Kai Himstedt

From Chessprogramming wiki
Jump to: navigation, search

Home * People * Kai Himstedt

Kai Himstedt [1]

Kai Himstedt,
a German computer scientist and Ph.D. from the University of Hamburg on parallel game-tree search and Grid computing [2] [3]. Along with Ulf Lorenz, he is principle author of GridChess and Cluster Toga, cluster chess programs based on the open source chess programs Fruit by Fabien Letouzey, and the Fruit derivate Toga. GridChess is the doctorate of Kai Himstedt and a cooperation between the Paderborn University and the University of Hamburg. GridChess played the WCCC 2007, Cluster Toga the WCCC 2008 [4]. His further cluster chess projects GridProtector and GridGinkgo, playing the PT 49 in October 2015, and the WCCC 2016, incorporate Raimund Heid's Protector, and Frank Schneider's Ginkgo respectively.

Jugend forscht

As a teenager during the late 70s, Kai Himstedt already developed a chess program along with Wolfgang Stahn and participated at Jugend forscht, a German youth science competition, initiated in 1965 by Henri Nannen, then editor-in-chief of the Stern magazine. In 1980, they won the second prize in the mathematics and computer science section [6]:

Inspired by the chess computer boom and not impressed by the programs offered at the time, the two young friends developed their own computer chess program. As a comparison, the modern American chess program Chess 4.8 was available, in which the two young mathematicians discovered certain weaknesses. 
By using the FORTRAN programming language and the TR 440 computer, the two young mathematicians set about developing their new program. The considerations for the development took about four weeks. Then in April 1979, first games could been played, but were unsatisfactory in the eyes of young researchers. Further improvements of the program made it possible that one of the best micro-computer chess prograns, the Chess Challenger 7 could be beaten. 
Their program proved to be so good that even the computer chess world champion Chess 4.8 could be beaten when using a suitable computer. A special feature of the new program is that an exchange of pieces with its effects on the overall course of the game was not completely calculated by search, but evaluated statically, yielding in a considerable shortening of the processing time. 

See also

Selected Publications

[7] [8]

External Links


Up one level