Software Development
After retiring from a very successful academic career at the age of 65, Professor Rzevski, in partnership with Petr Skobelev and Mike Peters, raised venture capital from Eurovestech, London and in 1999 founded a multi-agent technology company named Magenta Corporation, www.magenta-technology.com. Magenta is based in London and its software development unit comprising now around 120 programmers is located in Samara, Russia.
At the time, George was one of the oldest entrepreneurs in the UK. He initially led Magenta as Chairman of the Board of Directors and Chief Executive and later continued as Chief Scientist.
Magenta multi-agent technology is based on Professor Rzevski’s original work further developed in close collaboration with Magenta co-founder, Dr Petr Skobelev, who manages technology development in Samara.
Key components of Magenta software are:
[1] Ontology, which contains conceptual domain knowledge organised as a semantic network
[2] Virtual Market, which represents the environment in which thousands of Agents cooperate or compete whilst attempting to solve a given problem
[3] A Multi-Agent Engine, which provides all supporting runtime code
Magenta products are multi-agent schedulers for ocean and road transportation logistics; knowledge discovery engines; engines for text understanding applications and domain-specific semantic search engines.
Magenta also develops applications in specialised fields, including some aspects of airplane wing design and car production scheduling.
A considerable part of Magenta revenue derives from application development work in accordance with client’s specification, which may or may not encompass proprietary agent technology.
To protect its intellectual property Magenta lodged several patent applications, including:
[1] G Rzevski, P Skobelev. Managing a Virtual Environment. - Patent Application No. 0202527.8, 2003.
[2] G Rzevski, I Minakov, P Skobelev. Data Mining. - Patent Application No. 0403145.6, 2004.
[3] G Rzevski, I Minakov, P Skobelev. Automated Text Analysis. - Patent Application No. 305634, 2004.