Designing Heuristics: Hybrid Computational Models for Teaching the Negotiation of Complex Contracts

Abstract:     
The negotiation of even the most straightforward real-world contracts tends to be quite complex. A contract with only 25 distinct issues with two alternatives each presents the parties with more than 33 million possible contracts, far too many to be evaluated exhaustively within feasible time constraints. Furthermore, contract issues that exhibit high levels of interdependence result in highly nonlinear utility functions with the possibility of many local optima. This paper employs hybrid computational models, integrating both simulated annealing and tabu list optimization, to aid in the design of social heuristics and institutional mechanisms that may serve to improve the effectiveness of human negotiators.

Working Paper Series