Settling the Complexity of Arrow-Debreu Markets under Leontief and PLC Utilities, using the Classes FIXP and \Exists-R

Abstract: This paper resolves two of the handful of remaining questions on the computability of market equilibria, a central theme within algorithmic game theory (AGT). Our results are as follows: 1. We show FIXP-hardness of computing equilibria in Arrow-Debreu markets under Leontief utility functions, and Arrow-Debreu markets under linear utility functions and Leontief production sets. We note that these are the first FIXP-hardness results ever since the introduction of the class FIXP and the hardness of 3-Nash established therein. 2. We note that for the problems stated above, the corresponding results showing membership in FIXP were established after imposing suitable sufficiency conditions to render the problems total, as is customary in economics. However, if all instances are under consideration, then in both cases we prove that the problem of deciding if a given instance admits an equilibrium is \Exists-R-complete, where \Exists-R is the class of decision problems which can be reduced in polynomial time to Existential Theory of the Reals. 3. For Arrow-Debreu markets under Leontief utility functions and a constant number of agents, we give a polynomial-time algorithm for computing an equilibrium. This settles part of an open problem of Devanur and Kannan (FOCS'08). We note that PLC utilities are about the most general utilities of interest in economics and several fundamental utility functions studied within AGT are special cases of it. Several important problems, which have been shown to be in FIXP, are waiting for proofs of FIXP-hardness. In this context, our technique of reducing from 3-Nash to Multivariate Polynomial Equations and then to the problem is likely to be useful in the future.