The main focus of this work will be to investigate negotiation between agents. We will assume that negotiation is actually a suitable method for resolving conflicts. This assumption is justified by the fact that to resolve any conflict, there will be some kind of data flow between agents, independent of the method used. So even if a different method of conflict resolution, such as having a mediator, is used, the same effect may be accomplished by having direct communication between the agents. Also, the SiFA model assumes that each agent has all the knowledge it needs to carry out the design process, and that agents can talk to each other directly. With these considerations, direct negotiation between agents follows as a very natural choice.
In order to identify all conflict types in a design system, we will try to fill in all entries of a matrix with agent types on both axes, i.e., what are the conflicts that might arise between an evaluator and a selector?
Once all meaningful conflict types have been identified, we will investigate strategies that the involved agents can use to resolve the conflict. These strategies will consist of rules that define the behavior of an agent in all aspects of a conflict situation. At this point, we will study which aspects of the current state of the system should be considered to define a conflict situation. This conflict situation will be used by the agent to decide on a negotiation strategy.
These negotiation behaviors will be implemented as an addition to the already existing SINE platform. This will allow experimentation by which we can answer one of the above questions, namely, trying to see if patterns of communication exist between pairs of agents.