Truth-telling, dominant strategies, and iterative Groves mechanisms

This essay demonstrates that if an organization uses a price-decomposition algorithm to solve a decentralized resource allocation problem, in which rewards are those of a Groves mechanism, then truth-telling remains a dominant-strategy equilibrium. Earlier work shows that for a general class of non-linear programming algorithms, truth-telling is a Nash equilibrium that weakly dominates all other Nash equilibria, but it is not necessarily a dominant-strategy equilibrium.