Limited Assignment Number Search Algorithm?

Abstract. A nonsystematic iterative search algorithm for hard or over-constrained problems is proposed. This linear time complexity algorithm seeks a (partial) assignment of the problem variables. One step of the algorithm is a special incomplete version of chronological backtracking with constraint propagation. Subsequent searches try to improve the last computed partial assignment. This is done by developing variable and value heuristics based on the results of previous iterations. The algorithm was applied to solve random problems and a large scale timetabling prob-lem from Purdue University.