Skip to content

shambakey1/coevolutainary_strategy_based_adversial_search

Repository files navigation

adversial_project

This is an implementation of the Adversial Search project of Udacity AI Nanodegree. The implementation uses co-evolutionary strategy optimization algorithm to find the sub-optimal move of current player.

The co-evolutionary strategy is similar to Monte Carlo Tree Search (MCTS) in that it runs many simulations of the game to decide the next move of current player. Also, the co-evolutionary strategy is similar to Genetic Algorithm and Evolutionary Strategy optimization techniques in that it uses crossover, mutation and selection.

PS: The implmenetation uses files from the DEAP python module (https://github.com/DEAP/deap) with some modifications. I modified the "eaSimple" algorithm in the DEAP library to "eaSimple_mod" to suit the project's knights isolation problem. Unfortunatelly, the implementation consumes too much memory than what is required by the project (maybe due to the numpy library)

Releases

No releases published

Packages

No packages published

Languages