Solving small and large scale constraint satisfaction problems using a heuristic-based microgenetic algorithm

Microgenetic algorithms (MGAs) are genetic algorithms that use a very small population size (population size < 10). Recently, interest in MGAs has grown because, for some problems, they are able to find solutions with fewer evaluations than genetic algorithms with larger population sizes. This paper introduces two heuristic-based MGAs which quickly find solutions to constraint satisfaction problems. Both of these algorithms outperform a well-known algorithm, the iterative descent method, on most instances of the N-queens problem. We compare these three algorithm on the basis of the mean number of evaluations needed to find solutions to several instances of the N-queens problem.<>