q-DHI (Diffie–Hellman Inversion). Given (g,gx,gx2,…,gxq), compute g1/x.
q-BDHI (Bilinear DHI). In a pairing group e:G1×G2→GT: given the same power-tuple, compute e(g,g)1/x∈GT. The decisional version (q-DBDHI) asks to distinguish e(g,g)1/x from random.
q-wBDHI (weak BDHI). Given g,h,gx,…,gxq, compute e(g,h)xq+1 — the assumption behind hierarchical IBE with constant-size ciphertexts.
Relations and attacks
q-DHI ⇒q-SDH-style flexibility; all imply CDH; all fall to [[discrete-log]].
Cheon’s attack applies as for [[q-sdh]]: with q large and q∣p±1, recover x in O~(p/q) — security degrades with the tuple length. Quantum-broken via Shor.
Dodis–Yampolskiy VRF/PRF (2005, ~1000 citations): Fx(m)=e(g,g)1/(x+m) — the standard exponent-inversion PRF, used across anonymous-credential and e-cash literature.
Sakai–Kasahara IBE under q-BDHI is deployed: MIKEY-SAKKE (IETF RFC 6508/6509), the UK-government-mandated standard for secure voice (Secure Chorus).