A Convex Optimization Framework for Public Resource Allocation under Equity and Parameter Uncertainty.

Source abstract

In this paper, we develop a compact framework for allocating a fixed public budget across competing sectors under spending bounds, equity considerations, and parameter uncertainty. Sectoral benefits are modeled by increasing concave functions, yielding diminishing marginal returns. Existence and uniqueness are established, and Karush-Kuhn-Tucker conditions characterize the baseline optimum. Equity is incorporated through a convex quadratic penalty, while a robust counterpart protects against uncertain benefit coefficients without sacrificing concavity or separability. The model provides a transparent decision-support structure in which priorities, weights, and constraints remain policy inputs.