본문 바로가기
  • Home

Genetic Algorithm for Number Collecting Maze Game

  • Journal of Internet of Things and Convergence
  • Abbr : JKIOTS
  • 2025, 11(6), 17
  • Publisher : The Korea Internet of Things Society
  • Research Area : Engineering > Computer Science > Internet Information Processing
  • Received : November 25, 2025
  • Accepted : December 17, 2025
  • Published : December 31, 2025

Sangwook Lee 1

1목원대학교

Accredited

ABSTRACT

The Number Collecting Maze game involves players navigating a grid-shaped maze to collect specific numbers. The Number Collecting Maze game has been applied in educational and puzzle design fields to enhance mathematical and cognitive abilities, and has also served as a benchmark problem for testing the performance of algorithms such as greedy algorithms, beam search, and Chokudai search. This paper proposes a design method for applying genetic algorithms to the Number Collecting Maze game. We present a solution representation for applying genetic algorithms to the Number Collecting Maze game and present design methods for genetic operations such as crossover, mutation, and selection. The genetic algorithm, employing simple genetic operators, outperform the greedy algorithm but underperform the beam search and Chokudai search algorithms, which are specialized for the Number Collecting Maze game. To improve this, we design genetic algorithm enhancement techniques such as initialization mixed with greedy individuals, selection strategy mixed with elite strategy and tournament strategy, and adaptive mutation that has high mutation probability in the early stage of search and low mutation probability in the late stage of search. When applied to a number-collecting game, the proposed reinforced genetic algorithm shows superior performance compared to the beam search and Chokudai search algorithms.

Citation status

* References for papers published after 2024 are currently being built.