimize their own costs to the facility, while we mainly consider altruistic agents who consider the group costs incurred by all agents in their groups. Accordingly, we define Pareto strategyproofness to account for this new type of agents and their multiple group memberships with incomparable group costs. We consider mechanisms satisfying this strategyproofness under various combinations of the planner's objectives and agents' group costs. For each of these settings, we provide upper and lower bounds of approximation ratios of the mechanisms satisfying the Pareto strategyproofness.