Incentive Alignment Through Secure Computations

We present a game-theoretic approach to analyzing the incentive structure in formal models of inter-organizational businesses processes. In such processes, the choices of each participants influence the outcome of others. A potential participant may be torn between the prospect of a highly preferable outcome on the one hand (e.g., a bonus on timely delivery), and the possibility that another player may make a choice (e.g., reallocation of the fast trucks) which renders that outcome impossible to achieve. We propose (a) an analysis which given the preferences of participants determines if the collaboration is at all meaningful; (b) an algorithm for modifying the incentive structure of such a process using both fines and outcome re-distribution to increase the benefit for all participants; and (c) a practical way of computing this algorithm while concealing the preferences of the collaborators for each other using secure multi-party computation.