An Adaptive Learning Method in Automated Negotiation Based on Artificial Neural Network

Negotiation is an important activity most related to the decision-making process in the e-business. It involves the interaction between different parties and usually goes through a number of iterations. Similar to the negotiation process in the real world, software agent working in the virtual environment performs automated negotiation in this way. This paper proposes an agent-based learning method in automated negotiation based on artificial neural network. The aim of it is to implement interactions between agents and guarantees the profits of the participants for reciprocity. In the system, each agent has a learning capability implemented by an artificial neural network to generate sequential offers and can be trained by the previous offers that have been rejected by the other agent. With the negotiation model, software agents can negotiation with each other over a set of different issues of a product on behalf of the real-world parties they represent. The experiments have been conducted to evaluate its performance and the results show the efficiency and promise of the proposed system.