peridynamic

Dear lammps users,
I am using LAMMPS (May 2014) and have some questions about peridynamic:
1.For example, when using pair_style peri/lps, we usually need to set some parameters, such as:
pair_coeff <type 1> <type 2>
My question is: how to determine and for different materials in order to get a correct simulative result?

2. is relative to horizon and is usually set to three times the lattice constant? Why three times ,not two times or four times, five times …

3.From some literatures, I know the lattice constant determines the particle grid in peridynamic. Will the result be more accurate and the cost of calculation increase as long as the lattice constant is small enough ( it means the density of grid points in peridynamic is high enough )?

Thank you very much for your answer
Best regards,

I think those are all Qs for the PD literature. The mail list,
nor LAMMPS, can teach you about PD.

Steve