Solving Kakuro Puzzle using Self Adapting Harmony Search Metaheuristic Algorithm

Abstract

In paper presents a Self Adapting Harmony Search (SAHS) metahauristic based solution to the grid-based Japaneese game know as Kakuro. It has been shown that our approach can provide a time-efficient solution to obtaining a correct positioning of the numbers in the cells of a puzzle, with proper previous filtering of the combination of the numbers in the sets of sums. The pre-computation phase has given great improvements to the SAHS algorithm in order to reduce the search space. The proposed solution has been experimentally evaluated on various grids with different size and the results shown that the algorithm is very robust and always converged to the solution.