Póster: Comparing evolutionary algorithms to solve the game of MasterMind
Download as ODP, PDF0 likes2,189 views
The document compares two evolutionary algorithms, best-worst search (BS) and Evo++, for solving the game of MasterMind. BS search works by first searching for a consistent combination, then searching within the set of consistent combinations. Evo++ searches for a consistent combination simultaneously. Both algorithms are tested using different scoring methods such as most parts scored, best worst case scored, and entropy scored. The results show that the entropy score offers better results overall and that Evo++ needs fewer evaluations than BS to find solutions, though BS is faster in terms of time.
Póster: Comparing evolutionary algorithms to solve the game of MasterMind
- 1. Comparing evolutionary algorithms to solve the game of MasterMind J. Maestro-Montojo, J. J. Merelo, S. Salcedo-Sanz UGR, UAH Combination played Consistent! Always play consistent! Not consistent! Optimization algorithm based on distance to consistency (for all D=2 combinations played) Not all consistent combinations are Testing two different algorithms the same: use partitions Most parts. Score = 5 Best worst case. Score = -3 Entropy. Score = 1.67 0b-0w 0b-1w 0b-2w 0b-3w 1b-0w 1b-1w 1b-2w 2b-0w BS search consistent → search within AAA 2 0 0 0 3 0 0 3 1.31 BBB CCC 4 4 0 0 0 0 0 0 4 4 0 0 0 0 0 0 0.96 consistent set. Score: Entropy 0.96 ABC 0 0 0 1 4 1 1 1 1.58 CBA 0 1 2 0 3 0 2 0 1.52 AAB 1 0 2 0 1 1 0 3 1.67 AAC AAD 1 2 0 2 2 0 0 0 1 1 0 0 0 0 4 3 1.42 1.52 Evo++ →Search simultaneously. Score: Most BCA 0 1 2 1 3 0 1 0 1.67 parts or Entropy Entropy score offers better results. Evo++ needs less evaluations (or more). BS is faster (time-wise) and its number of evaluations increases less.