Abstract:
This work devises an approach for co-evolving negotiation strategies of agents that have different preference criteria such as optimizing price and optimizing negotiation...Show MoreMetadata
Abstract:
This work devises an approach for co-evolving negotiation strategies of agents that have different preference criteria such as optimizing price and optimizing negotiation speed Whereas many works on e-commerce negotiation define utility functions in terms of price only, this work defines a utility function in terms of both price and negotiation speed. Different emphases on these two criteria can be modeled by placing different weights on them. Hence, in this work, negotiation agents can be price-optimizing, speed optimizing, and P-S-optimizing. Additionally, this work is one of the earliest works that adopt an Estimation Distribution Algorithm (EDA) for finding best response strategies for negotiation agents with different preference criteria. Empirical results show that the EDA can evolve price-optimizing, speed-optimizing, and P-S-optimizing strategies that generally achieve high utilities for negotiation agents. Furthermore, empirical results show that the EDA can evolve to a near optimal strategy for price-optimizing negotiation agents.
Published in: 2008 IEEE Congress on Evolutionary Computation (IEEE World Congress on Computational Intelligence)
Date of Conference: 01-06 June 2008
Date Added to IEEE Xplore: 23 September 2008
ISBN Information: