Developments in Monte-Carlo Proof-Number Search
Jahn Takeshi Saito, Guillaume M. J. -B. Chaslot, JOS W. H. M. UITERWIJK, H. Jaap, Mark H. M. Winands, H. M. Winands · 2006
Over the years, proof-number search has successfully been applied to many game domains. Only recently, the combination of Monte-Carlo sampling and proof-number search was introduced. This proof-number enhancement applies Monte-Carlo sampling to ini-tialize proof numbers. This article outlines Monte-Carlo Proof-Number Search and presents two refine-ments to the algorithm: (1) novel tuning of standard parameters, and (2) static pattern evaluation. Ex-periments yield that (1) tuning of parameters can lead to better results than recorded previously, and (2) static patterns are difficult to apply directly. Fu-ture work will focus on examining how Monte-Carlo sampling can be carried over to depth-first proof-number search. 1