Optimizing Strategy in Agent-Based Automated Negotiation

Digital Business Agents (DBAs) can assist human buyers and sellers in electronic markets by strategically conducting automated negotiation to minimize transaction costs. However, the resulting information systems are complex environments, which are hard to assess analytically. The DBAs’ strategies will thus need to incorporate heuristics, which adapt to ever changing environment conditions using machine learning algorithms. This article compares the performance of an evolutionary algorithm, numeric optimization methods and a hybrid, economics-based mechanism, using the electronic market setting of the multiagent system AVALANCHE as an example.