Cookies help us deliver our services. By using our services, you agree to our use of cookies.

PhD Thesis Defence (Borko Bošković)


Revision as of 06:15, 28 May 2010 by Borko (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

Date: April 13 2010

Work:

Borko Bošković has on April 13 2010 successfully defenced a PhD Thesis called: "Differential evolution for the Tuning of a Chess Evaluation Function".

Digital library (pdf)

Congratulations!

Abstract

We present a Differential Evolution (DE) based approach for chess evaluation function tuning. DE with opposition-based optimization is employed and upgraded with a history mechanism. The opposition-based optimization improves efficiency of DE because of noise in the evaluation of individuals. Additionally introduced history mechanism reduces noise, because potentially good individuals played more games through several generations. New history mechanism uses an auxiliary population containing potentially good individuals from history of the evolutionary process. This mechanism ensures that potentially good individuals remain within the evolutionary process, even though they died several generations back and later can be injected back into the evolutionary process. In such manner, developed algorithm reduces noise in the evaluation of individual, reduces the possibility of overfitting and consequently improves efficiency of whole tuning process.

Gallery