Convolution, Non-transferable Utility, and Money

Sjur Didrik Flåm

Convolution plays important roles in nonsmooth analysis and optimization. It does, however, not quite fit instances that feature "non-transferable utility." Such instances are common in economics and game theory. Additional problems emerge upon convoluting data or decisions that are distributed across diverse agents. This paper attempts to get around some of these hurdles by "monetizing" individual objectives and by using direct exchange as main vehicle.