A geometrical bound for integer programming with polynomial constraints

Let F1, …, Fs ϵ Z[X1, …, Xn] be quasiconvex polynomials of degree bounded by d ≥ 2. Let L be an upper bound for the binary length of their coefficients. We show that the system F1 ≤ 0, …, Fs ≤ 0 admits an integer solution if and only if there exists such a solution with binary length bounded by (sd)...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Bank, B., Krick, T., Mandel, R., Solernó, P., Budach L.
Formato: SER
Materias:
Acceso en línea:http://hdl.handle.net/20.500.12110/paper_03029743_v529LNCS_n_p121_Bank
Aporte de:
Descripción
Sumario:Let F1, …, Fs ϵ Z[X1, …, Xn] be quasiconvex polynomials of degree bounded by d ≥ 2. Let L be an upper bound for the binary length of their coefficients. We show that the system F1 ≤ 0, …, Fs ≤ 0 admits an integer solution if and only if there exists such a solution with binary length bounded by (sd)cn · L. (Here c > 0 is a constant independent on s, d, n and L). We obtain a similar geometric bound for the corresponding minimization problem. The simply exponential feature of our bound is intrinsic to this problem. © Springer-Verlag Berlin Heidelberg 1991.