Impact-based search strategies for constraint programming

摘要: A key feature of constraint programming is the ability to design specific search strategies solve problems. On contrary, integer solvers have used efficient general-purpose since their earliest implementations. We present a new general purpose strategy for inspired from techniques and based on concept impact variable. The measures importance variable reduction space. Impacts are learned observation domain during we show how restarting can dramatically improve performance. Using impacts solving multiknapsack, magic square, Latin square completion problems shows that this criteria choosing variables values outperform classical strategies.