Representative agents and the cold start problem in contract negotiation

Principal Agent Theory (PAT) seeks to identify the incentives and sanctions that a consumer should apply when entering into a contract with a provider in order to maximise their own utility. However, identifying suitable contracts-maximising utility while minimising regret- is difficult, particularly when little information is available about provider competencies. In this paper we show that a global contract can be used to govern such interactions, derived from the properties of a representative agent. After describing how such a contract can be obtained, we analyse the contract utility space and its properties. Then, we show how this contract can be used to address the cold start problem and that it significantly outperforms other approaches. Finally, we discuss how our work can be integrated with existing research into multi-agent systems.