An evolutionary algorithm coupled with the Hooke-Jeeves algorithm for tuning a chess evaluation function

In a previous paper presented at CEC'2011, we reported the implementation of a chess engine based on evo- lutionary programming with a selection mechanism that relied on grandmaster's chess games. The objective was to decide the virtual players that would pass to the following generation. Here, we use these same techniques to adjust a larger number of weights (29 in this work instead of the 5 used in the previous one). The aim was to improve the rating of our chess engine. We also introduce here the use of a local search scheme based on the Hooke-Jeeves algorithm, which is adopted to adjust the weights of the best virtual player obtained in the evolutionary process. As our results indicate, this produced a further improvement in the rating of our chess engine. As in our previous work, the material values of the additional pieces considered here are similar to the values known from chess theory.

[1]  David B. Fogel,et al.  Further Evolution of a Self-Learning Chess Program , 2005, CIG.

[2]  Nathan S. Netanyahu,et al.  Expert-driven genetic algorithms for simulating evaluation functions , 2010, Genetic Programming and Evolvable Machines.

[3]  D. Hunter.  MM algorithms for generalized Bradley-Terry models , 2003 .

[4]  Claude E. Shannon,et al.  Programming a computer for playing chess , 1950 .

[5]  Feliu Sagols,et al.  An evolutionary algorithm for tuning a chess evaluation function , 2011, 2011 IEEE Congress of Evolutionary Computation (CEC).

[6]  Zbigniew Michalewicz,et al.  Genetic Algorithms + Data Structures = Evolution Programs , 1992, Artificial Intelligence.

[7]  Lawrence J. Fogel,et al.  Artificial Intelligence through Simulated Evolution , 1966 .

[8]  Janez Brest,et al.  A Differential Evolution for the Tuning of a Chess Evaluation Function , 2006, 2006 IEEE International Conference on Evolutionary Computation.

[9]  H. Nasreddine,et al.  Using an Evolutionary Algorithm for the Tuning of a Chess Evaluation Function Based on a Dynamic Boundary Strategy , 2006, 2006 IEEE Conference on Cybernetics and Intelligent Systems.

[10]  Robert Hooke,et al.  `` Direct Search'' Solution of Numerical and Statistical Problems , 1961, JACM.

[11]  Donald E. Knuth,et al.  An Analysis of Alpha-Beta Pruning , 1975, Artif. Intell..

[12]  Moshe Sipper,et al.  Evolution of an Efficient Search Algorithm for the Mate-In-N Problem in Chess , 2007, EuroGP.

[13]  Claude E. Shannon,et al.  XXII. Programming a Computer for Playing Chess 1 , 1950 .

[14]  H. J. van den Herik,et al.  Information in transposition tables , 1997 .

[15]  Janez Brest,et al.  An Adaptive Differential Evolution Algorithm with Opposition-Based Mechanisms, Applied to the Tuning of a Chess Program , 2008 .

[16]  Moshe Sipper,et al.  GP-EndChess: Using Genetic Programming to Evolve Chess Endgame Players , 2005, EuroGP.

[17]  Janez Brest,et al.  History mechanism supported differential evolution for chess evaluation function tuning , 2010, Soft Comput..

[18]  David B. Fogel,et al.  The Blondie25 Chess Program Competes Against Fritz 8.0 and a Human Chess Master , 2006, 2006 IEEE Symposium on Computational Intelligence and Games.

[19]  D.B. Fogel,et al.  A self-learning evolutionary chess program , 2004, Proceedings of the IEEE.

[20]  H. Jaap van den Herik,et al.  Simulating human grandmasters: evolution and coevolution of evaluation functions , 2009, GECCO.