Multi-Criteria Evaluation in User-Centric Distributed Scheduling Agents

SRI author:


Berry, P.M. and Gervasio, M. and Peintner, B. and Uribe, T. and Yorke-Smith, N. Multi-Criteria Evaluation in User-Centric Distributed Scheduling Agents, in Proceedings of AAAI Spring Symposium on Distributed Plan and Schedule Management, AAAI Press, pp. 151–152, Mar 2006.


This position paper discusses the problem of locally evaluating and comparing candidate schedules, in the context of a distributed scheduling task operating in unbounded environments in which each agent selfishly serves the desires and preferences of its own user.

Distributed scheduling systems have used a variety of mechanisms to maximize the quality of the global solution. Techniques include negotiation frameworks based on market economies [Kis et al. 1996], game theoretic algorithms, and global or shared evaluation functions [Modi et al. 2005]. Our position is that these techniques do not adequately address the situation where self-interested cooperative agents maintain schedules on behalf of users operating in unbounded environments. Here, satisfaction is more important than optimality [Palen 1999], and personalized preferences are paramount to users [Berry et al. 2005a; Faulring and Myers 2005]. In our approach, global utility is secondary; the agent aims to maximize the utility of its user, but folds the positive utility of cooperating with others into that utility. Thus, we treat the utility of others as a single component of a personalized and context-dependent multi-criteria evaluation function, which the agent learns through interactions with its user.

Read more from SRI