The crepant resolution conjecture of Ruan (2006) states that the orbifold cohomology of a Gorensteinorbifold is isomorphic to a semiclassical limit of the quantum cohomology of a crepant resolution.
In 2 dimensions, crepant resolutions of complex Gorenstein quotient singularities (du Val singularities) always exist and are unique, in 3 dimensions they exist[1] but need not be unique as they can be related by flops, and in dimensions greater than 3 they need not exist.
A substitute for crepant resolutions which always exists is a terminal model. Namely, for every variety X over a field of characteristic zero such that X has canonical singularities (for example, rational Gorenstein singularities), there is a variety Y with Q-factorial terminal singularities and a birational projective morphism f: Y → X which is crepant in the sense that KY = f*KX.[2]
Notes
^T. Bridgeland, A. King, M. Reid. J. Amer. Math. Soc. 14 (2001), 535-554. Theorem 1.2.
^C. Birkar, P. Cascini, C. Hacon, J. McKernan. J. Amer. Math. Soc. 23 (2010), 405-468. Corollary 1.4.3.
Reid, Miles (1983), "Minimal models of canonical 3-folds", Algebraic Varieties and Analytic Varieties (Tokyo, 1981), Advanced Studies in Pure Mathematics, vol. 1, North-Holland, pp. 131–180, ISBN978-0-444-86612-7, MR0715649
Ruan, Yongbin (2006), "The cohomology ring of crepant resolutions of orbifolds", Gromov-Witten theory of spin curves and orbifolds, Contemp. Math., vol. 403, Providence, R.I.: American Mathematical Society, pp. 117–126, ISBN978-0-8218-3534-0, MR2234886