Complete, or systematic, search methods can be considered the exact opposite of incomplete, stochastic methods. While incomplete methods rely on randomness in exploring only a certain part of the search space, complete methods systematically evaluate every possible solution that can be constructed. This means that if a solution exists it will be found, although the time required is often impractical except in very small problems. However, if the problem is highly constrained, complete search methods can use these constraints to eliminate the infeasible solutions and leave a considerably smaller feasible space to search through. Among the commonly used systematic search methods are Linear Programming (including Integer Programming) and Constraint Programming (or Propagation).