Issue No. 05 - October (1994 vol. 6)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/69.317698
<p>Backtracking search is frequently applied to solve a constraint-based search problem, but it often suffers from exponential growth of computing time. We present an alternative to backtracking search: local search with conflict minimization. We have applied this general search framework to study a benchmark constraint-based search problem, the n-queens problem. An efficient local search algorithm for the n-queens problem was implemented. This algorithm, running in linear time, does not backtrack. It is capable of finding a solution for extremely large size n-queens problems. For example, on a workstation it can find a solution for 3000000 queens in less than 55 s.</p>
operations research; search problems; minimisation; constraint handling; local search; conflict minimization; n-queens problem; backtracking search; constraint-based search problem; exponential growth; computing time; benchmark constraint-based search problem; local search algorithm; linear time; large size n-queens problems; workstation; nonbacktracking search
R. Sosic and J. Gu, "Efficient Local Search with Conflict Minimization: A Case Study of the n-Queens Problem," in IEEE Transactions on Knowledge & Data Engineering, vol. 6, no. , pp. 661-668, 1994.