“…Strikingly, even the weakest level-1 representations unify and dominate a vast array of published formulations for different classes of mixed 0-1 linear and quadratic problems (Adams and Johnson, 1994;Adams andSherali, 1986, 1990). Lower-level forms have motivated very efficient algorithms for 0-1 quadratic (Adams and Sherali, 1986) and mixed 0-1 quadratic programs (Adams and Sherali, 1993), as well as for bilinear programs (Sherali and Alameddine, 1992) and linear complementarity problems (Sherali, Krishnamurthy, and Al-Khayyal, 1998). Various linear reformulations of classic 0-1 optimization problems have been characterized (see, e.g., Adams and Johnson, 1994, Hahn and Grant, 1998, Hahn et al, 2001, Johnson, 1992, and cut strengthening methods (Lougee-Heimer and explained in terms of conditional-logic (Sherali, Adams, and Driscoll, 1998) operations.…”