Tree Pruning for New Search Techniques in Computer Games

Kieran Greer

doi:10.1155/2013/357068

Abstract

This paper proposes a new mechanism for pruning a search game tree in computer chess. The algorithm stores and then reuses chains or sequences of moves, built up from previous searches. These move sequences have a built-in forward-pruning mechanism that can radically reduce the search space. A typical search process might retrieve a move from a Transposition Table, where the decision of what move to retrieve would be based on the position itself. This algorithm stores move sequences based on what previous sequences were better, or caused cutoffs. The sequence is then returned based on the current move only. This is therefore position independent and could also be useful in games with imperfect information or uncertainty, where the whole situation is not known at any one time. Over a small set of tests, the algorithm was shown to clearly out perform Transposition Tables, both in terms of search reduction and game-play results. Finally, a completely new search process will be suggested for computer chess or games in general.

Highlights

This paper describes a new way of dynamically linking moves into sequences that can be used to optimise a search process
Dynamic move chains would be retrieved only during the α-β search and for possibly 1/3 to 1/2 of the nodes searched in that part
More intelligence and reasoning appear to be part of new search processes in any case. These tests are based on a relatively small number of examples, but they show the potential of the algorithm and confirm that dynamic move sequences are at least reliable. They can be added to a search process and used without returning unreliable moves

Summary

Introduction

This paper describes a new way of dynamically linking moves into sequences that can be used to optimise a search process. Move sequences are returned during the searching of the chess game tree that cause a cutoff, or determine that certain parts of the tree do not need to be searched further These move sequences are usually stored in Transposition Tables [1, 2], but instead, they can be stored in a dynamic linking structure and reused in the same way. Computer chess can be used, as a domain for testing AI-related algorithms It is still an interesting platform for trying to mimic the human thought process or add human-like intelligence to a game-playing program.

Chessmaps Heuristic

Dynamic Move Sequences

Related Work

Testing

Future Work

Conclusions