查看文章

We propose a mechanism for negotiating multiple contracts simultaneously which can be applied to many different problems. It utilizes a genetic algorithm that employs one of two different rank-based fitness functions. The mechanism is evaluated by means of the complex contract negotiation problem (CCNP), a clause-based contract model with interdependent issues. Furthermore, a transportation request assignment problem (TRAP) where pickup and delivery requests are to be assigned to carriers is solved. For the CCNP, the mechanism is able to find optimal solutions for many small instances. For the TRAP, solution quality depends highly on the tested instance class. Independent of the actual planning problem, the mechanism maintains a high level of information protection for all agents.