Genetic algorithms for automated negotiations: a FSM-based application approach

Abstract

An approach to implement strategies for automated negotiations in electronic commerce applications is presented. It is based on genetic algorithms (GAs) that evolve FSMs (finite state machines). Each of these FSMs represents a negotiation strategy that competes against other strategies and is modified over time according to the outcome of this competition by using GA principles. The paper gives an overview of negotiating agents and work related to the paper. Then the application of genetic algorithms to FSMs is presented and relevant details on the implementation are given.