Automated Economic Reasoning with Quantifier Elimination

Опубликовано 6 марта 2018, 18:40

Many theorems in economics can be proven (and hypotheses shown to be false) with available decision algorithms. This paper assembles a few dozen classic examples from economic theory that demonstrate why CAD-based algorithms are practical, and how the infamous double-exponential characterization of their complexity can be misleading in practice. The discrepancy between reputation and practice comes from a combination of: the approximation error in asymptotic complexity theory, the distinction between projecting formulas and projecting polynomial sets, and the algebraic structure of the applied problems. The results also suggest possibilities for (a) finding resource-conserving sequences for eliminating variables and (b) enhancing SMT-solver and other non-CAD algorithms for application to economic reasoning.

See more at microsoft.com/en-us/research/v...