Contingent Payment Mechanisms for Future Coordination

My research is broadly situated at the interface between economics and computer science (especially artificial intelligence) and draws on concepts from multi-agent systems, planning, and game theory. I am particularly interested in mechanism design for coordination when assigning resources or tasks to agents, when choosing a plan for a future event, in the presence of uncertainty, ex-post decisions, self-interest and private information. Consider assigning time slots for a neighborhood electric vehicle charging station in a way that they are utilized by the residents. At the time of planning, each resident is uncertain about her values and availability of using the station at different times (e.g. they may get stuck in traffic so can't arrive at the station on time, or emergency may happen so they need to drive the cars,) but has private information about their distributions. When the time comes, the values are realized and the assigned residents will then make decisions on whether to use the stations. Similar coordination problems exist when selecting consumers to prepare for reliable consumption reduction to balance supply and demand in electric power systems, or a group of students choosing a time for a project meeting. In each case, the decision is associated with an intended action (to cut consumption, and to show-up at the meeting) and the mechanism needs to elicit information about uncertain values for different alternatives, make a plan for the future along setting payments that may be contingent on people's future actions. We seek truthful mechanisms in which people will voluntarily choose to participate. The rich private information (all possible distributions), the possibility of different payments contingent on different actions, the voluntary choice of actions the agents have and the resulting utilities that are non-quasi-linear (NQL) in payments, impose big challenges for information elicitation and incentive alignment. Moreover, the design objective depends on the outcome that is determined by the ex-post actions taken by the agents, which