Multi-agent Negotiation for Distributed Production Scheduling Problems

We extend our previous work of applying automated negotiation to two self-interested agents each with a local but linked production scheduling problem by including multiple agents. To this end, we first present a negotiation protocol built on existing work in the multi-agent negotiation literature before turning to upgrade the two negotiation strategies proposed in the previous work. The first strategy concedes on utility, based on an approach widely used in the negotiation literature. The second strategy concedes in a metric space while maximizing an agent’s local utility. Experimental results show that on small local problem instances, the metric-space negotiation strategy outperforms its utilitybased counterpart on both agreement quality and computational effort, a finding consistent with previous ones.