Impact-Based Search in Constraint-based Scheduling

Abstract

A novel adaptation of impact-based search strategies for constraint-based resource scheduling is presented. Search based on impacts applies a general purpose search strategy originally from Linear Integer Programming and recently adapted to Constraint Programming. To my knowledge it is shown for the first time that this strategy is properly applicable to constraint-based scheduling and performs well on the class of job-shop scheduling problems. Evidence is given empirically by comparison with a problem-specific and a random strategy.