Abstract:  Effective strategies for dealing with patent licensing have been attracting greater interest amongst enterprises due to acceleration in high-tech development and rapidly expanding international trade. In a modern society, patent licensing is a widely accepted approach, viewed as a form of formal transaction. It allows for a licensee to implement the patented technology for future applications, and in return, pay royalties to a licensor for the potential value of the patents provided. Until recently, most patent licensing activities were face-to-face talks, however, due to complexities and uncertainty during the negotiation processes of a patent licensing transaction, asymmetric information between licensors and licensees exists. This can hinder a fare and optimal solution for licensing transactions. This study attempts to design a multi-agent negotiation system that employs Genetic Algorithms (GA) for automatically choosing best-fit strategies for patent licensing matching. The system, which serves as a patent intermediary, relies on experts from industries and professional organizations to establish the licensing negotiation conditions. Based on these realistic conditions provided by experts, this automated negotiation system is able to conduct preliminary negotiation activities between licensees and licensors. Through the automated process, lower costs and higher efficiency can be achieved. We hope that this system has the potential for handling multi-attribute based negotiations for real situations in the future.

Keywords: Patent licensing; Multi-agent; Negotiation; Genetic Algorithm (GA); Support system

 

The patent marketplace has expanded significantly due to an increasing demand for patents. In a public patent marketplace, patent licensing is a widely accepted approach, and is viewed as a form of formal transaction. Support systems for the patent market have been attracting greater interest due to accelerated high-tech development and rapidly expanding international trade. However, most current support platforms work as a patent information publishing platform. Despite the fact that some of them provide professional information analysis for customers, no intelligent and dynamic system currently exists that deals with automated transaction activities. This study attempts to design a multi-agent negotiation system that employs Genetic Algorithms (GA) for automatically choosing best-fit strategies for patent licensing matching.