A deep improver of two-move chess mate problems

Yaakov HaCohen-Kerner, Fridel Fainshtein

    Research output: Contribution to journalArticlepeer-review

    6 Scopus citations

    Abstract

    Computer composition of high-quality chess mate problems is relatively an uninvestigated research domain. A previous model, called an Improver of Chess Problems, which is based on a hill-climbing search, improved slightly the quality of 10 out of 36 known problems (about 28%). In this article, we describe an improved model, called Deep Improver of Chess Problems. This model uses an improved version of Bounded Depth First Search. The experiment, we carry out on the same problems, show that the quality of 32 problems (about 89%) has been improved. Moreover, some of the improvements are meaningful and impressive.

    Original languageEnglish
    Pages (from-to)443-462
    Number of pages20
    JournalCybernetics and Systems
    Volume37
    Issue number5
    DOIs
    StatePublished - 1 Aug 2006

    Fingerprint

    Dive into the research topics of 'A deep improver of two-move chess mate problems'. Together they form a unique fingerprint.

    Cite this