Searching for Solutions in Games and Artificial Intelligence

L. Victor Allis

ISBN 90-9007488-0

Thinking games are investigated from the perspective of research in Artificial Intelligence. In the introduction the relevance of such investigations is discussed. Then, two new search techniques, proof-number search and dependency-based search are described, which have been developed while solving the games connect-four (see M.Sc. thesis by the same author), qubic and go-moku . Finally, the complexity and state-of-the-art in computer game playing is described for each of the games in the so-called Olympic List of 15 well-known games, leading to a prediction for the future of games .

The thesis consists of the preliminary (16 pages), which include a table of contents, list of tables, list of figures and preface, followed by chapter 1 (12 pages, introduction), chapter 2 (52 pages, proof-number search), chapter 3 (30 pages, dependency-based search), chapter 4 (26 pages, qubic), chapter 5 (34 pages, go-moku) and chapter 6 (30 pages, which games will survive?), an appendix (2 pages), a English summary (2 pages), a Dutch summary (2 pages), a curriculum vitae of the author (1 page), a bibliography (10 pages) and an index (5 pages).

A PDF-version of this thesis, made by Francois Grieu, can be found here .

If you would like to obtain the thesis in book form (for a picture of the cover of the book, see above), you can order a copy by sending $US 20 (or the equivalent in any other currency) in CASH in an enveloppe to: Victor Allis, Beneden Boukoul 23, 6071 AH Swalmen, The Netherlands. Make sure to state the postal address to send the thesis to. Checks are not accepted, since the bank fees in cashing them is prohibitive.

Victor Allis is currently a Senior Consultant at Bolesian , the market leader in knowledge technology in the Netherlands.

Searching for Solutions in Games and Artificial Intelligence / L. Victor Allis /