Solving House Allocation Problems with Risk-Averse Agents

This paper solves the fair and optimal house allocation problem (Sun and Yang, Econ Lett 81:73–79, 2003) when the agents preferences are represented by nonlinear utility functions using techniques for global mixed integer nonlinear optimization. A small simulation study indicates that if quasi-linear specifications are adopted as approximations to nonlinear utility functions and if the fair and optimal allocation is identified based on this approximation, then the prices are typically higher on average and the resulting allocation is typically non-fair.