Mechanism design is a subfield of economics that studies the design of optimal incentive systems. It is sometimes also referred to as reverse game theory. Where game theory asks the question "Given the rules of a game, how will players play if they want to maximize their own returns?", mechanism design asks the question "If we want the interaction of self-interested agents to lead to a particular goal, what should the rules of the game be?"
Mechanism design has been fundamental to the design of spectrum auctions and for the auctioning of on-line advertisement space in search engines. At Xerox applications include transport demand management (for example for on-street parking) and optimal outsourcing.
New applications often are large scale, supported by digital means and involve challenging computational aspects. Algorithmic mechanism design is sometimes used to refer to the field that looks at these problems that are at the interplay of economics and computer science. The research at Xerox is unique because it combines mechanism design with machine learning: this allows for continuously improving systems, leveraging all insights from (real-time) data, while ensuring that human participants find it in their best interest to optimally participate.