References

1

D. Erbach, Computers and Go, in The Go Player's Almanac, ed. R. Bozulich (The Ishi Press, 1992) 205–17

2

W. Press, B. Flannery, S. Teukolsky, and W. Vetterling, Numerical Recipes in C (Cambridge University Press, 1988)

3

S. Kirkpatrick, Optimization by simulated annealing: quantitative studies, J. Stat. Phys. 34 (1984) 975–86

4

M.P. McLaughlin, Simulated annealing: This algorithm may be one of the best solutions to the problem of combinatorial optimization, Dr. Dobb's Journal (Sept. 1989) 26–37, 88–91

5

K. Binder, D. Heermann, Monte Carlo simulation in statistical physics: an introduction (Springer, 1992)

6

To determine the influence of stones on a board, a potential function can be computed very much like the electric potential due to charged conductors. This is a good anology, since all connected parts of a perfect conductor are always at the same potential, very much like in go where each stone of a chain of stones has the same number of liberties assigned to it. Physics may be able to teach as more about how such potentials are computed.

7

N. Metropolis, A. Rosenbluth, M. Rosenbluth, A. Teller, and E. Teller, Equations of state calculations by fast computing machines, J. Chem. Phys. 22 (1953) 1087-92

8

P.W. Frey, The alpha-beta algorithm: Incremental uptdating, well behaved evaluation functions, and non-speculative forward pruning, in Computer Game Playing: Theory and Practice, ed. M.A. Bramer (Ellis Horwood Ltd., 1983) 285–289

9

The author is not aware of any previous attempts to apply simulated annealing to game trees as they are present in go. There is an isolated comment in [2] about random number generators suitable for ``Monte Carlo exploration of binary trees''. Any references are welcome!

10

D. Fotland, The program G2, Computer Go 1 (1986) 10–6

11

R. Azencott (ed.), Simulated annealing: parallelization techniques (Wiley, 1992)

12

Rob H. Tu, Developing a VLSI go game processor, a Berkeley class project (newsreader post 3/93)