May 14 (Thursday) 10:00-11:00, room C.510 (new building, 5th floor)
Szilvia Pápai (Concordia University, Canada)
Pareto-Efficient Matching with Gradual Acceptances
We introduce the class of Gradual Acceptance matching rules, which assign objects to agents through a gradual procedure in which applicants are accepted selectively, rather than immediately accepting all current highest-priority applicants at each step, as in the Immediate Acceptance rule in school choice. The central idea is to slow down the acceptance process so that applicants can be accepted without creating priority violations whenever possible, while making permanent assignments at each step and hence maintaining Pareto-efficiency. We develop the secure-set concept and a priority-profile reduction procedure, leading to the Secure Gradual Acceptance rule, which is Pareto-efficient at every profile and selects a stable and Pareto-efficient matching whenever such a matching exists. Our main result uses these ideas to characterize, through a sequential procedure that imposes joint restrictions on preferences and priorities, the profiles at which Pareto-efficiency and stability are compatible. We also show that gradual acceptance can improve the stability and incentive properties of the Immediate Acceptance rule while preserving Pareto-efficiency.
The talk is based on joint work with Israa Hashem.